Some Thoughts on the State of the Tech Market

I’m in a contemplative mood this week, as this is my final column for Techpinions, and as such I thought I’d share a few big picture thoughts on the state of the tech market in late 2017. It strikes me that from a consumer perspective, in many ways we’ve never had it so good, but at the same time there are new threats and concerns which are also unprecedented. We will therefore be tempted to seek regulatory remedies and limits on the power of tech companies and technology, and while some of these may be worth pursuing, there’s also a danger that we politicize technology and undermine progress even as we seek to protect consumers and startups.

We’ve Never Had it so Good

First off, I’d argue that as consumers of personal technology we’ve never had it so good – the devices we have access to are unprecedented in both their raw power and in their specific capabilities, from cameras to connectivity to displays and audio. And the key thing here is that no single manufacturer either dominates sales or has far and away the best devices: one of the best things about the current state of the market is that consumers have a number of great options in key categories from smartphones to PCs to tablets and TVs. On the smartphone front, Apple and Samsung make the most and arguably the best premium smartphones, but new players like Google and Essential are creating promising new entrants, while the old guard including LG and others continue to produce interesting devices too. On key features like cameras, Apple, Samsung, Google and others all have great performance and it’s mostly a matter of personal preference rather than objectivity which is best.

In the smartphone market in particular, it’s also notable that consumers don’t have to spend the $700-plus that’s now required to buy a top-of-the-line smartphone in order to have a great experience. There are less powerful but still serviceable smartphones available at nearly every price point from $50 to $800, making this technology available to consumers throughout the world and thereby transforming lives and economies. All of this is also true in other categories like tablets and PCs, though low-end PCs still tend to prove the maxim that you get what you pay for more than other categories of consumer hardware.

Technology is an Enormous Force for Good

That last point is worth expanding upon: not only is our technology great, but it has done great good in the world, connecting people with each other and other resources as never before, opening up a world of information and content to anyone, anywhere, on the device of their choosing. The Internet has both allowed even the smallest publisher to reach massive audiences and allowed tiny interest groups to find comradeship across the globe. Technology is connecting families, giving opportunity to poor and otherwise marginalized populations, including the disabled and ethnic minorities.

But It Has Also Created Worrying Side Effects

None of that is to say that technology has created unalloyed good in the world. Many of the same enablers that have permitted innovation, positive communication, community building, and more to flourish have also fed conspirators of various stripes, trolls, and other bad actors and their ability to do nefarious work. Platforms designed to allow people to connect in positive ways have also enabled the spread of misinformation, harassment and abuse, and more recently even meddling in elections. It’s clear that we’re only beginning to discover the scope and potential of some of the negative effects of technology in our lives.

Meanwhile, tech as an industry is characterized by other unpleasant characteristics, notably a lack of diversity and a tendency to downplay or ignore the potential of new technologies for evil as well as good. Too often Silicon Valley demonstrates its lack of diversity in its lack of understanding of how its inventions will impact marginalized populations or even the population as a whole. Its self belief is one of its greatest strengths but also one of its greatest weaknesses. I’ve also pointed out that, with few exceptions, the largest companies in the industry are dominant and threaten to continue to squeeze out innovators.

Regulation is a Tempting Solution

In light of all this, voices from both sides of the political spectrum in the US and beyond are calling more loudly for regulation of big tech companies, whether on antitrust, content, advertising transparency, or other grounds. Some of these calls have obvious merit, and would bring the tech industry in line with older industries that provide similar functions. But my biggest worries with tech regulation are always that those writing the laws have an imperfect understanding of the market and that the process is so slow as to be ineffective in dealing with real problems while often creating unintended consequences. I’m also increasingly aware that in some of these debates a key constituency – media – has an inherent conflict of interest because it’s threatened by some of the very platforms it covers.

I’m hoping that we don’t see knee-jerk, often politically-motivated calls for regulation resulting in laws that would limit the ability of companies to innovate while not really solving the underlying problems. I have little faith that the current US political leadership will get anything meaningful done here without screwing it up, while the bigger threat to US tech currently comes from the EU and its efforts to punish big US tech companies for underpaying taxes and squeezing out local competitors.

A Promising Future

I’m inherently an optimist, and that optimism extends to the tech industry and the role of technology in our lives. I’m not naive enough to think that all the issues will merely go away, but on balance I think the positive benefits will be greater than the drawbacks, and humanity as a whole will continue to benefit enormously from the advances that will be made, especially in areas like healthcare, where consumer tech companies are just starting to scratch the surface of what’s possible. AI and machine learning bring their own threats and downsides, but I tend to think the more apocalyptic voices here are off the mark, while there could be significant benefits from smarter technology in our lives too.

Shifting Dynamics in the US Wireless Market

As I’ve mentioned before, one of the markets I follow closely is the US wireless market, with a focus on the four largest network operators. These operators continue to be by far the largest channel for smartphone sales in the US, and what I’ll share today is a mix of insights on the wireless market itself and the implications for the smartphone market. The data I’m sharing here comes from a much larger data set I maintain on the market, and the charts are also part of the quarterly wireless deck in my decks subscription service.

T-Mobile Continues to Win in Phones

Phones continue to make up by far the largest single chunk of the wireless market, for all the talk about an exploding IoT industry, so we’ll start there. The postpaid model continues to dominate phone subscriptions in the US, and T-Mobile continues to lead the market in growth there, as it has for the last several years:

T-Mobile has added the most postpaid phone subscribers of any US operator for the last four years straight, with Verizon the only operator really giving it a run for its money. Verizon’s growth, though, dropped precipitously during 2016 and Q1 2017 as Sprint and T-Mobile pushed unlimited offerings and Verizon and AT&T resisted that trend. Since reintroducing unlimited offerings at the end of Q1, Verizon has recovered well. Interestingly, AT&T hasn’t recovered nearly as well, and has continued to see shrinkage in its postpaid phone base for the last two-plus years. Sprint saw a big recovery starting in 2013, and has now plateaued around the middle of the market.

T-Mobile Also Sells More Than Its Fair Share of Smartphones

One of the other interesting outgrowths of T-Mobile’s growth leadership is that it punches well above its weight in phone sales. The three main drivers are its outsize share of new postpaid subscribers, strong performance in the prepaid segment, and the higher device upgrade rate among its base, as shown below:

As you can see, here too AT&T comes in last, and by a significant margin, with T-Mobile first for most quarters, though pipped by Sprint in Q3. That disparity and the mismatch in phone subscriber growth translate into a dramatically different share of smartphone sales versus installed base, given that T-Mobile and Sprint are still much smaller in terms of subscribers than Verizon and AT&T:

As you can see, although Verizon has over a third of smartphone subscribers in this group, it sells just 28% of the smartphones, while AT&T’s mismatch is even bigger at 30% to 22%. T-Mobile, meanwhile, sells sells the most smartphones of any of the four, off the back of just a 21% share of the market, and Sprint’s sales share is also well above its base share. All of this means that, if you’re a smartphone vendor, you shouldn’t be swayed by each operator’s share of the smartphone base as much as by its share of sales, which might be considerably higher or lower.

Smartphone Sales Overall are Down Due to Longer Upgrade Cycles

Beyond the competitive dynamics, the other big thing to note from a smartphone perspective is that upgrade rates continue to lengthen. In the chart below, I’ve averaged out the upgrade rate from one of the charts above to create a single number for the four major operators, and I’ve also added in total postpaid smartphone sales for each quarter. As you can see, there’s a strong correlation between the two:

Both are trending downwards, as people hang onto their phones for longer, and therefore buy fewer new phones each year. There are still spikes in sales in Q4 of each year, but the spikes have been getting lower. Those spikes are, of course, driven in part by Christmas present buying, but also to a great extent by new device launches towards the end of the year, notably new iPhones going on sale in late September. The 2014 “super-cycle” driven by Apple’s first large-screen phones drove by far the biggest Q4 of sales ever, and it’s quite likely that we’ll see the first meaningful growth in several years this quarter with the launch of the iPhone X.

But the overall trend will continue to be downward, driven by both that slowing upgrade cycle and by the fact that there are fewer and fewer new smartphone customers to be had – the industry added just 9 million new postpaid smartphone subscribers in the past year, versus 19 million three years earlier. That, in turn, means the market will be more competitive and closer to zero-sum, with any gains by individual vendors necessarily coming at the expense of others. In that context, we’re going to continue to see the push up-market by those vendors with credibility in the super-premium smartphone space, mostly Apple and Samsung, while other vendors are forced to fight for the mid-market.

Snapchat’s Strategic Failure

This week’s Snap Inc earnings call was an indictment of the strategy pursued by the company in regard to both its core Snapchat app and its Spectacles hardware. The company has failed to drive two of the three major metrics that are key to success in the space, and it reversed its long-standing strategic stances on several key topics during a single earnings call. Having resisted calls for change for months, it appears Snap is now trying to change everything at once.

The Multiplier Key to Ad-Based App Growth

There’s a simple formula which is key to growing revenues for any ad-based online or app company:

user growth x engagement growth x rising ad prices

That’s the formula that’s served Facebook so enormously well over the last few years, and it’s also the one which Twitter and others have failed to implement effectively, with Snapchat seemingly the latest company to do so. Snapchat is executing on just one of these adequately – growing engagement and time spent – though it doesn’t consistently report the metrics needed to measure that progress over time. We do know that at the time Snap filed for its IPO, its users spent an average of 25-30 minutes per day in its app, and that number has grown since, but we don’t have any sense of the longer-term trajectory here.

On the other two metrics – user growth and rising ad prices – Snap has fallen woefully short. The user growth it saw in early 2016 was clearly something of a temporary phenomenon rather than a reliable predictor of future growth rates, though the company argued otherwise in its S-1 and clearly expected stronger growth than it’s actually seen since its IPO. Instead, what we’ve got is Twitter-like incrementalism rather than the strong growth that should characterize a social app in its prime:

The line across the chart hits at 10 million users per quarter, a milestone the company beat three times in late 2015 and early 2016, but hasn’t crossed again since, with the most recent quarter at just half that pace. That kind of user growth clearly isn’t going to get Snapchat where it needs to be from a revenue perspective.

Turning to prices per ad, we have no way to measure those directly, but it’s abundantly clear that they’ve been falling rather than rising as Snapchat has introduced programmatic ad buying. That change has lowered the entry point for advertising on Snapchat by three orders of magnitude, per management’s earnings call commentary, and demonstrated in the process that Snapchat’s earlier fixed rate cards were priced vastly above what supply and demand would have dictated. Now that advertisers have a choice in the matter, they’re paying vastly less, largely because there’s so little competition to fill those ad slots.

As such, instead of rapid user growth and rising engagement being multiplied by rising ad prices, Snap has seen revenues grow anemically, with average revenue per user in the US very low at around $2, and outside the US vanishingly small at under 50 cents in Europe and just 30 cents in the rest of the world. That won’t turn Snap into the kind of advertising powerhouse it clearly wants to be, and something needs to change.

The Dam Breaks

All of this has been fairly clear to those of us watching the company with keen eyes from the outside for some time – indeed, I pointed out the terrible timing of Snap’s IPO back in February, citing already apparent slow user growth. And yet within Snap there’s been a resistance to change, which I can only assume has its roots in CEO Evan Spiegel’s conviction that he’s a product genius. He’s been reported to resist data-based approaches to product management, instead favoring his own instincts, which have undeniably created a phenomenon but don’t seem to be serving Snap as well recently.

Now, it appears that the dam has finally broken, and rather than subtly embracing some of the changes that have been called for, Snap appears to be doing a 180 on almost every aspect of its strategy:


  • Design – Snapchat has long been criticized for its unfriendly UI, while Snap’s management has defended it as part of its unique value proposition and argued that those who don’t get it aren’t its target users. Now, it appears poised for a major redesign, likely along the lines described here.
  • Creators and celebrities – Snapchat has famously eschewed the courtship of creators and celebrities common to nearly every other big social platform, arguing that its organic tools were enough to attract and retain them. There’s been evidence for some time that this attitude was leading it to lose share among these groups (and their followers) to Instagram and other platforms. Now it appears that it will belatedly embrace them, though I’d argue it may well be too late for that.
  • Curation/AI – Snapchat has also resisted the pull of AI-based curation of content, instead serving up a relatively small but relatively universal set of content to users in particular geographies, as a result of which it’s lacked the diversity of content and personalization found in other apps. It now appears ready to embrace this strategy too, though it acknowledges the risks inherent in moving away from the relatively confined form of content sharing it’s focused on to date.
  • Pursuing other demographics – Snapchat had for a long time prioritized its iOS apps because that’s where it saw the highest user engagement, something it acknowledged in its S-1 filing. It has belatedly embraced Android, and after months of small tweaks to its Android app is now apparently working on an overhauled version that should fix some of the issues it’s had until now. It’s also trying to broaden the scope of its audience beyond the teenaged and young adult audiences it’s dominated in the past, but that also brings inherent risks as it invites the parents of its current users into what’s felt like an intimate and separate space.
  • Spectacles – Snapchat launched its first hardware product late last year, seemingly out of nowhere, with a clever marketing strategy that focused on artificial scarcity. However, management seemed to completely misinterpret the early sales of the device and ordered way too many units, leading to a write-down of inventory that’s twice the size of its revenue from Spectacles in its first three quarters on sale. This feels like yet another example of gut winning out over data.

The big question is whether any or all of this will actually help to turn around the two metrics where Snap is currently failing, without damaging the one metric – engagement – where it’s been succeeding. I’m certainly skeptical, especially in the context of ongoing inroads by Instagram and others into Snapchat’s once unique mindshare among younger people. The biggest risk is that Snapchat damages the current user experience as it pursues these new features and the users they’re designed to attract, eroding its core strategic advantage: its penetration of its core demographic.

The Big Six in Q3 2017

Every quarter, one of the decks I put together as part of my Jackdaw Research Quarterly Decks Service is a comparison of financial and operating metrics for the “big six” consumer tech companies – Alphabet, Amazon, Apple, Facebook, Microsoft, and Samsung. As I’ve done for several previous quarters, I’m also doing a quick run-through here of some of the highlights from the deck for Q3 2017. You can learn more about the Jackdaw Research Quarterly Decks Service here, and I embedded the full deck for Q2 in last quarter’s post, which you can find here.

Apple Remains Ahead on Long-Term Revenue Despite Samsung’s Gains

Apple has returned to revenue growth over the past year, while Samsung has experienced some of its best growth (and profits) ever during this period. But although Samsung’s revenues have pipped Apple’s during individual quarters, the latter remains ahead on an annualized basis, and will almost certainly remain there given the monster December quarter Apple is predicting.

Meanwhile, Amazon is coming up very quickly behind Samsung, with some of the fastest growth of any of these companies (aided a little this past quarter by the acquisition of Whole Foods and to a lesser extent Middle-Eastern e-commerce property Souq. That’s particularly remarkable because, for much of the period from 2012 to 2014, Amazon and Microsoft were almost neck and neck, but the two took dramatically different paths from 2015 onwards. Microsoft’s own recovery, though, mirrors Apple’s: a strong return to growth in recent quarters, albeit helped a little by LinkedIn.

As the chart below shows, though, Facebook is still the king when it comes to percentage revenue growth, with very strong growth despite its repeated warnings of a coming slowdown driven by saturation of ad load. Meanwhile, the acceleration of both Apple and Microsoft’s growth rates is also easy to see in this second chart:

Apple Also Retakes Single-Quarter Operating Profit Crown from Samsung

In addition to its continued leadership on revenue, Apple this quarter also retook the quarterly operating profit crown from Samsung, after Samsung took a rare lead in Q2 thanks to the strength of its memory business:

On an annualized basis, of course, Apple is still miles ahead given the outsized contributions to its profits of December quarters, and that lead will be extended if it meets its outsized guidance for the current quarter. Further down the list, Amazon continues to be in one of its periodic investment phases, where it sacrifices a few points of margin in pursuit of faster growth, and its operating profits have dropped to almost nothing from already low levels. Meanwhile, Facebook  continues to generate enormous profits, with its first $5 billion operating profit quarter ever, and a net margin of nearly 46%, almost 20 percentage points across the second most profitable company, Microsoft, at 27%. Despite its much smaller revenues, Facebook has several times approached Microsoft’s dollar profit number in the past year and a half.

Samsung Continues to Outspend the Rest on Capex

When it comes to investment in property, plant, and equipment (usually referred to as capital expenditures), Samsung continues to lead the pack by a wide margin in dollar terms:

That rapidly-steepening trajectory over the past year or so is down to the success of Samsung’s memory business, with nearly 70% of capex in Q3 going to its semiconductor unit, a number that’s risen from under half of its capex just over a year ago. It seems to be building capacity as fast as it possibly can to continue to take advantage of the supply constraints in the memory market and the resulting price hikes.

Apple continues to be in second place, with one of the more unpredictable investment trends over time, largely due to its strategy of securing capacity and equipment opportunistically and often ahead of big advancements in its product line. Alphabet’s capex spend, which moderated significantly from 2015 to 2016, has slowly begun got climb again, though mostly in line with revenue growth, maintaining a roughly constant spend level of 10-12% of revenue.

On R&D, Facebook Continues to be the Most Generous

Facebook continues to invest very heavily, just behind Samsung in terms of percentage of revenue spent on capital investment, while its R&D spend is by far the highest of the companies in this group which report it:

Its “R&D intensity” (percent of revenue spent on R&D) has come down quite a bit over the past couple of years, but it’s still well ahead of the other companies here, at over 20%. Alphabet and Microsoft spend at very similar levels of around 15% of revenue, while Samsung and Apple spend smaller percentages. Of course, the dollar spend picture looks very different, with Facebook actually in last place, and Alphabet in first, though four of the five companies – Alphabet, Apple, Microsoft, and Samsung, all spend within a roughly billion-dollar range each quarter (ranging from around $3-4 billion).

On Headcount, Amazon is Entirely in a Class of its Own

Lastly, when it comes to headcount, Amazon has been off in its own class for some time already, but with the Whole Foods acquisition has leapt into the ranks of the top few private employers globally, with over half a million employees.

Even without the Whole Foods acquisition, which added around 90k new employees, Amazon’s organic hiring rate was 47% over the past year, matching Facebook’s. Since the growth rate of the other companies is hard to see on that chart dominated by Amazon, here’s a version of it without Amazon:

On that chart, the underlying trends for the other companies are much easier to see – for three of them (Facebook, Alphabet, and Apple) the trend has been broadly upward, with some slight changes in trajectory over time. For Microsoft, on the other hand, the ride has been bumpier – a big spike driven by the Nokia acquisition in 2014, followed by a fairly rapid decline as it consolidated operations and then thinned them in the wake of the failure of the phone business. More recently, it’s begun hiring again, while the LinkedIn acquisition late last year provided another much smaller inorganic bump. Facebook has actually eclipsed Alphabet for the last two quarters in quarterly hiring as both companies ramp up spending in engineering and sales, and these two companies lead the rest of the pack on this basis, behind Amazon’s incredibly aggressive expansion of its workforce.

Facebook Puts Investors’ Money Where its Mouth is on Security

This week has seen both Congressional hearings on the issue of Russian use of tech platforms to influence US politics and a quarterly financial report from Facebook which made that meddling a theme. There was a stark contrast between the relatively defensive attitude on display in the hearings and the more contrite and proactive approach indicated by Facebook’s earnings call. The social networking company seems far more willing so far than others to put real money behind its efforts to ensure undue political influence is curbed, for better or worse.

Political Fallout Remains Unclear

This week’s hearings were certainly uncomfortable for the representatives (mostly lawyers) sent by the big tech companies to testify in front of Congressional committees. They were on the receiving end of frustrated chastisement from lawmakers, and though they defended and dodged as best they could, they still likely came away licking their wounds. But it’s still far from clear what if any action Congress might take as a result of these hearings – the threats have been general and fairly vague rather than specific enough to warrant real concern on the part of the tech companies.

Certainly, none of those tech companies should consider itself immune from political fallout given the growing unease about their role from both the political right (over perceived bias) and the left (over antitrust and related issues). But for now they may well escape any concrete attempts at additional regulation given their commitments to increased transparency and reporting around political advertising and the like.

Facebook’s Response Stands Out

Though the attitude demonstrated by Facebook in the hearings wasn’t markedly different from that of its peers, its earnings call on Wednesday certainly stood out from the others held by the tech companies appearing before Congress this week. Whereas the others have largely downplayed the financial impact of the current political quagmire in which they find themselves stuck, Facebook’s leadership made a proactive commitment to trying to solve the problem: it committed to spending enough additional money on “security” in the coming months to have a material impact on its profitability going forward.

What’s interesting here is that there is no immediate pressure on Facebook to take these steps, and the decision to take them seems motivated as much by Mark Zuckerberg’s recent conversion to a more responsible and community-centric strategy for Facebook as by any political imperative. The CEO has been through such an abrupt about-face on the issue of Facebook’s role in political meddling over the past year that it’s almost as if he’s a different person from the one who reacted dismissively to suggestions fake news on his platform might have played a role in the outcome of the 2016 elections.

That community-orientedness was evident not only in the security commitments but also in Zuckerberg’s comments on Wednesday’s call about making interaction on Facebook more meaningful – an attempt to return to the company’s roots as a social network rather than the content hub it’s become in recent years. And of course all of this builds on the company’s new mission announced earlier this year to focus on community building and bringing people closer together.

People and Technology but no Change in Business Model

Facebook’s approach to solving what it frames as the security issue involves both people and technology: Facebook will double its ten thousand strong moderation team (many of them likely outside contractors) in the coming year, but will also continue to apply its AI and machine learning skills to the challenge. The cost of additional employees and contractors will therefore have by far the biggest financial impact on the company.

But one thing the company is notably not proposing to do is change its business model, which is interesting because it’s that business model that’s been at the root of the company’s problems with regard to political interference. Making its platforms far less open, and its algorithms more tightly controlled would certainly do a great deal to mitigate election meddling, but it would also have a massive effect on its top line, which continues to grow rapidly despite its dire warnings over the past year of coming slowdowns.

Facebook says that the investments it will make will precede the improvements to the user experience and therefore healthier user and revenue growth in the longer term, but that the latter will come eventually. However, there’s a real risk that all the dollars it spends on neutralizing political criticism with stronger moderation and smarter AI don’t really solve these business model-driven problems and leave it both poorer and in the same place politically as it is now, while big competitors like Google and Twitter try to ride out the political turmoil without making such drastic changes.

Facebook is here putting its investors’ money where Mark Zuckerberg’s mouth is – there was a palpable sense of disappointment at the financial commitment being made here among financial analysts on its earnings call, though also some acknowledgment that the moves might be necessary. The risk here will be borne by those investors, while the upside is still far from clear, and the real motivation seems to be a personal desire from Zuckerberg to feel better about the role his platform plays in the world. The next few months may alternatively demonstrate his wisdom in taking these bold steps or reveal them to be misguided and expensive.

Parsing Google’s Other Revenues

Ever since Google became Alphabet and broke out its “Other Bets” as a separate reporting line, that segment has received a great deal of focus and scrutiny externally. But the fact remains that the Other Bets collectively contributed just 1% of Alphabet’s revenue in the past quarter. Meanwhile, a much larger “Other” reporting line – that belonging to the core Google segment – posted 12% of total Alphabet revenues, and continues to grow rapidly. This non-advertising portion of Google’s business merits a closer look.

The Fastest-Growing Part of Google’s Business

Google’s Other revenues are its fastest-growing segment, and likely a pretty profitable one too, in contrast to the slightly faster-growing Other Bets business, which continues to be massively loss-making. The chart below shows the breakdown of revenues for Alphabet in the quarter just reported:

As you can see, Google Other accounted for 12% of revenues, and that was up from 9% three years ago. The growth rate has fluctuated quite a bit during that period, but for the last two years or so it’s averaged just over 40% year on year. Meanwhile, Google as a whole has managed just 20% or so year on year growth during that period, with even the part that’s focused on selling ads on its own websites, including YouTube and search, only growing by around that same percentage.

Something of a Black Box

However, like the core Google ad business, this part of the company remains something of a black box from a reporting perspective: the company neither breaks out the component revenue lines nor the profitability of this business, so we’re left having to guess at them. We do know that the three most significant components of Other Revenue at Google are the various Google Play stores, its enterprise cloud services business, and its hardware efforts. Google has never provided any meaningful guidance around those numbers, but we can make some reasonable assumptions to arrive at decent estimates.

Other is Still Dominated by Google Play

Although there are three main components to Google’s Other revenues, the reality is that one still dominates: Google Play. That revenue segment – itself dominated by app sales – continues to be by far the largest component of Other revenues, with over 80% of the total, as shown in the chart below:

App store revenues continue to drive the lion’s share of growth too, following something of a stall in 2013-2014. Analytics firm App Annie estimates that app sales through Google Play in 2017 will hit $21 billion in 2017 on a gross basis, of which Google’s cut should be around $6-7 billion on a net basis:

That number continues to be much lower than Apple’s App Store revenues, despite being driven by multiple times the number of downloads. But it’s growing at a rapid rate – around 25% by App Annie’s estimate from 2016 to 2017. In addition to these app revenues, Google’s Play business includes other content stores, including movies, TV shows, books, music, and so on, with music subscription services likely one of the faster-growing contributors to overall revenues in this category.

Cloud is Finally Growing Faster but Still Sub-Scale

Google’s various enterprise and cloud services businesses have been around for years, but they’ve never contributed a very meaningful amount of revenue, though that’s starting to show signs of changing. Google still doesn’t break out the numbers for its enterprise and cloud business, but estimates suggest they are now growing rapidly, and other signals provide further evidence that Google is taking this business much more seriously since the arrival of Diane Greene to run the unit. Commentary on earnings calls – including the recent call covering Q3 earnings – indicates that hiring in this cloud business is a big component of overall hiring, which has recently been at record levels at Google.

Despite all that, though, Google’s cloud revenue continues to lag well behind its two major competitors, Amazon and Microsoft. Those two companies each define their cloud services business slightly differently, but Microsoft’s business has finally overtaken Amazon’s in reported size in the past couple of quarters following rapid growth this year. Meanwhile, Google’s cloud revenues remain a small fraction of either of the scale of the other two:

There’s certainly scope to grow that business more rapidly now that Google is really investing in it, and it’s had some big wins recently. But at the very least it will take years for it to approach the scale of its two big competitors.

Hardware is the Newest and Most Unpredictable Component

The newest and most unpredictable component in Google’s Other revenues is first-party hardware. Google has always reported sales of Nexus devices it sells through its stores in this bucket, and more recently has reported its sales of its own hardware there too, with Chromecasts the major driver until about a year ago, when it dramatically grew its portfolio. As such, Q4 2014, the quarter in which it launched its Pixel phones, its Home speaker, and its WiFi routers, was by far the company’s biggest quarter ever for hardware revenues.

But even then, those revenues were tiny – I debated including here a chart comparing two of the largest hardware businesses – Apple and Samsung’s – with Google’s, but the latter’s revenues wouldn’t even have registered on the chart in that context. Even Microsoft’s revenue from its Surface and Xbox lines vastly outweighs Google’s, at an average of over $3.5 billion to Google’s few hundred million per quarter. Here, too, of course, there’s room to grow in future, but for now Google’s hardware business remains massively sub-scale, something it will have to see to change in the coming years.

Margins are Tougher to Parse

I’ve provided you above with my estimates for revenue from these three major categories, but margins are much harder to discern. It’s likely that Google is losing significant amounts of money on its hardware efforts for now, especially with the recent acquisition of people and IP from HTC. Its app store, on the other hand, is likely highly profitable, though probably not quite as lucrative as Apple’s, given its larger scale and smaller revenues. Its cloud business, meanwhile, is probably profitable, but at a much lower rate than the higher-scale Amazon and Microsoft businesses, each of which is now contributing very healthy profits to its parent company.

Overall, though, these three components of Google’s Other revenues remain three of its most interesting and fastest-growing businesses, and arguably some of those most likely to drive growth and profits for the company in the long term, with hardware likely the furthest from achieving that goal.

The Challenge of Breaking into the US Smartphone Market

This week saw one of the newest entrants in the US smartphone market, Andy Rubin’s Essential, drop its price by $200 from $699 to $499, undoubtedly a response to poor early sales. We’ve also recently seen the debut of the second generation of Google’s Pixel smartphone line, after a first year in which one of the biggest names in consumer technology was unable to shift more than a couple million of its flagship devices. Why is it so tough for even seemingly high-quality smartphones to break into the US market?

Carriers Dominate Sales

The first obvious answer is that the major wireless carriers dominate smartphone sales in the US, and that’s especially true when it comes to premium smartphones, which are predominantly attached to carriers’s postpaid wireless plans. Well over 70% of overall US smartphone sales go through the carriers, and in postpaid that’s likely even higher. Even when sales go through third party retailers like Best Buy, they’re often selected from the limited range of devices stocked by the carriers anyway, so the same effects apply.

That matters because the carriers stock a very limited number of brands on their shelves. For the most part, that means stocking Apple, Samsung, and LG phones essentially across the board, with limited support for other brands at each individual carrier. You might find the occasional BlackBerry, Kyocera, HTC, Motorola, or ZTE device, but shelves in your local store will be dominated by the three big brands. Carriers have little incentive to take risks on the smaller brands when the vast majority of demand is for the big three anyway, especially when the smaller brands offer little positive differentiation. As such, they’re often used to fill small niches but rarely to flesh out a high-end portfolio.

The Big Brands Dominate Share

Partly as a result of the carriers’ limited support for other brands, but in large part also because people have now established preferences and habits which make it hard to get them to switch, the big brands dominate market share. Counterpoint Research estimates that 75% of sales in Q2 were of phones from the three big brands, and 86% from four brands, including ZTE, which is strong in the prepaid segment. At three of the carriers, the big three account for over 80% of sales, and again on the postpaid side alone the dominance would be even more complete.

The reality is that the vast majority of customers are limited in the options they have available, and fairly happy with those options, providing few incentives for them to choose new offerings, or for the carriers to Mae them available.

Exclusives Remain Attractive but Ultimately Fruitless

One of the few ways to break through the carrier resistance to the new is to offer exclusives for attractive new phones, thereby giving the carriers incentives to range a phone in hopes of winning customers. We’ve seen this happen with both the Essential Phone and both generations of Pixels now, with Sprint getting the exclusive on the former and Verizon the exclusive on the latter.

Neither, though, appears to have served either the carriers or their device partners well – both devices have sold poorly despite strong promotional activity from the mobile operators, and both have also shut themselves off from the benefits of being sold by all the major carriers, which every major device vendor has long enjoyed. I’ve written previously about the dubious merits of this strategy for Google, which surely doesn’t need the brand recognition extra promotional activity provides.

AT&T and Nokia long ago proved that exclusives can’t overcome the fundamental resistance to change in the market, and these new entrants should give up on the strategy too. At least the Pixels have the advantage of being sold by the largest of the US postpaid operators, while the poor Essential Phone is stuck on the smallest and weakest of the big four.

Slow and Steady Wins the Race

Both Google and Essential will be hoping that they can change minds in time, even if their early efforts have been fairly fruitless. And there’s some evidence from other brands that a long-term strategy can pay off, though the starting point should likely be somewhere other than premium smartphones. ZTE and TCL-Alcatel have long proven that there is room in the market for low-priced brands targeting the prepaid segment, while Blu has been successful in selling prepaid phones direct to consumers rather than going through carriers.

ZTE’s persistence in particular over the course of many years has finally earned it a slot on a carrier’s postpaid roster for the first time, albeit with what’s likely to be a marginal phone, the Axon M, in AT&T’s portfolio. There is no real precedent, though, for a premium brand to break in by means of this strategy. Brands have instead had to start at the bottom and work their way up – a strategy that arguably worked for Samsung in several markets in the past too. That doesn’t bode well for Essential or Google, though both are deep-pocketed enough to continue to pursue the harder path to success for several more years if they choose to do so.

Analyzing Apple’s Retail Growth

This past week saw the opening of Apple’s newest flagship retail store, this one on Chicago’s Michigan Avenue. The new store is enormous and strikingly designed, and is part of the company’s new model for roughly one-fifth of its nearly 500 stores. It also reflects the shift Apple is trying to make in how people perceive its stores, from which it’s now dropped the “Store” moniker and which it would like customers to think of as “town squares” or gathering places as much as retail experiences. Today, though, I’d like to take a step back from Apple’s reinvention of its stores and look instead at the growth of its store footprint over the last few years, and what it tells us about the role of retail and Apple’s regional focus.

Approaching 500

Any day now, Apple should open its 500th store, a milestone it’s been approaching for some time now. The chart below shows how that number has grown over the past eight years, from 273 stores in September 2009 to 499 at the end of September 2017:

As you can see, the growth rate has sped up and slowed at various points during that period, and in the last year in particular seems to have slowed overall, as the company focused instead on what the new experience should look like and revamped a number of existing stores.

Regional Distribution Favors the US and Europe

Those stores, though, are far from evenly distributed across the globe, with the US still very over-represented and other regions under-represented. The pair of charts below shows the mismatch between the contribution of each region to Apple’s revenues and the mix of stores in those markets. The first chart shows the percentage share of revenues and stores, while the second shows the ratio between the two:

As you can see, the Americas as a region has a share of retail stores which vastly outweighs its share of revenues – 61% of stores, but only 42% of revenues. At the opposite end of the scale is Japan, whose 8 stores (2% of the total) belie its 8% contribution to revenues. Europe is very close to parity between the two measures, while Greater China still has a 2x mismatch between its revenue share and its share of stores, and store presence in the rest of Asia-Pacific outside of Greater China also lags its revenue share.

Growth Over the Last Three Years Favors China, the US, and AP

It’s interesting, then, to look at where Apple has opened new stores over the last few years, with the country breakdown shown in the chart below:

As you can see, a single country stands out starkly here: China. It has seen 31 new stores since May 2014, or over 40% of all the 75 new stores opened during that period. Together with the US, where 16 new stores were opened during that time, it accounts for 63% of total stores opened. Many countries saw either a single net new store or none at all during that period, including Canada. Of the other countries where more than one new store was opened, three are in Europe (France, Italy, and Germany) and two are in AP (Hong Kong and Australia), with one in the Middle East, also reported as part of Apple’s Europe region.

Many New Countries in the Last Few Years

Apple currently has retail stores in 23 countries, with ten of those added since 2011. Of those, three have been in Asia Pac, and all but three have been in regions where Apple has had little penetration previously, including Latin America and the Middle East.

So the story of expansion over the last few years has three parts:

  • Continuing to expand in markets where Apple’s retail presence has been strong, notably the US and to a lesser extent Europe
  • Expanding massively in China, a region that’s suddenly become very important to Apple, and which now has more Apple Stores than any country after the US, passing the UK in the past year
  • Continuing to add a presence in new countries, by way of testing the water – Apple has ten countries with three or fewer stores, and five with just a single store.

Overall, that expansion has slowed a little over the past year as Apple has launched its new strategy and store concepts, but I would guess that as the company returns to growth and the strategy is implemented in a core set of stores, we’ll see it speed up that expansion and continue to focus on those three main sets of markets in much the same way.

Samsung Aims for Connected Thinking at Developer Conference

Samsung is holding its annual Developer Conference this week in San Francisco. At the day one keynote on Wednesday, it pushed a vision centered on “Connected Thinking” as its major theme for not only the conference but its strategy in relation to its software and services in the coming year. That was reflected in a range of moves designed to bring what have been disparate parts of Samsung together, but it’s apparent that this will be a tall order.

A Single Cloud Platform and Bixby as Connective Tissue

Samsung’s major announcements focused on three key topics:

  • Consolidating Samsung’s disparate Internet of Things cloud offerings
  • Iterating on Bixby, by improving the technology, extending it to new devices, and opening it up more
  • Going all-in with Google on AR through ARCore support on all of this year’s flagship phones.

The Internet of Things moves are focused mostly on using the SmartThings brand (now without Samsung as an umbrella brand) as the consumer lead, while consolidating three separate cloud IoT platforms into one, also now tagged with the SmartThings brand. ARTIK survives as a separate IoT brand, but now focused mainly on modules, while its cloud platform along with the Samsung Connect Cloud announced earlier this year will be folded into the SmartThings Cloud.

The way I see this is that the SmartThings Cloud will be the invisible connective tissue on the back end, while Bixby 2.0 eventually becomes the visible connective tissue in the front end as part of a much more coherent and connected vision for Samsung’s range of devices. Samsung executives pointed out during the keynote that it has arguably the largest number and range of devices in use of any company in the world, but the reality is that it’s always been a pretty disparate range of devices, with only fairly superficial integration between them. A big reason for that is Samsung’s operational structure, which has separate CEOs for each product-centric business unit.

The vision Samsung is pushing now is one where a variety of services on these devices will all be powered by the same cloud back-end, and Bixby will become a cloud-based voice interface which works on more and more of them over time. Bixby 2.0 will shift its personalization and training from the device to the cloud, and will therefore start to build profiles of individual users which can be exposed on a variety of devices, including shared devices like TVs and fridges. In addition to its own devices, it’s going to try to extend Bixby support to a variety of third party devices through modules and dongles as part of what it called Project Ambience, which will Bixby-enable existing home devices, both smart and dumb ones, and connect them to each other.

Significant Challenges Lie Ahead

What’s interesting here is that, even though Samsung controls the operating systems on several of its devices, because it doesn’t control by far the biggest – Android on its smartphones – it is instead building the connective layer between its various devices at the interface level. That means pushing Bixby to become far more than it’s been so far, acting not only as a way to perform tasks previously done through touch on a phone, but increasingly allowing for integration with other Samsung devices like TVs and control of smart home gear through SmartThings integrations.

In reality, though, voice can’t be the only interface and therefore can’t be the only connective layer between these various devices – in time, the integration therefore either needs to grow beyond Bixby, or Bixby itself needs to evolve to the point where it’s more than just a voice interface. In the meantime, the SmartThings brand, now decoupled from the Samsung brand to foster a sense of openness, will nonetheless become the brand for Samsung’s own connected home ecosystem too (replacing Samsung Connect), which may cause some customer confusion.

But those aren’t the only barriers to making this vision work: Samsung needs to overcome both internal and external hurdles if it’s to be successful in creating a truly connected ecosystem. The biggest internal barriers continue to be structural – hearing Samsung executives talk about this week’s announcements both on stage and one-on-one, the language is still far more that of separate companies “partnering” rather than a single team working together. The integration announced this week represents progress, but there’s a long way still go go and huge cultural barriers to overcome.

Externally, Samsung needs to convince developers and hardware partners that Bixby is ready for use as a voice platform beyond its smartphones, at a time when it’s got big shortcomings even there. Deeper integration of the Viv technology will certainly help to improve its functionality, as will opening up version two earlier to developers so that the integration can be deeper when it launches to consumers. But the leap Samsung is contemplating here is a huge one, one which other platforms have approached much more gradually and incrementally than Samsung is proposing to do. Samsung would arguably be better served by tackling either third party integration or cross-device support first and then pursuing the other second, rather than trying to do it all at once. The current approach risks over-promising and under-delivering.

The last big challenge is one of adoption – unlike earlier voice assistants, Samsung can’t simply add Bixby to existing hardware, because little of it was designed with voice interfaces in mind. What that means is that it can only grow the Bixby base to the extent that it can grow the base of devices which offer it. In categories like TVs and fridges, that means waiting until next year to even start selling them, and with long refresh cycles, it’ll take many years before penetration is meaningful. Even in smartphones, where Samsung has an installed base of hundreds of millions, it has just 10 million users of Bixby, and we don’t even know how many of those use it daily or weekly. Even if the new SmartThings and Bixby ecosystems work exactly as intended, it will be quite some time before any significant number of consumers actually get to benefit from them.

Q3 2017 Earnings Preview

We’re about to kick of earnings season for Q3 2017, and so I’m doing my usual quarterly preview. My focus here isn’t so much predicting what we’ll see as suggesting the things to look for when these companies report. As usual, I’ll tackle the main companies I track in alphabetical order.


Alphabet’s last set of earnings was pretty impressive, with strong performance pretty much across the board and no obvious areas of weakness. Key trends continued, with ad revenue from  Google’s own sites continuing to grow much faster than ad revenue from third party sites, though growth in the latter has accelerated recently. In some ways, the most interesting revenue line to look at is the “Other” bucket within the Google segment, because that’s where a lot of Google’s new focus areas sit, including its first party hardware push and enterprise cloud services. Both have been major focus areas for Google in the past year, but Q3 should have pretty low hardware revenue given it will have been the lull before new hardware was launched earlier this month, so if there’s strong growth here, that’ll be a good sign cloud services are finally starting to grow commensurate with the investment Google is now making in this area. As I noted last week, one other thing to look out for is traffic acquisition costs in the Google sites business, and how these are tracking relative to revenue.



Amazon demonstrated conclusively last quarter that it’s in the midst of another period of higher investment and therefore lower margins – its revenues grew strongly, but its profits were way down. Investments in AWS capacity, fulfillment infrastructure and employees, and even heavier hiring in higher-paid headquarters roles like software engineers and sales people for AWS and advertising all drove up costs and drove margins down to levels we haven’t seen in a couple of years. It’s likely we’ll see some of those same trends this quarter for the same reasons, with Amazon likely hiring significantly ahead of the holiday season. One of the biggest things to watch for this quarter is how Amazon will report the financials of the Whole Foods business – my guess it that it will simply be an additional reporting line along the lines of AWS, but we’ll have to wait and see. It has historically been more profitable than Amazon’s core business, so it should provide something of a boost to overall margins, again like AWS.


To my mind, by far the most interesting thing to look at in Apple’s earnings will again be its guidance. Its overall revenues and profits for this past quarter should be fairly predictable and in line with the outlook it provided a quarter ago, but there’s a small possibility it will be at the low end of its guidance if iPhone 8 sales in the first few days on sale were less than expected. However, the December quarter is entirely unpredictable at this point, with the iPhone X going on pre-sale right before the earnings call and on retail sale right after. Apple will certainly know what kind of supply it’s likely to have in the remainder of the quarter for that device, and that in turn will to a large extent determine how Apple’s overall December quarter goes. Weak supply could depress overall iPhone sales as many would-be buyers wait out the supply constraints, while strong supply would give Apple a massive quarter off the back of both strong sales and much higher ASPs (I wrote about all this in detail in an earlier piece). Early indications of Watch Series 3 sales and an ongoing reduction in the rate of revenue decline from China are also worth watching for.


Coverage of Facebook in the news recently has been dominated by things that have nothing to do with its financials, at least from a direct perspective, and I’d expect its earnings call to feature a few questions about Russian influence and whether the measures Facebook is taking to mitigate that will have any longer-term impact on its ad business. But ad load saturation and its predicted effect on ad revenue growth is the thing many investors will be watching for, and I’d also expect lots of questions about Facebook’s big video push and the effect that will have on margins as the company invests heavily in content and projects lower margins due to revenue sharing. We got some hints about that on last quarter’s call but I’d expect more detail this quarter as Facebook moves this project along. It’ll also be interesting to see how many new hires it had in Q3, given that it promised faster hiring in the second half of the year.


Microsoft continues its transition from a product-driven to a services-driven company, but the headline on all of its earnings releases for the last two years has been all about the cloud. Microsoft’s growth rate in cloud services ticked up significantly last quarter, and one of my big questions then was whether that was a blip or a sign of a change in trajectory – my guess is the former. Meanwhile, the phone business is finally far enough in the rearview mirror that it should no longer be a drag on the business, while Surface revenue growth following recent product launches should turn positive again this quarter after some declines driven by shifting release cycles. The PC business overall, which of course is a major driver of Windows revenue, continues to be somewhat unpredictable animal from quarter to quarter, and foreign currency has been an ongoing drag on Microsoft’s results overall too.


Netflix will kick off earnings season on Monday afternoon, and its guidance was for just under 4.5 million new subscribers, the vast majority overseas. Netflix’s guidance, though, has been somewhat poor lately, missing on both the high and low side, and it’s always possible that it could see significantly fewer or more net adds. My guess is that it might overshoot its guidance slightly in both geographies, but there’s no reason to expect a significant departure. The bigger question is what its guidance for Q4 looks like, given that it’s just announced price increases which will come into effect in the quarter. I wrote about those price increases here last week, and overall I’d expect them to take a hit to net adds in the US (the only region where the price increases are happening) in the quarter, but still to generate positive net adds there given that it’s usually a healthy quarter for growth. However, the impact this time around will likely come all in the fourth quarter, unlike Netflix’s last increase, so it’s possible that we’ll see a bigger and more concentrated impact this time around. I’d also expect management to be asked about the mix of customers between Netflix’s three service tiers – SD, HD, and 4K – given that the price increases affect these three bands differently.


Samsung has already pre-announced very strong earnings, as well as the impending departure of its CEO. But as usual we’ll have to wait for the full results before we know how the different business units fared. Based on recent results and overall market trends, it’s very likely that both strong demand for and increasingly high prices for memory were major drivers, with the smartphone business likely also having a good quarter off the back of the Galaxy S 8 launch earlier this year. I’d expect there to be some questions about how Samsung will replace its CEO, who suggested in his resignation letter to employees that he felt it was time for some younger blood at the helm.


Snap has been struggling ever since it went public to generate rapid user growth, while some market observers have recently reduced their forecast for its ad revenue growth rate as well. There’s been nothing to suggest any of the underlying trends will have changed much in Q3, with no new obvious revenue generating or growth-inducing features released in the quarter. Ongoing rollout of Snapchat’s self-service and automated tools for ad buying should continue to help drive revenue per user, and I would hope that the company will also be more transparent about some of its engagement metrics, which it hasn’t updated consistently as a public company. CEO Evan Spiegel has recently acknowledged that he needs to be more communicative now that Snap is public, and I hope we’ll see evidence of this during earnings – Snap’s earnings releases so far have been utterly spartan affairs, and management has been cagey and standoffish during the calls.


Twitter continues to be more or less stuck in the same difficult spot as last quarter, with user growth likely not much better, and revenue per user likely continuing to be fairly stagnant. The company says it’s in the midst of yet another strategy shift and revamp of its ad tools, but has shown little concrete evidence that the shift will generate better results going forward. Meanwhile, it’s tinkered at the edges of its product, making changes which aren’t likely to change engagement or user growth meaningfully while leaving larger issues such as abuse and the complexity of on-boarding as a new customer unfixed. Live video has been a big focus, and Twitter will likely update the metrics it has shared previously on this topic, including the number of unique viewers. But we still need to see more information from Twitter about how much time those viewers spend watching video, and whether those views are generating meaningful revenue. I also live in eternal hope that Twitter might at some point finally provide daily active user numbers!

Google and the Disintermediation of Search

This week, the growing amounts Google pays phone makers and other companies to carry its search engine have been in the news as financial analysts have expressed concern over margin pressure. The growth in those traffic acquisition costs is certainly worth watching, but I’d argue that by far the larger strategic threat to Google comes from the growing disintermediation of search, something that’s also been in the news this week.

Google’s Growing Traffic Acquisition Costs

There’s no doubt that Google’s traffic acquisition costs have been growing, not only in absolute terms but as a percentage of revenue. By far the biggest driver of that increase is the increasing cut Google has to pay to Apple, Samsung, and others who give the Google search engine prime placement in their browsers. The chart below shows the percentage of revenue from Google’s own sites which it has paid out in TAC to these partners:

As you can see, that number has risen in various phases, notably from 2011-2013, and again starting in 2015 and continuing through the first half of this year. Overall, the percentage has nearly doubled from 6% to 12% during this eight-year period, and the trajectory continues to be dramatically up and to the right. That reflects the fact that an increasing proportion of Google’s search traffic and revenue now comes through smartphones and especially the iPhone, which likely constitutes a big chunk of its overall TAC payouts.

Disintermediation May Be the Bigger Issue

However, all of this only affects the search revenue Google actually generates and the margins it can drive off the back of that. Certainly, if TAC continues to rise in this way, that should squeeze margins, but the threat of disintermediation could undermine the revenue base on which those margins are generated in the first place. What do I mean by disintermediation here? The fact that many of what would once have been Google searches are now pre-empted by other apps and services before the user ever reaches Google. Here are just a few longer-term examples:

  • Apps: whereas users once used Google as a starting point to reach a variety of websites, they’re far more likely today to visit smartphone apps associated with those sites. To the extent that there’s any searching going on, it likely takes place within the narrower confines of those apps or perhaps an on-device search engine.
  • E-commerce: for online retail specifically, past studies have shown that some 55% of searches now originate not on but on, again cutting Google and its search and ad revenues out of the picture (and in the process allowing Amazon to quietly build its own search advertising business.
  • Voice: people are increasingly using voice interfaces to search for information they once used a text search for, both on mobile devices and increasingly on smart voice speakers like Amazon’s Echo and Google’s own Home products. In many of these cases, even on Google’s platform, there’s currently no ad revenue opportunity associated with that.
  • Bots: Facebook and Microsoft have now both announced integration of AI-based virtual assistants into their messaging platforms, with Microsoft finally launching Cortana in Skype this week after trailing it at last year’s Build conference. These bots will increasingly pre-empt searches because they give users the information they need when they need it in proactive ways.
  • Contextual information: even if AI-powered bots aren’t serving up this information in a messaging context, there are a variety of other ways in which information previously provided reactively is now being provided to users proactively. Snapchat’s addition of Context Cards this week is the latest example of this, offering up restaurant reviews and ride sharing services in the context of Snaps with location tags.

Google clearly recognizes all of this, which is why it’s been one of the biggest proponents of progressive web apps and other approaches which try to reassert the pre-eminence of the web, though it hasn’t had much success with that approach against the continued growth of native apps. But it’s also clearly aware that it may as well try to play in secondary roles where it can, which explains its recent reappearance as the back-end of Apple’s Siri search functions in iOS and macOS, which likely resulted from a bigger financial incentive, which in turn will drive up traffic acquisition costs further. But such concessions are going to be increasingly necessary if Google is to maintain its search and ad revenue growth in the face of these multi-faceted threats.

Context For Netflix’s Price Increase

The price increases Netflix announced this week come around a year after it finished implementing its last set of price increases, a process it spread over several years. Those last price increases occurred at a time when Netflix’s margins were already expanding despite its growing content spending, but this time around, the increase follows pressure on margins from ongoing growth in content spending.

Netflix’s Highly Predictable US Streaming Margins

Netflix is one of the few companies I know of which has set specific long-term margin targets and then made consistent progress towards attaining them. It set a 40% margin target for its US streaming business some time ago, with the intention of hitting that mark in 2020, and for a long time its progress towards that goal was almost linear. Recently, there’s been a little more volatility in that number, but it’s essentially still on track and even ahead of target:

The volatility in the last couple of quarters was due to the timing of new series launches, which were delayed a little this year from the first to the second quarter and therefore moved some costs around a little. But you’ll note from the chart above that the last several quarters have all been within spitting distance of 40%, while Q1 actually exceeded it.

The last set of price increases certainly contributed to that margin progress, hitting in 2014 and 2015 for new customers but in 2016 for existing customers and thereby pushing average revenue per US paid streaming subscriber per month up from $8 to $10, even as cost of revenue per subscriber was falling from $5.30 to $5.00 or so per month. However, since then, cost of revenue per subscriber has actually begun rising, albeit not dramatically, going up around 23 cents per month on an annualized basis over the past year. That, in turn, threatens to slow the progress towards the 40% margin goal and even reverse it, hence the price increases just announced.

Lessons Learned From the Last Increase

As I mentioned above, Netflix chose to stagger the introduction of the last set of price increases, “grandfathering” existing subscribers at the earlier price until prices went up by $2 in 2016, while two single-dollar price increases hit the price for new subscribers in 2014 and 2015. The thinking here was presumably to spread the impact of higher churn from those unhappy about the price increases over a larger number of quarters rather than taking the impact all at once.

In practice, though, the customers that churned did so starting when the price increases for existing customers were announced, earlier in 2016, rather than when most of the increases actually hit later in 2016, as the chart below shows (2016’s numbers are shown in light blue):

As such, Netflix seems to have learned its lesson this time around and has decided to implement the price increases all at once, with the price rising for new customers immediately and for existing customers starting in November. That means it’ll take the whole hit in once quarter – Q4 2017 – rather than over several quarters. Given that Q4 is normally the company’s second strongest for subscriber growth after Q1, that’s probably smart timing, as it’ll likely still manage positive growth overall even with several hundred thousand fewer net adds.

Increasing 4K Impact

One interesting wrinkle in the new price increases is that they affect each of the three tiers of service Netflix offers on the streaming side differently. There’s no price increase for the basic, standard definition, single-stream service; there’s a one dollar price increase for the most popular HD, 2-stream service; and a two-dollar bump in the price for the Ultra HD, 4-stream service. Given that the revenue per US streaming user has historically tracked very closely with the middle tier’s pricing, it’s very likely that this is by far the most popular service, and that the numbers on the UHD and SD plans are small and largely cancel each other out.

However, as more and more people buy not only 4K TVs but also streaming boxes from the major players, all of which have recently been updated with better 4K support, that could start to change. We could therefore see higher uptake of the UHD service at $14 start to push average revenue per user above the $11 we’d see if the past pattern continued. That will make the choice to raise that price by two dollars instead of just one perhaps the most consequential change of all, and one which will likely accelerate the progress towards 40% margins and beyond.

Google is Clearly Serious About Hardware, But Not About Selling Phones

Google’s announcement of its second generation of first-party hardware made clear something that’s been increasingly apparent over the past year: the company is very serious about this business. However, it also reinforced what continues to be a strange paradox: Google doesn’t actually seem to be very serious about selling at least some of that hardware, namely phones.

Another Solid Set of Hardware

Google’s first set of hardware was announced a year ago this week, and was already an impressive start. The Pixel phones were a decent pair of devices with some clever features, good cameras, and exclusive access – at least temporarily – to the Google Assistant which was then just launching. The Google Home was a prettier and in some ways more powerful answer to the Amazon Echo, at a lower price. Daydream VR was a prettier and more usable answer to the Samsung Gear VR. And Google WiFi leaned on Google’s partnership with a couple of traditional WiFi router vendors to create the first mesh WiFi system from a big brand. All told, the hardware released was solid, attractive, and competitive in its various markets.

This year’s hardware builds nicely on last year’s, with refreshed Pixel phones, two new Google Home devices either side of the first in the lineup, and new entrants in several other categories too. The new Pixel phones look like good upgrades on last year’s, and offer some interesting new features while not embracing some of this year’s big trends either entirely or at all. The new Google Home devices will help Google compete more effectively with the low end of Amazon’s lineup while also targeting a part of the market Apple had looked like having to itself. The Pixelbook laptop is as baffling a product as the first Pixel was – a premium device in a category notable mostly for its low prices. Google’s two new categories are wireless earbuds, where its PixelBuds serve as BeatsX competitors at an AirPods price point with a unique AI-based twist, and Google Clips seeks to create a new category which took a somewhat unwarranted chunk out of GoPro’s share price yesterday.

Again, all told this looks like a solid lineup. Google’s recent HTC deal, the fact that it’s updated and broadened the hardware line, and its ongoing public statements all confirm that it’s really serious about making its own hardware, offering a coherent set of products, and driving the same integration and optimization benefits as other integrated vendors have done before.

A Unique and Somewhat Puzzling Approach to Hardware

Google continues to approach the hardware categories it enters in a unique way, however. It downplays the role of hardware itself while playing up the role of software, clearly playing to its strengths as first and foremost a software company, and one which has had little control until now over the details of its hardware design in phones in particular. It’s the argument we would expect Google to make, and it’s clearly capable of achieving in software what others pursue through hardware, which is impressive.

Some of Google’s offerings also feel experimental, and the AI features in both the PixelBuds and Clips seem like good examples of that. Real-time translation and the idea of a camera that automatically takes the shots you want are both impressive demonstrations of Google’s AI chops, but neither is a product people are clamoring for, nor are the features ones people are actually likely to use regularly. How often do you need real-time language translation? And how likely is it that both you and your conversation partner will have the necessary hardware and software to make it happen? Google is honest about the fact that it doesn’t expect Google Clips to be a big seller, and that’s a good thing – creating new hardware categories is tough for anyone, but especially for companies without a significant presence or history in the space.

Given that ChromeOS has actually done pretty well in the segments where it feels relevant, while Android Wear continues to languish, some of Google’s choices about where to pursue a first party hardware strategy seem a little puzzling. Why not show the Android OEMs how to make a decent smartwatch, rather than ceding much of the market to two platforms – Apple’s and Samsung’s – it doesn’t control? Why pursue premium Chromebooks instead? Showcasing both Android apps and Google Assistant on a Chromebook would be just as possible at $500 as at $1000 if that’s the intention here.

Marketing Continues to be the Biggest Challenge

Above all, though, it feels like marketing continues to be Google’s biggest challenge when it comes to smartphones in particular. The products are there, but Google’s approach to the other 4 P’s of marketing continues to be lacking:

  • Price – last year Google matched iPhone pricing exactly, but this year it has a $200 price differential between phones which it explicitly said don’t have any feature differences, in a market where $100 size differentials have become the norm. That makes the Pixel 2 XL more expensive than Apple’s base iPhone 8 Plus, which seems an odd decision.
  • Promotion – as with pricing, much of Google’s marketing last year seemed aimed directly at the iPhone, not surprising from a company which doesn’t want to be seen to undermine its OEM partners. And yet all the evidence suggests that people tend to be fairly loyal to the two big ecosystems, and there’s relatively little switching from iOS to Android. Everyone else knows that Pixels are for Google-centric people, and Google needs to embrace that in its marketing rather than making sarcastic digs at the iPhone. Its advertising should be about what its phones do uniquely well for people who love Google and its services, because that’s the niche it’s really going after.
  • People and Placement – ever since the launch of the first Android phone, Google has underestimated the role of people in selling and supporting phones, and that still doesn’t seem to have changed. Its own channel is exclusively online, and it really hasn’t invested in the kind of third party retail presence necessary to effectively market a phone. But from a US perspective in particular the biggest stumble Google has made continues to be its exclusive relationship with Verizon on the carrier side. Carriers are by far the biggest channel for smartphone sales in the US, and limiting itself to one carrier – generally not the one seen as the most forward looking either – has been a huge mistake, one Google has repeated this year. (Indeed, I was told today that the Google-Verizon exclusive is a three-year deal, and if that’s true it means Google can’t extricate itself from this mess anytime soon.)

All of this adds up to a bizarre picture of a Google which is enormously serious about building the best possible hardware, but doesn’t seem very serious at all about actually selling it. Given the scale of both its organic investment in the Pixel line and now its billion-dollar-plus HTC deal, Google is pouring massive resources into this project, but it will never see a reasonable return unless these devices sell. It’s unclear to me whether this is a deliberate strategy on Google’s part to limit the negative impact on its hardware partners, or the result of organizational schizophrenia, but it simply doesn’t make sense. Either Google is serious about this market or it isn’t, and if it is it needs to bite the bullet and go ahead and compete with its OEM partners more directly. If that pushes them to do better, that’s good for Google too, and if it doesn’t Google gets a greater share of the premium Android smartphone market and gets to put its own services front and center. At this point, there’s no viable alternative to Android for independent phone makers, so there’s little if any risk to the strategy.

Roku’s Big Shift

Roku went public this past week, and saw its stock soar in its first couple of days on the NASDAQ. The timing of its IPO is closely tied to a shift in its business model the company implemented a few years ago and which is now beginning to bear fruit. Understanding that shift is critical both to understanding the company at this stage, and understanding whether or not the big bets many investors are making on it are sensible.

From Players to Platform

The big shift Roku has made strategically, and is still in the process of executing on now, is a pivot from making money from the sale of smart TV set top boxes to making money from the usage of its TV platform. Roku’s IPO filing provided limited historical financial data, but in the first half of 2015, for example, the company derived 85% of revenue from selling its players and just 15% from its platform, including licensing of that platform to third parties, advertising, revenue share from subscriptions sold through its platform, and so on. By the first half of 2017, by contrast, the revenue split was 59/41, and in Q2 2017 it was 54/46, nearing an even split between player and platform revenue.

Big Implications for Margins

This shift in business model comes with big implications for margins: gross margins on players have never been very high, and Roku now explicitly says that it isn’t optimizing for margins but for creating the largest possible audience for its platform. Platform revenues, however, are pretty high margin, because there are few costs and a growing revenue stream associated with it.

As you can see, gross margins in the platform segment are much higher than in the player segment, and that’s likely to continue to be the case. As a result of the shifting revenue mix, overall gross margins have improved over the past couple of years, but this hasn’t yet flowed through to sustainable profits at the operating or net income level.

As such, to bet on Roku is to bet that it can continue this transition from being focused on selling set top boxes at a profit to maximizing its audience and then monetizing that audience, so that it can finally turn a profit regularly. It’s certainly made good progress on both those fronts so far, growing its active user base from 4 million in early 2014 to 15 million today, while driving average revenue per user from under four dollars annually to over eleven dollars in the past year.

Roku’s Future Remains Uncertain

But there isn’t yet a clear linear correlation between the transition and operating margins, in part because even as gross margins have improved, the company has increased R&D spending from 16% to 20% of revenue in the last two years, to fund the innovation necessary to remain competitive in a market which also included three of the largest consumer technology companies in the world: Amazon, Apple, and Google.

And that’s easily the single biggest challenge to Roku in the future: it’s succeeded in large part by offering decent boxes at competitive prices without tying users into any particular ecosystem, at a time when Apple TV has been missing Amazon Prime Video, Android TV has failed to take off, and Amazon has taken its time adding Alexa voice features to Fire TV. But all of those things either are now changing or could soon change, with Amazon’s video app coming to the Apple TV for the first time, Nvidia and Sharp devices running Android TV getting Google Assistant support, and Amazon adding Alexa to the Fire TV lineup.

As that happens, competition in the smart TV set top box space will intensify, and features like voice control will become table stakes, forcing Roku to hire and attempt to catch up with those very deep-pocketed competitors, which are fighting for dominance of far more than just TV boxes. At the same time, Roku’s future is unusually tied up in advertising-based revenue streams for a company in the streaming and cord-cutting space. To be sure, advertising will have a role there as it has in the legacy TV space, but much of streaming today either eschews advertising entirely or makes it optional, with the largest ad-based streaming property – YouTube – not paying Roku a dime.

As such, there’s plenty of potential upside in Roku’s transition to a platform company, but there’s also plenty of risk. It’s got a decent strategy, and may yet carve out a profitable niche for itself among the consumer tech giants, but it’s going to have to both execute pretty flawlessly and catch its fair share of luck in the next few years.

Amazon Matures its Echo Lineup

Amazon on Wednesday held an event to update and expand its Echo lineup, and in the process demonstrated how the product line is maturing. There are more products, but they’re also more clearly targeted at specific segments of the market, while still undercutting pretty much all major competitors on price. Amazon clearly isn’t ready to cede the market to Google, Apple, or anyone else likely to enter in the near future.

Increasing Segmentation in the Echo Lineup

The first thing to note about the new Echo lineup is the increasing segmentation of the market. Though I’ve seen some reacting to Amazon’s new lineup as if it’s essentially random and experimental, I think there’s a lot more structure to it than it might at first seem. The illustration below lays out Amazon’s Echo product line before and after this week’s announcements:

Before this week’s announcements, the Echo portfolio looked like a bit of a hodgepodge, with two specialty devices, one device with a screen, and one cheap and one relatively expensive straight voice speaker. After this week’s announcements, the picture is a lot clearer, especially in the first two columns in the chart above:

  • A good, better, best lineup in the core voice speaker space, with list prices at $50 intervals and likely discounts to even more competitive prices. The core Echo is now also $50 cheaper, significantly better looking (and with more design options), and with better audio.
  • A two-size lineup in the currently much smaller (and pricier) speaker-with-a-screen segment, doing to that segment what the Dot has long done in the core voice speaker space and offering the videoconferencing and other video benefits in a smaller device, likely for secondary rooms in the home
  • There remain two specialty devices, one for fashionistas and the other for those that want to take Echo-like functionality on the go, which will likely continue to account for a very small proportion of sales
  • There are now also two accessories, which expand the addressable market and activities associated with the Echo line, into voice calling using a landline and family gaming.

At this point, the vast majority of customers will mix and match products based on how much they care about audio quality and video, in a variety of rooms in the home, using Dots for secondary spaces and Echo or Echo Plus units in primary spaces, with the occasional device with a screen for customers who care about video conferencing or another endpoint for Prime video.

Addressing a Broader Market

The range of activities enabled by the Echo line has been expanded mostly through software since the original launch, but over the last few months Amazon has added considerable specialization through hardware too, from the screens and cameras on the Show/Spot and Look to the smart home hub in the new Echo Plus to the accessories.

By addressing a wider set of use cases, Amazon is clearly looking to expand the market for which Echo devices will be attractive beyond those looking for voice-driven kitchen timers, user-friendly but low-quality and unattractive speakers for music, or an alternative to turning on the radio in the morning. As such, these devices start to compete indirectly with more products which would previously have set in fundamentally different categories, including game-centric TV boxes like the Apple TV, VoIP services and devices, smart home hubs like those from SmartThings and others, and so on.

Amazon’s ambitions with the Echo have always been broad, but they’ve been achieved largely by taking on a new, voice-driven interface with the limited set of tasks that are well suited to voice control through a device with finite capabilities. Its broader ambitions are now more easily realized as it leverages its early dominance in voice speakers into these new market segments, and I would expect it to take a greater share of the segments of the market it addresses in the coming year than in the past year, despite Google’s likely introduction of a Dot competitor next week.

The High End is Still Up for Grabs

That last sentence, though, raises the question of what will happen in the segments of the market Amazon doesn’t address with its first party lineup, especially the high end market. That market doesn’t really exist today for voice speakers specifically, and is mostly limited to standalone speakers from premium audio companies and Sonos, which will hold its own event next week, one that’s widely expected to feature voice as a major new feature.

Apple, of course, will also shortly enter the market at the high end, emphasizing many of the same historical strengths that have driven Sonos’s dominance of the mid-tier whole-home audio market: quality, ease of use, and a focus on music. Amazon still seems relatively uninterested in going after the premium part of the market, in part because that’s a strategy for those who want to drive high margins from their hardware rather than for those who are using voice speakers as a loss leader for building a voice ecosystem. Both Amazon and Google seem likely to try to address that premium end of the market mostly through partners, though Amazon has moved ever so slightly into higher quality audio in recent months. Google is reportedly working on a higher-end Home speaker, but we’ll have to see whether that actually launches and where in the market it’s pitched.

The diagram below indicates how I see this market playing out over the next few months, with Sonos and Apple both entering towards the premium end of the market, but Apple likely pricing above Sonos based on the pricing of Sonos’ current lineup:

Amazon doesn’t have a strong incentive to pursue the premium part of the market for today, while Google’s participation in that part of the market is still theoretical at this point. That means that we may well see the same dynamic playing out in voice speakers which we’ve seen in smartphones, tablets, and even connected TV boxes, with Apple capturing the small but highly profitable premium segment while Google (and in this case Amazon) captures the lower-margin mass market. Given that there’s likely to be close alignment between those who already favor Apple’s devices in those other categories and those who will prefer the premium approach in the voice speaker market, that likely won’t present problems for Apple’s ecosystem. Sonos, meanwhile, may find itself squeezed between increasingly high-quality voice speakers from Amazon and Google (and their partners) and Apple’s HomePod, and will have to ensure that it’s unique value proposition around multi-room audio really stands out in that mid market.

iPhone Demand and Apple Guidance

This past week, there was lots of coverage of perceived demand for the iPhone 8 models in the media. Shorter lines at retail stores were seen as evidence of poor sales, and there was the usual handwringing about what it might mean for Apple. To put all this in context, it’s helpful to look at Apple’s financial guidance for the September quarter and see what that tells us about what Apple was expecting by way of early iPhone sales, and whether it’s still on track.

Apple Guidance: Growth of 5-11%

Apple’s guidance for the September quarter is as follows:

  • revenue between $49 billion and $52 billion
  • gross margin between 37.5 percent and 38 percent
  • operating expenses between $6.7 billion and $6.8 billion
  • other income/(expense) of $500 million
  • tax rate of 25.5 percent.

We’ll focus here on the revenue portion for the most part. That $49-52 billion range represents a year on year revenue growth range of 5-11% compared to the September quarter in 2016, when Apple reported total revenue of $46.85 billion. For context, Apple’s last four reported quarters had growth rates of -9%, 3%, 5%, and 7% respectively. So the midpoint of the guidance, at 8%, would be roughly on track with the recent trend of rising revenue growth. In dollar terms, the range suggests between $2.15 and $5.15 billion year on year revenue growth.

iPhone Versus the Rest

Now, given that the iPhone is so dominant in Apple’s overall results, it’s tempting to look at it in isolation, but of course it isn’t the only product Apple sells, and Apple’s year on year growth is also heavily dependent on what’s happening in the rest of its business. One under appreciated facet of Apple’s period of negative growth during the iPhone 6s cycle was that several of its other products also experienced revenue declines during that time, exacerbating the iPhone shrinkage. Conversely, the June 2017 quarter was the first one in which every product line reported revenue growth in several years. So Apple enters the September quarter with the wind behind its back.

We can make some reasonable assumptions about how other products will have performed in the September quarter from a growth perspective as a foundation for talking about iPhone growth and guidance. Based on recent trends, I would forecast a roughly $1.5 billion year on year growth in revenues from products other than the iPhone in the September quarter. That would get Apple 70% of its way to the low end of its revenue guidance by itself, leaving just $650 million of growth to come from the iPhone. Hitting the high end of its guidance, meanwhile, would require $3.7 billion in additional revenue growth from the iPhone. This is illustrated in the chart below:

What the iPhone Has to Do

Now that we know what the iPhone has to do in order to help hit guidance, we can parse that out a little more. We’ve already said the revenue range is $0.7-3.7 billion dollars, but let’s put that in percentage terms. That translates to a range of 2-13% of last year’s September quarter iPhone revenue. So the iPhone has to achieve somewhere in that growth range if it’s to help Apple hit its overall revenue guidance for the quarter.

How hard is that? Well, the revenue growth for the iPhone for the previous three quarters was 5%, 1%, and 3%, so it’s been within that range for two of those three quarters, but very much at the low end of it. More importantly, Apple’s revenue in those quarters was driven entirely by demand, while during the very small portion of the September quarter when new phones are on sale is historically driven entirely by supply – i.e. how many iPhones Apple can manufacture. Normally, then, whatever the percentage growth rate in iPhone revenue is required, Apple has to either make that many more iPhones or raise the average selling price to make up the difference.

This quarter, predicting that mix is particularly challenging, for three reasons:

  • iPhone demand – demand for the iPhone 8 is being impacted in a unique way by the announcement of the iPhone X, which seems to be causing at least some who would normally buy a new iPhone during the early sales period to wait
  • Pricing – Apple raised the prices of the iPhone 8 models relative to the iPhone 7 equivalents, by between $30 and $50 for the base configurations, which could theoretically drive ASPs up
  • Mix – but because some of the normal premium buyers are holding out for the iPhone X, the mix of early buyers likely won’t skew as much to the high end as it normally does.

Demand and ASP Projections

As such, though there is a driver that could push up ASP – the new prices – there’s another driver that could push it down – mix – while demand is also less predictable than usual. However, given that Apple typically sells out of its entire launch inventory and remains extremely supply constrained in the first ten days or so of sales that falls into the September quarter most years, we can reasonably assume that unless there has been a massive drop-off in demand for these models relative to last year’s, Apple will still sell out and lower demand will not by itself depress sales in the September quarter. In other words, Apple will still sell all the iPhones it can make in the quarter, and the question is therefore whether it was able to manufacture more iPhone 8 models for launch than it did iPhone 7 models last year. Given that many of the components and the overall shape of the device are unchanged, that’s certainly possible.

Then the question becomes ASPs. One scenario is that they drop from last year because of the mix shift among early buyers. Say they drop from $619, where they were last year, to $600: that’s about a 3% drop, meaning that Apple would have to manufacture and sell around 5% more iPhones in total to hit the low end of its guidance for the quarter, and considerably more than that to hit its high end (or even the midpoint). On the other hand, last year’s mix was likely impacted by the fact that some Plus models (notably the jet black finish) were in very short supply early on, so it’s possible that the mix shift this year will wash out much the same, leaving Apple needing to boost sales by just 2% to hit the low end of its guidance. I think the stretch scenario is a rise in ASPs off the back of the new pricing – early orders will still come predominantly from people who care about the newest, best thing, and therefore will skew towards the new models, which all start at $700 or above, which could push ASP higher, leaving Apple a much lower bar to clear in terms of increased unit shipments. It’s simply impossible to predict with any accuracy, but it’s fair to say that what happens to ASPs will be the biggest determinant in whether Apple hits its guidance.

At the end of the day, Apple knew all this when it issued guidance – what the new phones would look like, how many it could likely manufacture, the fact that it would also announce the iPhone X at the same time, and so on. So it based that guidance on its best estimates of what would happen in that scenario, which would play out just a couple of months after the guidance was issued. The reality is, though, that this is an unprecedented launch for Apple, and one which it would likely have struggled to predict accurately too. At the end of the day, I’m inclined to lean on the fact that Apple has been relatively conservative in its guidance recently, and will likely clear at least the low end of its range, and might still end up somewhere in the middle of it. I think it’s relatively unlikely, on the other hand, that it hits somewhere in the high end of its range given what we’ve seen so far, barring some big surprise in sales of its other products, like another quarter of unexpectedly strong iPad growth.

Beyond the September Quarter

Finally, it’s worth looking beyond the September quarter, which after all is so dependent on the few days right at the end of the quarter when the new iPhones are even on sale, and to the December quarter, which is always by far Apple’s largest. It’s in that quarter when the vast majority of early sales of the iPhone 8 line will occur, but it’s also when sales of the iPhone X will begin, roughly one third into the quarter. To my mind, the question of whether there is adequate supply of that device to meet demand at some point in the quarter is the single biggest factor in how the quarter will go, given how many people seem to be waiting to see or buy it rather than simply buying the iPhone 8.

It’s possible that some people, having seen it in person, will simply decide they want the iPhone 8 instead, and that demand for that device will be very strong and easily carry Apple through its quarter. But it’s also possible that a large number of people will want but not yet be able to buy the iPhone X in the quarter, and that many sales will therefore be pushed back into the March quarter. iPhone 8 sales (and sales of the other devices which still make up a considerable portion of the overall sales mix for iPhones) will be strong throughout the quarter, but whether we see strong overall growth off the back of both higher unit sales and much higher ASPs will be entirely dependent on the share of iPhone X models in that mix, which in turn will be heavily dependent on availability. At this point, that December quarter therefore looks a lot more uncertain than the September quarter on that basis, and so I’m very curious to see what guidance and management commentary looks like a month from now when Apple reports its September quarter results.

All of this will eventually wash out over the course of the next year. Apple will reach the point, likely sometime in the March quarter, when it can meet demand for the iPhone X with adequate supply, and it’ll likely see a very strong year of growth off the back of the new products. But until then, given the uncertainties of supply, it’s going to be an unusually unpredictable couple of quarters for iPhone sales and revenue.

It’s All Change at Nest Except the Business Model

Nest this week make its biggest ever announcements and introduced its first truly new hardware category since it acquired Dropcam in 2014. Its product line is now extensive and with partner devices thrown into the mix covers much of the addressable smart home market, with all of its own products overhauled in some fashion in the past year. But one thing remains stubbornly unchanged at Nest: its business model. And that may now be the biggest thing holding it back.

An Explosion in Activity After Years of Minor Change

The past year has seen an explosion in activity from Nest after years of relatively minor change and incremental updates. The picture below illustrates what’s happened to Nest’s portfolio of products since its inception, and it’s clear how different 2017 has been:

From its inception in 2011 with the original smart thermostat through the acquisition by Google and acquisition of Dropcam shortly thereafter, Nest created or acquired three main product lines, and from 2014 to 2016 that didn’t change. The Dropcam products adopted Nest branding and got some updates and new variants, but there was no dramatic change. Then, in 2017, the cameras got big updates with much smarter technology, Nest introduced its first cheaper (and less obtrusive) thermostat, and this week it announced its first doorbell product and a home security system. All of this came after the ouster of founder Tony Fadell and although that’s likely in part a coincidence, it’s notable how much more quickly the company has appeared to be moving in the past year.

But the Business Model Remains the Same

However, for all the new products announced over the past year, Nest’s business model remains the same: this is fundamentally an off-the-shelf, pay-upfront, do-it-yourself model, the same as it’s been since the beginning. And as I’ve argued before, that model has severe limitations in terms of its addressable market. Just consider the prices of Nest’s top of the line products:

  • Nest Cam IQ outdoor: $349
  • Nest Cam IQ indoor: $299
  • Nest Secure: starts at $499
  • Nest Smart Thermostat: $249.

Several of these products will need to come in multiples to be useful, so those prices should likely be multiplied to get a real sense of what they’ll cost. That alone will make them cost prohibitive for many customers, but add in the intimidation factor of fiddling with thermostat wiring and the attendant risk of electric shock, drilling through walls to install a security camera, or trying to troubleshoot devices that won’t maintain a reliable connection to WiFi, and you further limit the addressable market.

The (DIY) Smart Home is Stuck

That’s why I’ve been arguing for quite some time now that the biggest thing the smart home market needs to go mainstream is a service model in which professionals install and manage the system and charge a monthly fee which recoups the cost of the hardware rather than charging for it upfront. That lowers the price barrier to entry considerably and means that those not willing or able to install or manage devices themselves can still participate in the smart home.

When new CEO Marwan Fawaz took over from Fadell, I posited that his background at Motorola and other companies which worked through carriers to support a service model might mean that we’d see more of this kind of thing from Nest going forward. But although Nest devices are included in some third party smart home services, Nest still hasn’t created its own, in contrast to players like Comcast, Vivint, AT&T, or the alarm companies. Indeed, at this week’s event it used the self-install model as a major selling point in contrast to having to wait around for a technician to come and spend hours installing a system. That may well be a selling point for the minority who feel comfortable with that model, but I worry that Nest is shutting itself off from a much larger addressable market by restricting itself to it.

A Foundation is in Place for a Managed Service

Nest already has a foundation in place for a services model, as it offers the Nest Aware service for monitoring cameras, and now has a partnership with Moni for 24/7 monitoring around its security system. It’s building a subscription model, but it’s entirely based on either third parties or automated systems today rather than incorporating the human touch in installation and management. Nest even offers to help you find an installer for your new Nest products, but that’s still an arm’s length relationship today and doesn’t offer the brand guarantee that could come from a truly integrated service.

To be sure, there’s still likely quite a bit of growth available for Nest in its current model, by expanding it to new markets and now expanding the line of products it sells. For now, that may be enough to sustain its business for the next couple of years, but there’s a much bigger opportunity out there if it takes many of the components it already has in place and turns them into a managed service.

Competition in the Smart TV Box Space

With Apple’s announcement last week of its new 4K Apple TV, it reinforced its positioning in the market, which remains remarkably distinct from the other three major players competing for US buyers. So far, that strategy has seen it take fourth place in market share, something that seems unlikely to change going forward. Why is that? And does it matter?

The Market Context

Shortly before Apple’s event, I put together some charts comparing prices for the smart TV boxes from the four major players in the US market – Roku, Google, Amazon, and Apple. The first chart shows the price ranges for each company in this space:

There’s a very clear pattern here – three players compete almost entirely in the $35-100 space, while Apple competes entirely in the $150-200 space, well above the others. The picture is even starker when we look at the discrete price points each company offered within that range before Apple’s recent announcement:

Two players stand out here – certainly, Apple again with its vastly different price range, but also Roku with its many different price points, with six options between $30 and $110.

Apple’s 4K TV Announcement

Leading up to Apple’s 4K box announcement, one might have concluded that the existing Apple TV would drop in price by $50 or so to get it down into that $100 range to compete more closely with Roku and Amazon’s high end, but instead it stuck with the same base pricing and introduced new models within the existing price bracket. Instead, the first of the two charts above remains completely unchanged following Apple’s event, with the old Apple TV remaining at $149. The second chart, meanwhile, only changes slightly, with an additional price point between the two former prices:

Apple discontinued the larger-storage version of its older box and slotted the two new 4K boxes in at $179 and $199, keeping the overall price range the same, but now with three options instead of two.

Business Models Determine Pricing

I’ve heard quite a few people suggest that Apple somehow doesn’t get this market, or that it’s failing because it’s pricing its product all wrong and avoiding the “stick” market in which the other three players compete. But the biggest difference here isn’t really pricing but the business model behind that pricing, and that in turn determines where and how the players compete. The reality is that at this point three of the four players are in this market for reasons other than making money on hardware:

  • Amazon wants to drive its broader ecosystem including its Prime Video service and its Alexa voice technology, and to do that it’s willing to sell various hardware products at breakeven or even at a loss across a number of different categories
  • Google wants to add value to Android smartphones and as such sells cheap dongles for TVs and stereos which act as outputs for audio and video from those devices
  • Roku made very clear in its recent S-1 IPO filing that its business model involves getting as many people as possible using its platform, and that it doesn’t actually prioritize making money on hardware anymore, instead licensing its platform essentially for free and pricing its boxes to sell.

The business models that drive these three players, then, all lend them to sell devices at far lower margins than Apple is willing to, because they each intend to monetize in other ways. Apple’s business model, meanwhile, has always been to craft and sell premium devices which it also sells as a premium, with any additional revenue from services a bonus on top. To provide a cheap but inferior experience on a TV stick would go against Apple’s DNA and require a major shift in how it thinks about hardware.

Apple’s Market Share Suffers as a Result

With all three other players pricing their products to sell rather than generate profits, it’s no surprise that Apple’s premium strategy has left it with by far the lowest market share among the four, as shown in the chart below, based on recent eMarketer data for the US market:

The big questions at this point are whether this will ever change, and whether it matters. The answer to the first question is “probably not” – changing Apple’s approach to the TV box at this point would be a huge shift and entirely uncharacteristic. Apple sees hardware as the key locus of value in the broader value chain, and to subsidize or give away its boxes in the TV space would be a signal that the real value lies elsewhere. I think that’s therefore very unlikely, even if Apple eventually begins selling its own premium subscription video service.

The answer to the second question is more complex, but I think still fairly clear. Firstly, Apple knows better than perhaps any other company that TVs are far from the only devices people use to watch video content, and its other devices including iPhones and iPads are gaining increasing share in the broader market, with many subscriptions to video services being both bought and being consumed on those devices. Its share of smartphone and tablet viewing is vastly higher than its share in the smart TV box market, and that’s arguably far more strategic. But it’s also worth noting that Apple has never been in the business of market share for its own sake, and that’s also unlikely to change. Looked at in their entirety, the smartphone, tablet, and PC markets all dwarf Apple’s market share, though it takes a far larger slice of the premium segment in each of those markets. And that’s pretty much analogous to Apple’s position in the smart TV box market, taking almost the entirety of the $100-plus segment.

Ultimately, Apple seeks to provide a premium experience in any category where it operates, and often concedes the majority of the market to others as a result. However, it also often captures a disproportionately high share of actual usage and especially revenue generated from its base of users, and that’s almost certainly the case in the smart TV box market too. And that should continue to serve Apple very well as it prioritizes generating subscription service revenue through both its own and third party services from customers willing to pay.

The Role of Marketing and Business Models in iPhone ASPs

With the launch of new iPhones this week, and specifically the creation of a fourth tier in the current generation lineup, I’ve been talking to people over the last couple of days about price quite a bit, and thinking about average selling prices for iPhones, and how Apple has been able to raise ASPs even as it’s extended the iPhone lineup down market, which is a pretty unusual feat.

A brief history of iPhone pricing

A good place to start this analysis is with a brief review of the last few years’ iPhone lineup and the price range implied by it, as well as the average selling prices that resulted from it. The diagram below shows how the iPhone lineup has expanded over the last few years from a single current model to four (ignoring the fact that older models have stuck around at discounted prices as well):

That expansion in the range largely reflects the maturation of the market – when markets are in their infancy, product lines can be simple because there’s plenty of addressable market to go around, but as they mature and become increasingly saturated, diversification in the product line becomes necessary to meet a wider range of use cases for increasingly sophisticated buyers. It was also clearly in part a response to competitive moves to offer larger phones in the case of the Plus line.

The chart below shows what’s happened to the price range and average selling prices as a result (in each case, the year refers to the year in which products were launched, with the ASP being the average price per shipment during the four quarters from launch). In this case, the lowest price available does include older phones sold at discounted prices, and I’ve included a pretty modest guesstimate for this year’s ASP, which could well be quite a bit higher depending on supply levels for the iPhone X.

The price range from the very cheapest iPhone being sold to the most expensive model has obviously expanded significantly, from just $200 in 2010 to $800 once the iPhone X launches, with the iPhone SE dropping to $350. That reflects the broadening range of iPhones available, as well as the increasingly large storage tiers the various lines offer.

Remarkably strong ASPs throughout

But to my mind the most interesting thing to look at is what’s happened to ASPs, because the pattern in Apple’s other three big product lines – Mac, iPod, and iPad – has been that as the product has matured and Apple has spread the lineup down market, ASPs have fallen, sometimes very significantly. Mac ASPs in the late 1990s were well over $2000, but have dropped by around half in 20 years even ignoring inflation during that period. iPods sold for an average of over $400 when they first launched, but dropped to an ASP of near $150 in the ensuing ten year period, while iPad ASPs have dropped from over $600 to closer to $400.

The iPhone is therefore a massive outlier even among Apple’s own product lines, with an ASP that has held constant or risen for much of the last ten years. Apple has changed the way it reports iPhone revenues since its inception, which makes it impossible to make true comparisons between early ASPs and those today, but as the chart above shows, ASPs have risen over the last seven years at least even as the lowest available price has fallen by $200. In other words, though Apple has reduced the lowest price of iPhones by nearly half since its inception, people are choosing to spend more and not less over time, even as the market for iPhones has expanded dramatically beyond the US and other mature and high income markets. In other words, that expansion in ASP has happened in spite of many sales of cheaper, older phones in less wealthy markets around the world. The big question is therefore why, and what’s different about the iPhone?

Marketing, business models, and the centrality of the smartphone

I’d argue there are three key reasons for the behavior of iPhone ASPs. The first is marketing, the second is business and sales models for the iPhone, and the third is the increasing centrality of iPhones in our lives. Starting with marketing, I don’t think we can overlook Apple’s successful marketing of new iPhones each year as big improvements over the last year’s phones and especially over those from two years before, given that the two-year upgrade cycle was for many years the default. Apple has convinced people to buy a top of the line iPhone roughly every two to three years very consistently, and the reality is that the price of the top of the line iPhone has risen significantly during that period without much evidence of price elasticity in mature markets at least. That’s a testament to Apple’s product management, positioning, and promotion of its new devices.

Secondly, the business and sales model for selling smartphones in many markets is a huge factor here too. There’s really no other major consumer electronics category where the default sales model in many markets is to pay for it in monthly installments rather than upfront. There are certainly financing models in other markets, but nowhere is the installment or leasing model as ubiquitous as with smartphones. That model has helped the iPhone enormously, as one of the more expensive smartphones on the market, because consumers are not exposed as directly to the price increases they’ve faced over the years. Without the wireless carriers and their subsidy and installment models, I’m very skeptical the iPhone would have seen the iPhone performance it has.

Thirdly, the willingness of consumers to pay more and more for their iPhones is a testament to the increasing centrality of smartphones to our lives and the role these devices now play for us. In many cases, they’re partial or complete replacements for personal computers, used for managing calendars, writing emails, online shopping, paying bills, and so on. Many of us run businesses in part or entirely from our smartphones. And so these devices have taken on enormous significance for us and we’re willing to pay more to get bigger screens, faster processors, and other features which reflect their centrality in our lives.

My bet is that we’ll see significant demand for the iPhone X this year, and if Apple is able to meet that demand with significant supply later in 2017 and into 2018, ASPs could again rise quite significantly. That would be a testament to these same three factors: great marketing by Apple, the enabling business and sales models for smartphones, and a recognition by consumers that these are some of the most important devices in their lives.

Streaming Bundles are Coming

This week saw two announcements of streaming bundles in the form of T-Mobile giving away Netflix’s service to its family plan subscribers, and Spotify adding the Hulu service to its US college student subscriptions. As streaming content services grow, we’re going to see more attempts to bundle them together both to grow subscriber numbers and increase the stickiness of existing memberships. But some of these bundles will be more natural fits than others, as data from a recent survey shows.

Bundles Will Become More Common

Both of the streaming bundles announced this week are technically giveaways by one provider of a service provided by another. T-Mobile is using free Netflix as a value-add for its family plan subscriptions, while Spotify is using Hulu as a great add-on for its student subs in the US. We don’t know the financial details of these arrangements, but it’s likely that T-Mobile is paying a slightly discounted rate for the $10/month subscriptions it’s subsidizing for its subscribers, while Spotify and Hulu are likely doing some kind of revenue share on their partnership.

In both cases, the partners have come together in recognition of the fact that customers will respond both to discounts and to consolidating multiple services on a single bill. That’s a strategy that’s long been applied by telcos and cable companies combining broadband, TV, and home phone services, and more recently adding wireless services into the mix as well. Numbers from cable operator Comcast show that under a third of its customers take just one service, while exactly a third took two services (most commonly broadband and TV) at the end of June, and over a third take three or more services from the company. Long experience with these bundles among these companies show that they compete more effectively against single product offerings and that churn is lower for each additional service taken by a subscriber.

Those lessons are now being applied to the streaming content market, with the same drivers presumably motivating the partnerships we’re starting to see. And I’ll bet we’ll see more of these. Some companies are already aggregating these subscriptions in more indirect ways, notably Apple through the App Store and its subscription model, and Amazon with its Subscribe with Amazon service. But we’re going to see more services partner directly with each other going forward.

Some Services Will be Better Matches and Drive More Growth Than Others

A few weeks ago, I shared some high level findings from recent surveys I’ve done on subscriptions to US streaming services. Today, I thought I’d share a slightly different set of data from that survey, which dives a little deeper on the cross-penetration between services. The table below summarizes the penetration of each of a handful of services within the base of subscribers to the others. In other words, the 50% shown near the top left means that 50% of Netflix subscribers also subscribe to Amazon Prime, while the 12% that appears near the bottom right means that 12% of Apple Music subscribers also subscribe to Spotify.

You’ll notice a very wide range of numbers across the board here, all the way from 10% Apple Music penetration among Spotify subscribers to 79% penetration of Hulu subscribers by Netflix. Some of these services are clearly more natural bedfellows than others. Importantly, subscribing to multiple video services – e.g. Netflix, Hulu, and Amazon Prime – is common, with penetration rising among those who already take one video service relative to the overall population, while taking multiple music services is far less common. That reflects the fact that video services are largely complementary, while music services tend to be very similar in the content they offer and mostly differ on additional features.

Of course, one of the partnerships we saw announced this week was between Spotify and Hulu, and in the version of the table below I’ve highlighted the cells that deal with the overlap between these services:

As you can see, cross-penetration between these services is relatively low, with 24% of Spotify subscribers taking Hulu, and 18% of Hulu subscribers also taking Spotify. It might be tempting to draw the conclusion from this that the services are a poor fit, but the critical context here is that both services have fairly low penetration among the overall population compared to much more popular services like Netflix and Amazon Prime. As such, the rates of cross penetration are actually a little higher than their penetration among the general population. And the other way to look at this is that there’s clearly an opportunity to raise both that cross penetration and the overall penetration among the population through bundles like the one these two companies announced this week. In fact, given the relatively low rates of cross-penetration combined with the complementarity of the services, this is arguably by far the best match among the companies in the survey.

Broader Potential for Bundling and Discounts

For now, that bundle is only available to college subscribers, and not to other users of either service. That allows the companies to make good inroads with an important target group of subscribers without exposing themselves to massive losses across the broader customer base, and I think it’s very unlikely that we’ll ever see Hulu thrown in for free with a standard Spotify subscription. But I wouldn’t be surprised if we see Hulu and Spotify bundled for a discounted rate, such as $12-15 per month instead of the $18 they’d cost if purchased separately. That would be very much in line with the cable and telco style bundling I mentioned earlier, and the companies could justify the discounts on the basis of the lower churn and customer acquisition costs they’d likely see from these subscribers.

Meanwhile, we’ll also continue to see bundling across industry sectors, of which this week’s T-Mobile-Netflix announcement is an example. AT&T already offers free HBO with some of its wireless plans, and as it completes its acquisition of Time Warner in the coming weeks it’s likely to do more of this sort of thing. Sprint has invested in music service Tidal and offers free access to Tidal for six months to its subscribers. Music bundles in particular have been very popular among mobile operators outside the US. For many streaming services faced with the prospect of expanding into new content areas – e.g. out of music and into video – the option to instead partner with a player in that market and achieve some of the same benefits will be very attractive.

As a counterpoint to all of this, there are those companies that already own services in multiple streaming content categories – Amazon and Google are perhaps the best examples today, offering both music and video subscription services, but Apple might well join the fray soon with its massive investment in original video content and the video service that will likely eventually house that content. And as I’ve already mentioned, both Apple and Amazon allow their users to bundle in third party subscriptions as well, making them powerful alternatives to these ad hoc bundles between individual services. Whether the individual services are compelling and whether there’s real benefit from integrating them will determine which of these players is ultimately successful.

Why Talk of a $1000 iPhone is Overblown

There’s been a lot of talk about Apple releasing a $1,000 iPhone next week, and a lot of pushback from financial analysts in particular on the idea that people would actually buy such a thing. Carolina shared some data yesterday from a recent survey about people’s attitudes towards buying a higher-priced iPhone, which showed a mixed picture at best. But the reality is that talking about a phone in these terms is a bit old-fashioned at this point regardless of the actual price, especially in a US context.

Installment Plans and Leasing are Now the Norm

In the US, the vast majority of premium smartphones are sold through the major wireless carriers, with the largest four being AT&T, Sprint, T-Mobile, and Verizon Wireless. Each of those companies has been through a transition over the past few years away from the traditional subsidy model, under which customers paid a portion of the price of a phone up front, to a combination of installment and leasing models, where the cost of the phone is broken up into monthly payments. The exact percentage of sales off the old subsidy model varies by carrier, but all four are now above 75% with several over 80%. This is how people buy phones now in the US, and it’s not about the full retail price.

Further, the new model gives the carriers quite a bit of flexibility with regard to how they actually set the monthly payments, because they can be spread over a variety of different time periods and discounted (and, yes, subsidized) in various ways to attract customers, especially in the context of a hot new phone. So a leasing plan from Sprint might have a very different monthly price from an installment plan from Verizon even for the exact same phone. But all of that will be expressed in the form of monthly payments and not a full retail price.

The Samsung Note8 Launch Gives Us a Sneak Preview

The recent Samsung Note8 launch gives us something of a sneak peak at how a $900-1000 phone actually gets priced and sold in the US market. Here’s a sampling of prices from the major US carriers:

  • AT&T: full retail price: $949; monthly price: $31.67 for a 30-month installment plan with option to trade in after 2 years; $39.59 to trade in after one year
  • Sprint: full retail price: $960; monthly price: normally $40 but currently $20 as part of a special offer, for a lease with option to trade in after a year
  • T-Mobile: full retail price: $930; monthly price: $210 up front and $30 per month over a 24-month term, with an offer to buy one and get one free during preorders
  • Verizon Wireless: full retail price: $960; monthly price: $40 for two years.

Compare that to current monthly prices for the base model iPhone 7 Plus, a phone that costs roughly $200 less to buy outright, which run from $25 to $36, and you’ll see that the real difference in price between a $770 phone and a $1000 phone isn’t $230 for most customers but a monthly price difference of anything from zero to $15. On top of that, bear in mind that the new iPhones are likely to be the biggest carrier switching event the US market has seen since 2014, so we’re going to see a lot of discounts, offers, and other promotions which lower the effective price even further.

So really any survey that asks about a thousand dollar iPhone is asking the wrong question: the real question is whether customers are willing to pay a little extra (or perhaps none at all) for a great new phone. Which is why Phil Schiller showed this slide during the iPhone pricing segment of last year’s event:

This is the framing you can expect to see from Apple next week: affordable-looking monthly pricing, with the new phone probably coming in at around $40, or $8-10 more than the iPhone 7 Plus. And that’s going to be a lot more palatable than the “$1000 iPhone” headlines will lead people to believe.

One Last Note on iPhone Pricing Surveys

One last thing that’s also worth mentioning is the very nature of surveys that ask about theoretical prices for as-yet unreleased devices: they’re historically terrible as predictors of what people will actually pay in the real world for devices they have seen. Ask someone whether they’ll pay $1000 for a new iPhone they in many cases know nothing about (and might reasonably assume will be an on-par successor to the current iPhone lineup) is always going to lead to unrealistic responses. But ask them next week whether they’ll pay a few bucks extra per month for the iPhone Apple just announced with all its usual fanfare, whiz-bang demos, and so on and you’ll get a very different answer that’s likely a lot closer to what we’ll actually see in stores. Now it’s just up to Apple to sell us all on that device on stage next Tuesday.

The Virtual Reality Price Squeeze

High-end, mainstream virtual reality is coming up on eighteen months in the consumer market at this point, with the launches of the Oculus Rift and HTC Vive early last year and the PlayStation VR in late 2016. Around the time of the first launches, I wrote a piece here about the emerging spread of VR products that existed at that time in two fairly distinct camps separated by quite different pricing, with the higher-end rigs mostly $400-800 on top of either a high-end PC or a console, and mobile VR down at the $100 price point or below. Microsoft and others clearly saw an opportunity in that market to create lower-cost but still powerful VR rigs in-between those two price buckets around that time. But the market has now moved on quite a bit, with recent price reductions squeezing the middle and leaving little room for differentiation where there was once a clear market opportunity.

At the time Microsoft launched its intentions around what it called Windows Mixed Reality headsets late last year, the VR market looked about like this:

There were two distinct markets at the high end and the low end, and an opportunity at around something like $300 to create a powerful and yet more affordable alternative to the premium PC-based rigs and to some extent Sony’s console-based setup. Microsoft consequently announced that it would have headsets in that price range available through partners based on its Windows Mixed Reality platform sometime this year, and we’re now starting to see those devices announced and available to the public, including at IFA this week.

And yet a lot has changed in the months since Microsoft first outlined its vision, notably the prices of those high-end rigs. At the time those mixed reality devices are actually launching, the market looks more like this:

Whereas Oculus Rift’s bundle and the HTC Vive both started at $800 early last year, in recent months those prices have fallen quite a bit, with the Oculus bundle currently selling at $400 and permanently reduced to $500, and the HTC Vive getting a recent price cut to $600. Sony, meanwhile, has dropped the effective price of its core VR system by $50 as part of various bundles. And those Microsoft mixed reality headsets that are starting to debut are mostly coming in higher than that eye-catching $300 price point, with actual prices mostly between $300 and $500. I’ve painted a pink band across the chart above to show how that price bracket compares to the rigs Microsoft originally intended to undercut, and you can see that there’s actually overlap with the current prices of both the Oculus Rift bundle and the PlayStation VR, while the HTC Vive is just $100 more.

Of course, Microsoft’s other big price argument is that its rigs won’t require such high-end PCs, offering at least a decent experience on more generic PCs and therefore providing a lower total cost of ownership than HTC and Oculus’s solutions. And I don’t mean to pick on Microsoft – others have pitched standalone VR headsets which are intended to sit between the mobile VR and high-end rig prices, but they’re also aiming at that same $300-500 price bracket that’s now being squeezed from the top.

All of this is great news for the consumer and for the mainstreaming of VR, at least in the short term: the cost of owning a VR system has never been lower, and the range of content (especially games) available for these platforms is growing all the time. All this will expand the addressable market for VR beyond the 4-5 million that have so far bought one of the mobile or premium systems, probably multiplying it several-fold over the next few years.

And yet I have to wonder how sustainable these price cuts will be for the companies: Oculus has always said it sells its hardware at a loss, while HTC was reportedly making a profit on $800 systems but likely isn’t at $600. Oculus’s price discounts may well presage new hardware coming soon which could see the price point of newer systems creep back up, but HTC may have a hard time rolling back its price discounts without new hardware. It can ill afford to sell VR systems at a loss given the company’s overall losses and the state of its smartphone business, hence recent reports that it’s exploring strategic options including a sale or spin-off of the Vive business, which after all has little connection to its phone effort. Facebook can certainly afford to fund ongoing losses for some time to come, but its platform has been the least adopted of the major systems so far, though the price discounts may help change that.

Meanwhile, the opportunity in the middle of the market is going to be increasingly challenging because the price differential versus the premium experiences is significantly less than it once was, making life harder for the Daydream, HTC, and Facebook standalone headsets due to hit the market soon as well as those Windows headsets. And of course at the same time the other part of the mixed reality spectrum – augmented reality – is going gangbusters with both major smartphone platforms creating tools for developers to leverage the largest installed base of consumer devices out there. It’s going to be an interesting period for the VR market.

Why Tech Companies Keep Buying Sports Rights No-One Cares About

This past week saw both Twitter and Facebook sign deals with recently-launched sports network Stadium for live coverage of various college sports. Both companies had earlier signed TV streaming deals for a number of other bits and pieces of sports content, none of them particularly compelling. In a world where sports content is one of the few slices of live TV still holding up reasonably well as viewing shifts to on-demand and streaming, why aren’t these companies buying more interesting stuff? The answer lies largely in the long-term deals signed by the major sports leagues in the US.

Recent Sports Rights Deals for Tech Companies Are Mostly Sub-Par

Twitter and Facebook’s Stadium deals are far from the only ones they or other tech companies have signed over the last couple of years. This year in particular has seen a big increase in investment by these companies as they look to fill their rosters of live video content and Twitter in particular tries to deliver on its commitment to have 24/7 live streaming video on its site. Some other examples include:

  • Facebook signed a deal with Major League Soccer and Spanish-language broadcaster Univision, another deal with the NBA’s D-League (the minor league of the NBA), showed pre-Olympic basketball exhibition games in summer 2016, will show 20 live MLB baseball games in the 2016/17 season on Friday nights, and secured rights for a number of second-tier European soccer tournament games for the 2017-2018 season, as well as over 5000 hours of e-sports content.
  • Twitter won the 2016-2017 deal to stream Thursday night NFL games online (a deal lost to Amazon this year), but other than that has mostly had content of less interest to most viewers, with few exceptions: the Wimbledon tennis tournament was another highlight, but it has also signed deals with the WNBA, the PGA Tour, and various other sports including lacrosse.

To be clear, there is an audience for each of these sporting events that these companies are carrying, but the vast majority of it sits outside of the sports that actually drive live viewing in meaningful numbers. Facebook’s live video metrics have all been vague and relative, so we have no way of measuring its success in absolute terms, but Twitter has provided the number of unique live video viewers each quarter, as shown in the chart below:

As you can see, the number of live video viewers has grown over time, but in Q2 it was just 17% of its total monthly active users, meaning that over 80% of its monthly active users never even watched any live video at all. Of that 17%, it’s entirely possible that many simply watched a few seconds, so the number that actively engaged and watched any meaningful amount of live video is likely far smaller still.

The Best Rights are All Locked Up for Years

Why, then, do Facebook and Twitter bother with these sub-par sports rights that drive little viewing? The simple answer is that the rights that might actually drive meaningful engagement are almost all locked up for years. The table below shows a summary view of the US TV rights for the four major sports leagues in the US, some of which are sliced and diced in many different ways, with the rest allocated more simply:

As you can see, Major League Baseball, the National Hockey League, and the NBA are all locked up until at least 2021, with most of the NFL rights packages also locked up for almost as long. The NBA won’t be available to new bidders until the 2025 season. This is why the tech companies – notably Facebook and Twitter – have been acquiring so many other rights: because they’re the only ones available. No matter how much these companies are willing to spend, they simply can’t get significant access to the major sports people actually watch in the US.

The one exception to all this has been the NFL’s Thursday Night Football package, which has had separate broadcast and digital rights deals for the last couple of years. Twitter won that deal for the 2016/17 season, but lost it this year to Amazon, which is likely to be another big bidder for sports rights in the next few years. Verizon, meanwhile, has the unique mobile rights to NFL games, which detracts from every other digital football package out there, but that deal will expire in 2018, so it will be interesting to see what happens then.

Two More Years of Dealing in Marginal Sports Content

All of this means we likely have two more years of tech companies mostly dealing in the same marginal sports content we’ve seen so far from them, grabbing a few games here and there from the major leagues, and then securing broader rights to sports of less interest to the mainstream US user. But a couple of years from now, as some of the rights negotiations for big deals starting with the 2021 season begin, I would expect a number of big tech companies, including not just Facebook and Twitter but also Amazon, Google, and Apple, to be major bidders and likely secure some big packages which have hitherto all been captured by broadcasters and cable operators.

In the meantime, the main focus for most of these companies will have to be on video content other than sports, meaning commissioning both scripted and unscripted shows or acquiring those being shopped around, and the investment in original content in particular will continue to grow, with Apple apparently spending a billion dollars this year, Netflix spending $7 billion, and others like HBO, Hulu, Amazon, and others spending somewhere in-between those figures. For now, that’s the only thing these companies can do, and it’s going to mean that the prices of content go up and there’s eventually a glut of video on the market, until such a point as the legacy TV industry starts reducing its spend in the face of accelerating cord cutting and cord shaving.

The Big Software Company Hardware Push

Recent weeks have been a great reminder that some big companies that once focused almost entirely on software and online services have become incredibly interested in and serious about making hardware. Microsoft, Google, and Facebook have each made major inroads into this space in the last few years, with mixed success, and show signs of getting deeper into the hardware market in the next few months. What has motivated this push into hardware, and where is it likely to lead?

Five Years of Significant Change

Go back just over five years to the beginning of 2012, and three big names in the consumer tech industry – Microsoft, Google, and Facebook – were largely absent from the hardware market. Microsoft made Xbox consoles and PC accessories, but allowed PC OEMs and smartphone vendors to make the hardware that ran its operating systems. Google worked with partners to create its Nexus phones to show off the best of Android but in reality the market was carried by OEMs there too. And Facebook was a social network with no aspirations in hardware at all. Fast forward to today and these companies each now have a significant presence in hardware. How did we get here?

The history is different for each of these companies:

  • Microsoft:
    • launched its Surface line of personal computers in October 2012 (the brand had previously been used for tabletop computers), and has launched several more Surface Pros as well as the Surface Book, Laptop, and Studio
    • announced its acquisition of Nokia’s devices business in 2013
    • unveiled HoloLens in early 2015
  • Google:
    • acquired Motorola Mobility in 2012 and owned it for a couple of years before selling much of it on to Lenovo in 2014
    • launched the Chromebook Pixel laptop in 2013 and updated it in 2015
    • launched the Pixel C tablet in 2015
    • partnered with TP-Link and Asus to make OnHub routers in 2015
    • unveiled the Pixel phones, Google Home voice speaker, and Google WiFi routers in 2016
  • Facebook:
    • partnered with HTC to create the HTC First, the “Facebook Phone”, in 2013
    • acquired Oculus in 2014
    • created the Building 8 group and hired Regina Dugan to run the team focused on new hardware projects in 2016.

The pattern, however, is the same: each company has steadily invested more and more in its own hardware over time, and I’ve only covered the history here. Looking forward, Google is expected to launch more Pixel phones, a possible new Chromebook Pixel, a smaller Google Home device, and possibly Google Assistant-powered earbuds later this year, and Facebook has recently been reported to be working on voice speakers, video conferencing hardware, and more. Both Facebook and Google have also consolidated their previously disparate hardware efforts under unified leadership.

Control and New Sources of Revenue the Drivers

What, then, are the drivers of this increased investment in hardware? The simple answer is twofold: these companies want more control over areas they’ve previously had to cede to others, who have often had their own priorities, and there are new revenue streams in providing hardware, which become particularly important as consumers stop paying for traditional software.

Steve Ballmer arguably kicked off the hardware investment era at Microsoft with his Devices and Services strategy, which was a recognition that those two categories and not software sales would be the sources of future revenue for the company, and led to its acquisition of the Nokia devices business and launch of Surface. As other companies – notably Apple and Google – created the expectation that operating systems and productivity software would be free, and funded their own efforts in these areas through device purchases and ad revenue, it became clear that Microsoft’s model would have to shift too.

But both the Surface and Pixel lines can also be seen as attempts to exert greater control over the end user experience than working through OEMs allowed either company. That was born out of frustration over both the lack of competitiveness in the hardware OEMs were creating, largely versus Apple, and also the way the companies’ software features and services were often buried behind layers of user interface and crapware provided by OEMs and their partners. Facebook’s hardware efforts also stem from its frustration at missing out on the last round of user interfaces and platforms, primarily mobile, and again dominated by Apple on the one hand and Google and its OEMs on the other. Investing early in VR was an attempt to stake a claim in what might become the next user interface and ensure a leading role, while the efforts rumored more recently are clearly a response to Amazon’s success in controlling another new user interface category.

These companies have also seen the benefits that accrue to those companies which are able to combine their own hardware and software in a tightly integrated fashion, designing and commissioning their own chips and other components to create highly optimized experiences, with Apple again the prime example of this once-derided strategy. Apple’s AR push with ARKit is just the latest example of how this strategy has allowed it to pull ahead of others by controlling the whole widget from start to finish.

Mixed Success at Best, So Far

For all that, the early efforts from each of these companies have met with mixed success at best so far. Facebook’s phone push with HTC never went anywhere, leaving it entirely without a stake in mobile platforms, while Oculus hasn’t been a huge seller in the VR market and Oculus’s business is still formally immaterial to Facebook’s overall finances.

Google’s hardware has a mixed record too, with the Motorola acquisition a failure, the Pixel laptop and tablets also-rans in their respective markets, and the Pixel phone a modest seller in its first year thanks to limited supply and distribution channels. The Google Home seems to be selling decently well as the only real competition to the Amazon Echo, but it’s far from clear what its long-term business plan is here – advertising seems the obvious revenue model, but is fraught with risk and likely to encounter significant resistance from paying customers.

Microsoft’s Surface is arguably the big success story here, though even then it’s a tiny player in the overall PC market, it took several years of work to figure out the right model for the Surface Pro hardware and software, and it has of course recently seen blowback from last year’s reliability issues. The Nokia acquisition, meanwhile, may have become an inevitability by the time it happened, but certainly hasn’t panned out well either. And HoloLens has an interesting role in the enterprise and education markets, but is far from a mainstream AR or VR product.

Where Do We Go From Here?

It’s clear that we’re going to see more of this stuff from these three companies in the next few months and years, with a launch event from Google this fall a certainty and some kind of new hardware products from Facebook early next year also looking increasingly likely. And yet it’s far from certain how these devices will perform given this patchy history. The categories Facebook is getting into seem like a poor fit either for its expertise or for customers’ perceptions of what Facebook is as a company – I’d argue the push into hardware for ad-based companies is particularly tough, because the products customers know them best for are all free. Thus, they either have to subsidize their hardware efforts with ad revenue or substantially change customers’ perceptions of their brands. The lack of experience with hardware itself can also create problems. To go even further, I’d argue that if we’ve learned anything over the last couple of years about Facebook, it’s that it’s pretty great at squeezing more revenue out of already successful social products, but pretty terrible at creating brand new ones, let alone anything in hardware.

All that said, it’s possible that we’ll see some really good stuff out of these companies too. Microsoft has demonstrated an innovative streak in hardware that’s been refreshing in a market dominated by a lot of me-too products, while the combination of hardware and software in Google’s Pixel was actually pretty compelling too. All of these are companies packed with smart people and impressive capabilities around software, AI and machine learning, and more skills which can be applied to their hardware efforts. In the process of pursuing these hardware ambitions, these companies have also pushed competitors and partners to do better, and that’s particularly clear in the PC business, where the Surface has arguably shaken some life into the OEMs.

Overall, there’s still quite a bit of uncertainty over the exact direction the hardware pushes at these erstwhile software companies will take over the coming years, and how well it will work out for them. But it’s abundantly clear that each is now very serious about hardware and going to spend a lot of money in the coming years trying to figure that out.

Parsing US Online Subscription Trends

I’ve recently done a number of surveys into online subscription trends in the US market, covering services such as Netflix, Amazon Prime, Hulu, Apple Music, and Spotify. Though the penetration and market share numbers that come out of these surveys for individual services are interesting, today I want to focus on some of the second-order trends that emerge from the numbers.

Many Still Don’t Subscribe to These Services

My questions were asked pretty broadly, so as to capture household subscriptions rather than merely individual subscriptions, since these services are often shared by families or roommates. But one of the striking things that comes out of the surveys is that those who subscribe to none of the services I listed above is still pretty high – 43% of those surveyed said no-one in the household pays for these services. That number is definitely influenced by age – respondents in the oldest age brackets were almost twice as likely to give this response as those in the youngest – but even among younger respondents, as many as a quarter replied that no-one in their household paid for these services. It’s possible that in some cases these individuals are using those services by taking advantage of an account belonging to someone outside the household, but it’s still surprising how high some of these numbers are. My previous research found that income levels also play a role when it comes to services like Amazon Prime in particular.

As broadband and smartphone penetration continue to increase, the number buying none of these services will continue to come down, but these figures certainly suggest that there is headroom for all the services to grow, by tapping into those households which have yet to see value in any of them. That may well mean offering cheaper and more flexible packages, as Amazon has done in the past couple of years by offering both a monthly option and a discounted service for low-income households.

Those Who Subscribe Often Take Several Services

On the other hand, among those who do subscribe to at least one services, many subscribe to more than one, with significant overlap between some of the major services. Among the group that took at least one service, 43% took more than one service, with 19% taking three or more. Notably, though Amazon and Netflix are often thought of as competitors given that their video services are similar in nature, it’s certainty not an either/or proposition for many households. Just over half of Netflix subscribers also subscribe to Amazon Prime, while 59% of Amazon Prime subscribers also subscribe to Netflix, so there’s significant overlap between the two bases. Given that only around a third of Prime subscribers use the video feature, that makes sense.

Strikingly, Hulu subscribers were among the least likely to take just one service – just 18% of Hulu subscribers only pay for Hulu, while 82% also pay for Amazon Prime, Netflix, or one of the music services listed. And the vast majority of those take Netflix in addition, suggesting that although these too are nominally competing services, they’re complementary for many households. However, this pattern is less true with music services, where there is far less overlap between Spotify and Apple Music subscribers. The difference in overlap is likely driven by the fact that whereas video services offer different content, the music on offer is essentially the same on all music services, with differentiation mostly at the feature level.

All of this also fits with the increasing trend towards subscription aggregation, with Apple and Amazon the leaders in the market thus far. Through the App Store, Apple offers subscriptions to a wide range of services including most of those listed, while Amazon has offered several of its own subscription services and an increasing array of third party subscriptions through its Prime channel. That’s likely to be a lucrative strategy in the years to come as the average number of subscriptions per household continues to rise.

Age and Gender Play a Role for Some Services

The age and gender of respondents also played a role in at least some of the responses. I already mentioned that older respondents were less likely to buy any of these subscriptions, and that effect was particularly marked over the age of 55. But age also played a role in the propensity to subscribe to individual services, notably the music services, where the percentage of those paying for either Apple Music or Spotify was much higher among the 18-24 and 25-34-year old respondents than among those over 35. This difference by age was far higher than for any of the other services.

Meanwhile, gender played a limited role for almost all the subscriptions, with men and women responding very similarly on all the services except one: Hulu. For Hulu, women were more than 60% more likely to subscribe than men, suggesting there’s something of a gender skew in Hulu’s subscriber base, something I haven’t seen reported previously. That also suggests there may be some upside for Hulu in making its service more appealing to male customers.

As I’ve said a number of times recently, I believe that the subscription business model is only going to grow in the coming years – it’s how many of us increasingly like to buy our content, because it offers predictability and a large amount of content for a single monthly price. Big content distributors which have yet to embrace this model will increasingly be left behind, while those that jump on board are likely to see significant growth in subscribership and revenue as a result.