Apple: destroyer of fragmentation

With the release of Android Wear, we now have a pretty good idea of how Google intends to target the smart-watch space, and it’s very much about extending a subset of smartphone functionality to peripheral devices. Although today these new devices will offer a very thin layer on top of the smartphone, it’s easy to see in time they could subsume more and more of the functions of the smartphone and eventually act independently with their own cellular connectivity. As such, Android Wear can be seen as positioning Google for a future beyond the smartphone.

We don’t know yet what Apple has planned for wearables, but from the rumors around Healthbook, it certainly looks likely it will be taking almost the opposite approach – namely, adding value to the smartphone experience through peripheral devices. You can think of Google’s approach with Android Wear as being “smartphone-out” – i.e. extending smartphone functionality out to wearables, and Apple’s rumored approach as “wearables-in” – i.e. using wearables to add functionality to the smartphone. This makes good strategic sense for each company – Google’s interests are best served by extending its services to all possible categories of devices, while Apple’s are best served by making its devices in key categories as compelling as possible.

Wearables - Apple vs GoogleOpportunities in aggregation – in wearables…

Interestingly, there are no credible rumors so far about a wearable device from Apple, though there is no shortage of interesting thinking about the topic from various people in the industry. But I continue to think it’s possible that Apple won’t release its own wearable device at all – at least not yet. I think its major play here may be acting as the glue to bring what is presently a very disparate and fragmented set of wearable ecosystems together. Note how many fitness and health devices Apple already sells in its online store. Many of these sync in some fashion with companion iPhone apps, but they all do it separately. Some of the vendors have their own ecosystems, which offer limited integration between their own devices and third party devices for analytics purposes, but they’re largely islands today.

Bringing the data supplied by all these sensors into a single app which would make sense of it all would be hugely more powerful and Apple is one of the few companies that could do it. Its ability to do so would, however, be somewhat hampered if it decided to enter the market itself – it would then be competing with would-be partners, and its position as a hub would be significantly less attractive to them. Meanwhile, the wearables market is made up of so many small niches that it’s almost impossible to imagine how Apple could drive significant revenue (in the context of its existing $174 billion a year business) without launching at least a handful of different products.

… and beyond

But this approach needn’t be restricted to the wearables space. There are at least a couple of other areas where Apple could offer the same approach of unifying a fragmented ecosystem and the most obvious one is the home. At CES this past year, the most obvious trends aside from 4K TVs were wearables and smart home. But the smart home space is at least as fragmented as the wearables space, with players from the home security, broadband router, appliances, mobile services and other industries all vying for a slice of the action and ahough there are some attempts by both vendors and third parties to bring elements together, no one has yet cracked the unified home experience. But Apple is in an excellent position to do so, with both routers (the Airport line) and controllers (iPhones and iPads) already in place, and a huge user base to start with. Apple’s position as an aggregator will be much stronger if it isn’t itself trying to build smart home products, which also seems unlikely given the broad variety of devices it would have to make.

The other obvious area for Apple to apply the theme of bringing order to chaos is payments, where there’s at least as much fragmentation at present. We currently have at least three separate domains today when it comes to payments – classic credit cards, online payments providers such as Paypal, and a plethora of players in the mobile payments sector, from mobile terminals such as Square and Paypal Here to Google Wallet to the US carriers’ Isis initiative. None of these has so far thrived, and fragmentation is again a big reason. Apple’s huge user base and the installed base of Bluetooth LE-capable devices it has in the market are an enormously strong starting point. But the play here would be a different one: creating its own payment system rather than aggregating third-party efforts. System-level integration would be a huge advantage over any third-party app, as iMessage and FaceTime have demonstrated. And Apple’s unique leverage over carriers would allow it to do things Google wasn’t able to with Google Wallet. Tim Cook has already indicated that with hundreds of millions of credit cards on file and the Touch ID system introduced in the iPhone 5S, Apple already has a great starting point for payments.

The big question then becomes, how does Apple make money from any of this? If it doesn’t enter the wearable device space, how does it get revenues growing again? The answer is twofold: firstly, and in some ways most importantly, Apple will strengthen its ecosystem and the appeal of the iPhone, driving more iPhone sales. Conversion, not capturing first-time smartphone users, will be the key driver of growth in the premium end of the smartphone market going forward and Apple has to do all it can to cement the position of the iPhone in that segment. Secondly, these aggregation efforts can each yield revenues in their own right, through licensing (the Made for iPhone program is an existing example of this sort of thing) and in the case of payments through taking a cut of revenues, which could be enormously lucrative in its own right. Neither of these requires Apple to create a new hardware category of its own, and I’m increasingly convinced Apple’s new product categories in 2014 may not be hardware at all.

Content Selection as a Competitive Advantage

Many books have now been written about how globalized our world is, so it’s easy to sometimes forget how local aspects of our digital culture still are. Nowhere is this more obvious than when looking at digital content stores, the disparity in the availability of different content around the world, and in the differences between demand for content in different countries.

Figures from the International Federation of the Phonographic Industry (IFPI) released on Tuesday include a useful reminder of how different tastes in content can be from country to country. One of the tables in its report on digital music showed the percentage of the top ten albums in certain non-English-speaking countries by domestic artists:

IFPI domestic repertoire

In these 13 countries, the range is from 50-100%, with a significant majority in each being local content. Also in the past few weeks, the BBC released its iPlayer Stats for January 2014, including the top 20 TV episodes watched on the service – all are British television programs, making the domestic rate 100% (source: BBC iStats):

iPlayer top 20 episodes

What does this mean for players in the consumer technology industry? It means in order to be relevant in providing content services in various countries, you need to not just have stores in each of those countries, but you need to secure local content. The chart below shows the number of countries in which four major players offer various content services (music downloads, movie downloads, TV show downloads, ebook purchases, app downloads and music subscriptions):

Content availability for big 4

There’s a huge variety here, with Apple having by far the largest number of countries for music downloads and movie downloads, Amazon (predictably) offering the largest number of countries for ebook purchases, and Microsoft leveraging its long-standing developer infrastructure to provide app stores in the most countries. Google comes out on top in Music subscriptions, where its All-Access Pass is now available in 25 countries, but otherwise lags its competitors.

Building up a set of international stores for content takes significant time, and it’s easy (especially for those of us in the US) to forget how few countries some of these providers offer content in, especially Amazon. That’s partly a question of acquiring the appropriate rights for distributing more global content such as apps and movies, but it’s also critically about securing the rights to relevant local content. Then there’s the infrastructure required to make all this work, preferably either in-country or somewhere close by. Building all that takes time, commitment and significant financial investment. It’s an easily overlooked part of Apple’s competitive advantage today (not to mention providing almost $10 billion in annual revenue).

In addition to geographic availability, all of the iTunes stores combined have the largest library of content, with 650,000 movies, 250,000 TV shows and 37,000 songs as of the last data released, and in most cases multiples of the catalogs available from other providers. As far as the iPhone and the iPad are content consumption devices, Apple has built a significant edge in allowing consumers to acquire content easily and legally and making it simple to get it onto devices.

Apple’s competitors also need to learn from its example. In other words, to the extent Amazon is serious about digital content, it has to not just launch new streaming services and new streaming devices, but extend its content catalogs in terms of both size and geographic reach to be competitive. The same is true for Microsoft as it pursues its Devices and Services strategy – tablets in particular are still content-centric devices, and it badly needs a better story around its content stores to make the Surface more appealing. And Google, too, appears to be investing in improving its content catalogs, announcing new countries for some services just this week. But all three still have a long way to go.

How Carriers’ Move Away From Phone Subsidies Could Hurt Them

Over the last several months, we’ve seen moves by the US carriers to introduce a new model for paying for devices in addition to the traditional subsidy model. The new model allows customers to pay for their devices over a period of time in monthly installments. It began at T-Mobile in May 2013, and the other carriers quickly followed over the course of the ensuing months. There are still significant differences in the details of these plans: T-Mobile has done away with service contracts (though you’ll still sign one for the installment plan), and T-Mobile, Sprint and AT&T offer service discounts for customers who are paying for their own device either outright or through installments, whereas Verizon Wireless stubbornly charges the same service fees regardless of whether it is subsidizing the device.

What the T-Mobile, Sprint and AT&T plans have in common is they separate two things that have historically been intertwined: service fees and device payments. Carriers traditionally charged about a third of the retail price of high-cost devices up front, recouping the rest through service fees over the course of a two-year contract. Many customers didn’t really understand the true cost of the device as a result, and the model also meant customers who held onto a device beyond the standard upgrade cycle were paying the carrier far more than the cost of the device. As such, the switch to installment plans is a good thing for consumers, because it introduces transparency over the relative costs of service and hardware.

It’s also good for the carriers, because they can slowly reduce the costs of subsidizing handsets while being more competitive on service pricing. Sprint is the only carrier that’s really broken out its costs of subsidy explicitly, and the impact in the first few months was significant, as shown in the chart below, which illustrates the cost of subsidies (i.e. the cost of equipment not paid for explicitly by customers) and the percentage of the total cost of equipment sold that’s covered by equipment payments from customers.

Sprint subsidy costs

As you can see, the total subsidy cost, which normally spikes hugely in Q4, when carriers sell many more handsets than in the other three quarters, didn’t spike nearly as much in Q4 2013, partly because Sprint sold fewer smartphones than it usually does in Q4. But the key thing to note is the line, which shows 41% of Sprint’s cost of equipment was paid for by customers, significantly up from 34% a year earlier, and also significantly higher than in any other recent quarter. The strategy worked, reducing subsidy costs and passing more of the cost directly on to consumers. It’s likely other carriers, especially those actively pushing customers towards the installment model, are seeing similar or even greater positive results. So this shift is good for carriers too.

What are the other implications of this move? Well, one that I’ve already mentioned is consumers will become much more aware of the true cost of a device (i.e. $649 for an iPhone 5S 16GB, and not just the $199 most carriers charge up front on a two-year contract). But to me the bigger implication is consumers will start to wonder why they should pay the carrier for a device at all. If the service fees and hardware fees are separate, even though the service fees are logically paid to the service provider, why shouldn’t the hardware fees go to the hardware vendor?

The answer today, of course, is carriers are offering zero percent financing on these devices – there’s no additional cost to paying for a device in installments over a period of 12-24 months instead of up front, and that’s a lot easier to swallow for most consumers. But what if hardware vendors started offering interest-free financing too?

The challenge to the direct sales model for hardware has always been multi-faceted. For example, hardware vendors lacked direct distribution in the form of retail stores where consumers could try out devices. But the biggest challenge was consumers in most countries simply aren’t accustomed to paying up front for devices, and even if they did, the service fees they paid to carriers would still implicitly include subsidy repayments, making it a really unattractive proposition. But now that three of the big four carriers are pushing reduced service fees in return for consumers paying for their own devices, that equation changes. And if hardware vendors started offering installment plans instead of forcing consumers to pay up front, that would be pretty attractive. If the hardware vendors threw in free annual upgrades too in return for giving back the old device (as some carriers do), that might make it even more attractive.

Consider this: Samsung offers you the option of always owning the latest member of the Galaxy S family for a flat monthly fee of $30. For consumers, this would offer far greater flexibility in their choice of carriers. Instead of being forced to stick with a carrier until their device was paid off, they could switch whenever they wanted to, using either contract-free postpaid plans or even prepaid plans. Their loyalty would be to the device vendor and not the carrier (though they might choose to stick with a carrier that worked for them). Device vendors would enjoy the benefits of eliminating the carrier middle-man, which would give them the option of reducing the price of devices on these installment plans, and develop direct relationships with their customers.

What are the downsides here? Well, for the carriers, this would be a step in the wrong direction: they’d lose the direct relationship with a consumer around their device, and potentially become much more expendable. At present, the device and service contract cycles, often out of sync especially within a family plan, create a perpetual lock-in which helps keep churn low. Remove the device upgrade cycle from the equation and suddenly it becomes much easier to switch when the contract is up (or at any time, if there is no contract).

What’s the downside for vendors? Well, the obvious answer is that, instead of getting a big payment up front they’d get the payments spread out over a period of time. That’s much less attractive from a revenue recognition perspective. And it also creates a massive potential bad debt problem, in that customers might fail to make their payments. This risk is one of the reasons for Verizon’s caution on installment plans, and it would be a big risk for any device vendor adopting this approach, without the benefits of a service contract to hold over the customer as leverage.

Which vendors would be most likely to take this approach? Well, since many device vendors struggle to make money as things stand, most of them are not likely to pursue this strategy. Those best placed are those which already dominate the market, namely Apple and Samsung. Both make very healthy profits from smartphones and have deep pockets to fund such an initiative. Both also have something of a retail presence, Apple a significant one with its retail stores, and Samsung a small but growing one with its stores-within-a-store at Best Buy and some standalone retail outlets. Two other players who don’t currently make smartphones but could afford to do something interesting with this approach are Google and Microsoft. Microsoft, of course, is acquiring Nokia, and so will shortly be in the handset business, and given its struggles to get carriers to support Windows Phone it might find the direct route appealing. Google already sells phones direct through the Google Play online store, and this might finally offer a way to get mass-market interest in Nexus and Play Edition phones, something it hoped (but failed) to stimulate when the first Nexus phone launched.

Beyond these big players, the Chinese vendors might also find the approach attractive. They have struggled to get tier 1 carriers in major western market to carry their devices under their own brands, but might use the installment approach as an alternative route to those markets, improving their name recognition and perhaps helping to bring the carriers around too.

There are significant barriers to this approach, but thanks to the carriers’ moves to get out from under the burden of subsidies, several of the barriers that existed in the past are slowly being removed. I would expect at least one phone maker to begin experimenting with this model in the coming months, and I wouldn’t be surprised if others followed.

Taking Apple TV beyond a hobby

Apple’s TV device made the headlines again this past week when Tim Cook mentioned at the annual shareholder event that it was a billion-dollar business for the company in the last fiscal year. Apple subsequently made clear that this included content as well as hardware sales connected with Apple TV. Though Tim Cook made clear that such numbers make it a little harder to describe it as a mere hobby, it’s still a tiny fraction of Apple’s overall revenues. It’s likely that the Apple TV in its current iteration (the small black puck) has sold around 20 million units in total at this point, at $99 a pop. To put that in context, consider that Apple hasn’t sold fewer than 20 million iPhones in a single quarter since 2011, and that it routinely sells that many iPads each quarter. The Apple TV generated under 1% of Apple’s revenues in the last fiscal year. At this rate, Apple is still selling two to three times as many Macs each quarter as it is Apple TVs.

Why won’t the Apple TV take off in the same way as Apple’s two biggest products? The biggest single reason is that, although it serves three useful functions, it doesn’t solve a fundamental issue in the category in the same way as the iPhone and iPad do. What is the fundamental issue in the TV category? Consumers want single-input, single-interface access to all their video content, and instead are presented with having to switch devices, inputs and apps to get from one show to the next. The things they want to watch are spread across traditional television services from cable, satellite and telco (CST) providers, online subscriptions like Netflix and Amazon Instant Video, free, ad-funded services like Hulu and networks’ own apps and websites, and pay-per-use services like iTunes, Amazon Instant Video and Vudu. There are apps dedicated to helping you figure out which of your many services might carry the particular show you want to watch – that’s how bad things are.

Apple TV does now carry lots of apps which allow you to watch a variety of content without switching inputs on your TV, but that’s only a partial solution. You’re still having to hop between apps, with no universal search to allow you to locate shows across them (unlike Roku). But there are other hassles too – people still have to maintain multiple subscriptions and authenticate themselves to the various apps they want to use. Apple has simplified this somewhat by allowing customers to subscribe to Netflix and other services through their Apple ID. But the authentication process you have to go through to watch HBO Go, the Disney channels and other TV Everywhere content is a significant hassle. Entering usernames and passwords on any TV device is a pain in the neck, and having to do it repeatedly is even worse. The ideal scenario in the current environment would be connecting your Apple ID to your CST provider ID and then having it automatically authenticate you across all your apps – something it sounds like Apple might have been trying to accomplish as part of its talks with Time Warner Cable.

Though adding universal search and one-time authentication would represent progress, it would be progress of a very incremental kind, leaving consumers with many of the existing hassles they deal with today, and not solving the fundamental fragmentation in the TV space. It would leave the Apple TV without a comprehensive live TV offering (beyond the specialized sports channels it offers today). That was likely the major focus of the Time Warner Cable talks, which are obviously now unlikely to come to fruition. But that, too, would have represented a short-cut to what has to be the end game for Apple in this space: offering a full-service TV subscription offering that would combine traditional live linear broadcasting with on-demand viewing of both recent and back-catalog TV shows and movies. That kind of offering would be truly disruptive in the way the iPhone and iPad were before, and would finally make the Apple TV a compelling offering, truly differentiated in the market and offering something really unique: a single-input, single-interface, single-account TV service.

Besides goosing Apple TV sales, this would help Apple in a couple of other ways too. One of the reasons why investors are so easily spooked when it comes to Apple, and why its valuation always seems way out of whack with its performance, is that its business model relies on huge, roughly yearly hits in its two major product categories. Annual iPhone and iPad releases have to succeed in a huge way to keep the company’s growth and profit trajectory going. The problem is that there is always the risk – entirely theoretical so far – that one of these hits will turn out to be a dud. Adding the kind of annuity revenue stream an Apple TV service would offer – several hundred dollars per year per customer – would provide the sort of predictable revenue stream that investors like to see, which might offset some of the perceived risk associated with the hardware business.

The other big thing this service would do is reinforce the appeal of the Apple ecosystem, and of an all-Apple portfolio of devices in the home. Many people today combine iPhones with Nexus tablets, iPads with Samsung Galaxy smartphones, and all of the above with a Roku, a smart TV or a Playstation or Xbox for TV viewing. An Apple TV service, available like most of Apple’s other services exclusively on Apple devices, would provide a powerful glue that would stick all those devices together and give consumers very strong reasons to buy only Apple products.

The big questions remanning are whether such a service would actually make any money, and whether it’s even possible for Apple to put such a service together. On the first point, there’s solid evidence that it is possible to make very good margins in this business. The closest parallels to what Apple would offer are the cable networks, since they also acquire content rights and sell them to consumers on a subscription basis, but without the delivery infrastructure that CST providers need. These companies in the US have very good margins, in most cases at or above Apple’s recent levels (these numbers are all specifically for the cable network divisions of these companies):

US cable network marginsAnother company with an analogous business model is Netflix, particularly its US streaming business, which has been generating steadily increasing margins which are now in the mid-20s, just below Apple’s recent levels. While neither Netflix nor the cable networks’ businesses are exact analogies for what Apple would do in this space, there is good evidence here that packaging and delivering high quality video content on a subscription basis can be very profitable.

The bigger challenge is content rights, which is no doubt why we’ve seen many stories over the last several years about Apple trying to acquire them. The kind of bundle of live, recent and back catalog content rights we’re talking about here is unheard of, and Apple would be charting new territory (something it has a good history of doing). But Apple has both the existing audience (Intel’s biggest problem), content relationships, and DRM and other structures in place to reassure content owners that it can make a success of this business. And DISH has reportedly just signed just the sort of deal Apple will need to sign with Disney, already a strong Apple partner, which is promising. However, this remains the biggest challenge to this strategy, and the Time Warner Cable talks may well have been a sign that Apple wanted an interim solution while it works to create the ideal product.

The other thing worth thinking about is broadband providers’ bandwidth caps and their impact on this sort of service. As long as people are only using Netflix and Hulu to supplement traditional CST-provided TV, their bandwidth consumption is likely to stay within the 250GB caps that are starting to be implemented by major broadband providers. But if such a service became the only way a household consumed video, it would significantly increase consumption and might lead to problems.

Overall, the only way I can see for Apple to turn the Apple TV into something more than a hobby is to truly disrupt the TV space, and the way to do that is to launch not a TV set, but a TV service, with Apple TV, iPhone and iPad as different ways of accessing that service. This would finally turn Apple TV into a truly unique product, it would provide a very different kind of revenue for Apple in the form of predictable, recurring monthly subscription fees, and it would cement the Apple ecosystem as a compelling option in consumers’ minds. And this would be a substantial new revenue opportunity for Apple – just those cable networks listed above generated over $70 billion in revenue last year, almost entirely in the US. Given the content relationships Apple already has in many other markets, it could expand the service beyond the US in the coming years, providing a significant ongoing source of revenue growth.


How to make Devices and Services work for Microsoft

Back in 2012, Microsoft’s then CEO Steve Ballmer started talking about Microsoft as a Devices and Services company rather than as a software company. This was a major strategic shift, and it was behind major moves like the launch of the Surface and the acquisition of Nokia’s Devices and Services business. This is a signal that Microsoft recognizes the threats to its hitherto core software businesses, but also a fundamental move towards hardware and services-based models for monetizing software. Apple is arguably in many ways a software company, but it now monetizes essentially all its software through hardware purchases. Google is a software company, but makes none of its money through traditional software licensing models.

Two big questions arise from this: firstly, does it even make sense for Devices and Services to be the basis of a strategy? and secondly, if it does, what are the keys to doing this successfully? Let’s take each of those in turn. (Note: Microsoft’s Devices and Services strategy applies both to its Consumer and Enterprise businesses, but I’m going to focus here on the Consumer side).

Does Devices and Services even make sense?

Ben Thompson recently argued that a Devices and Services strategy is fundamentally flawed:

The truth, as I’ve written multiple times … is that a “Services and Devices” strategy is fundamentally flawed. Either be everywhere with your services, or differentiate your devices.

I see what he’s saying, but I don’t believe it’s right. I think there’s an enormously powerful link between devices and services which the two most successful players in the industry today – Apple and Google – have both exploited. When you have compelling services, and when the best way to experience those services is to use them on your devices, then there’s a virtuous circle between your services and your devices which allows each to benefit the other. As people find the services compelling, they seek out devices which provide the best experience for those services, and become loyal to them. As more people buy those devices, more people use the services, which makes them better and creates loyalty to the broader ecosystem of devices which provide similarly optimized experiences.

Apple and Google have approached these things slightly differently – Apple considers that its services should not just be best on its devices but exclusive to its devices (with the exception of the desktop iTunes software), whereas Google benefits greatly from wide adoption of its services across platforms but optimizes them for its own as a driver to get people onto the platform it can monetize best (Android). Both companies, though, are effectively combining devices and services, and if you ask people why they buy either iPhones or Androids the answer is many times that they want easy integration with services from each company that they’re already using (such as iTunes or Gmail).

The counter-argument to all this is that these two companies have each chosen either Devices or Services as their core business, and only participate in the other as a way of adding value to that core business. That’s true to some extent, and it’s what’s behind Apple’s exclusivity approach, but I’m also reminded of the Alan Kay quote Steve Jobs was so fond of: “People who are really serious about software should make their own hardware”. Ultimately, the most successful companies in this space will combine hardware, software and services to create compelling experiences. That’s certainly been true for Apple, and I’d argue it’s increasingly true for Google too as it gets deeper into hardware through Nest and Google Glass even with the disposal of Motorola.

What are the keys to success?

If combining Devices and Services does make sense, then the next big question is what it takes to be successful in combining those two at the heart of a strategy in the consumer technology market. Again, we can learn from Apple and Google here. I believe the keys to success lie in answering certain questions:

  • How will you make money?
  • How far along the better/exclusive spectrum do you want to go?
  • What are the compelling services around which you will build your strategy?

Business models

Although companies don’t have to pick either Devices or Services as a core strategy, they do need to establish a core business model. Which will they attempt to monetize: hardware or services? Amazon sells hardware, but does so with thin or negative margins as a way to sell more services (content). Apple makes large margins on hardware and gives away most of its software and services. Google sells some hardware (Nexus devices) essentially at cost, and services for free, and monetizes them through advertising. One of the key questions for Microsoft is how they want to make money going forward in the consumer market. Will it be hardware alone? Mostly services? Does charging for software still play a role? A big part of Microsoft’s current problems stem from the fact that, despite the headline shift to a Devices and Services company,  it still acts as if it sees Windows licensing – software – as its core business. In order to pursue a focused Devices and Services strategy, Microsoft needs to resolve this conflict.

Only on Microsoft devices, or better on Microsoft devices?

Making the decision about how it wants to make money will help Microsoft make the next decision: does it want to take the Google approach (available on many devices but best on Android) or the Apple approach (exclusive to Apple devices), or land somewhere in between. If the main monetization strategy is hardware, then Apple’s exclusivity approach makes the most sense. If it’s services, then going big makes the most sense. Microsoft faces two fundamental challenges here which are somewhat unique to the company: firstly, it’s used to providing market-dominant services and software, and secondly one of its most compelling services – Skype – didn’t start life as a Microsoft service. But it needs to decide how exclusively tied Microsoft’s services will be to Microsoft devices, and if they’re not exclusive then how to make them better on Microsoft devices. The latter is a lot harder to figure out, but there are some lessons from Google’s approach:

  • Better features / functionality, or earlier access to new features – Google Maps’ navigation features, Google Now and many others both started on Android before moving to iOS and continue to enjoy some advanced features on Android which iOS doesn’t offer.
  • Single sign-on. I’m a heavy user of Google services on my iPhone, and there’s nothing worse than having to sign in to every separate Google app, especially if you’re using two-factor authentication. The Android experience significantly simplifies this by having single sign-on for all Google services and apps (much as Apple does with iTunes / iCloud accounts on iOS).
  • Unique features – Google Keep, Google Wallet, Google Play Movies and other apps are all exclusive to Android in their full form.
  • Tighter integration with core services – Chrome and Google Voice can be set as the default apps for web browsing and phone calls on Android devices, but can’t on other devices.

Microsoft is already doing some of this with its services on Windows Phone Windows 8, but there’s a lot of room for improvement.

Creating compelling services

But the biggest single question, and the one I think Microsoft is going to struggle with most, is what are the compelling services that will drive people to Microsoft’s devices? So far, Microsoft has made its two biggest money-spinners – Windows itself and Office – the centerpiece of its marketing around both Windows 8 and Windows Phone. But the fact is that neither of these is all that compelling to consumers, and they have nothing like the driving force of iTunes or Gmail or Google Maps in bringing people to the platform. So what’s the answer?

The two key things people use their personal devices for are communications and content, and the solution to Microsoft’s challenge is to develop, create or acquire compelling services in both of these areas. A quick look at Apple and Google’s most compelling services reinforces this point:

  • Apple – iTunes (content purchasing and management), iMessage and FaceTime.
  • Google – Gmail, Google Maps, Google Play.

These services are big reasons why people buy iOS and Android devices respectively, but what are the equivalents for Microsoft? Windows Phone’s biggest single problem is that it doesn’t have a compelling set of services that are driving people to buy it. Today, the best selling devices on Windows Phone are the cheapest, suggesting that the purchasing decision today is much more about the price/quality equation than any apps. The Surface continues to sell poorly, despite the tight Office integration, suggesting that this isn’t the solution. Microsoft has several assets in the communications space, not least Skype but also, and it has stores for purchasing apps and content. But it needs to go significantly further both in developing and creating compelling apps and making the experience on Microsoft devices better than on any others.

Until it does these things, Microsoft’s Devices and Services strategy is doomed to fail. Nokia makes some great hardware, but so does HTC. Neither company is selling bucket loads of devices despite the great reviews the hardware regularly gets, which is further proof that combining devices and services is critical for success in this space. Microsoft desperately needs to create compelling services, and then find ways to make those services perform best on Microsoft devices, if it is to spark interest in its devices and make a success of the Devices and Services strategy.

In this context, it’s useful too to think about the Nokia X line announced this week. How does that fit into a Devices and Services strategy? The whole thing is built on Android AOSP rather than Windows Phone. But if the strategy at Microsoft is around Devices and Services rather than necessary operating systems that may be OK, especially since the Nokia X line will reach parts of the market Windows Phone can’t address today anyway. The key thing is that (a) this is a platform that Microsoft can control, even if it doesn’t own it, and (b) it can put Microsoft services front and center, optimizing for them in the same way it would on Microsoft’s own platforms. As such, I think it fits fine with the overall Devices and Services strategy. The key problem is the same though: Skype, OneDrive and just aren’t all that compelling as a set of Microsoft services around which to drive loyalty, on the Nokia X line or with any other Microsoft devices.

The Windows 8 Mistake

The CEO succession at Microsoft and the approaching Build conference have made Windows 8 a hot topic again, raising lots of questions about where the OS goes from here. While there are lots of theories, few people seem to be going back to where Windows 8 came from in the first place, which is really useful if you’re trying to figure out where it might go next.

Windows 8 was about pulling tablets towards PCs and away from smartphones

Microsoft missed the boat on the current generation of smartphones. Though it was a significant vendor in the early history of the smartphone with Windows Mobile, it misjudged the entry of the iPhone, its impact on the market, and the resulting dominance of iOS and Android. That history has been well covered and doesn’t need to be rehashed here. But the important thing is that, as Microsoft saw the re-emergence of the tablet category following the launch of the iPad, it saw that tablets were being formed in the image of smartphones, not PCs, and that was enormously bad news.

Why was this so horrifying for Microsoft? I believe they saw even then that tablets had the potential to displace PCs, but the smartphone model had several disadvantageous features from a Microsoft perspective:

  • The smartphone space was already, by 2010, coming to be dominated by two models: a tightly-integrated hardware/software package exemplified by Apple (and by then less successfully by BlackBerry) and a free, open source software licensing model embodied by Android
  • Operating system licenses on smartphones are much lower than on desktops for Microsoft, and non-existent for essentially everyone else, threatening Microsoft’s Windows revenue and margins
  • Smartphones ran entirely different apps from the applications that ran on desktops, with no cross-compatibility, eliminating the opportunity to sell existing versions of Office
  • Smartphones were based on ARM architectures rather than on Intel, making them fundamentally incompatible with all the existing Windows software.

The diagram below illustrates the situation as Microsoft must have seen it in 2010:

Windows 8 tablet strategyTablets could, at that point, theoretically go either way, and Microsoft certainly had the power to decide which way it wanted to pull tablets, whether towards the smartphone model or towards the PC model. And it clearly decided that its future depended on applying the PC model, rather than the smartphone model, to tablets. A PC model applied to tablets would allow it to continue charging high licensing fees for Windows, make Office applications easily available on the devices, and make them compatible with existing Windows applications from third parties. But it’s important to note that this was a decision driven entirely by what was perceived to be best for Microsoft, but by what would be best for the actual users of the products (The one counter-argument is that Microsoft would make Office available on tablets in more or less fully-fledged form, and this could be user-friendly as well as helpful for Microsoft’s bottom line.)

This, then, explains Windows 8, as the supposed solution to these problems. Microsoft would make the desktop, and not the mobile, version of Windows the core of the tablet experience. But by doing so, it forced on the desktop experience lots of things that made sense in the tablet world (the Metro UI, touch screens and an app store) but didn’t make sense there. Instead of taking a consumer-led approach and unifying two fundamentally similar products, smartphones and tablets, with a single OS, Microsoft tried to bridge the gap between two fundamentally dissimilar products, the desktop and tablet. And all of this was in the service of establishing the PC operating system licensing model and not the smartphone OS licensing model on tablets. 

But of course when it came to tablets, Microsoft had to make compromises to compete on price and form factor: the choice of ARM rather than Intel for Windows RT, and the resulting limitations of RT devices, ironically made them a lot more like smartphones, while eliminating almost all the consumer benefits of creating tablets in the image of the desktop. As a result, whereas the other two major platforms have one big dividing line from an applications point of view – between smartphone/tablet on the one hand and PC on the other, Windows has two – between smartphone and tablet, and between tablet and PC.

Where do we go from here?

If Microsoft’s ambition was to pull tablets in the direction of PCs, it clearly failed. The iPad and now a range of Android devices have enthroned the two dominant smartphone operating systems as the OSs of choice for tablets too, and there is nothing Microsoft can do to stop this at this point. Its own tablets have sold poorly, and its OEMs’ tablets haven’t sold that well either. Part of the challenge is that Windows tablets are competing on an uneven playing field from a cost perspective – wrapping in a licensing fee for a full operating system significantly increases the price for Windows tablets over the price of iOS and Android devices which don’t have to cover licensing costs.

But that’s far from the only reason why Windows 8 has been a relative failure. The bad decisions described above, which flowed from a desire to bolster Microsoft’s two cash cows rather than to do what was best for users, have left it with an operating system (or two) which meets almost no-one’s needs well, and sales have reflected that. So, where should Microsoft go from here? The following should serve as a good to-do list for starters:

  • Merge Windows Phone and Windows RT, mirroring the existing iOS and Android structures, and rename Windows RT as Windows Tablet.
  • Make both flavors  of the merged mobile OS free for users and OEMs, eliminating licensing fees
  • Do much more to promote consumer services, notably Microsoft’s own Music, Video and Gaming stores and offerings, across its consumer devices (smartphones, tablets and Xbox)
  • Continue with Windows 8 as a separate operating system, making Metro an optional overlay UI for touch-screens, but allowing users to choose the old-fashioned desktop UI as their primary or only UI if they so choose.

I’ll address each of these below in more detail.

Windows RT has largely been a flop, precisely because of those compromises and the confusion created in consumers’ minds about what it is and how it relates to the full version of Windows 8. It shares much more with Windows Phone than it does with Windows 8 in its limitations and in its architecture, and as such the answer is to rebrand Windows RT as Windows Tablet and merge it with Windows Phone, while leaving Windows 8 as a separate entity that works primarily on desktops, laptops and a few convertibles including things that look a lot like tablets but behave more like laptops in their functionality. This would be a recognition that Apple and Google were right all along, and thus a painful one. But it would also open the door to a couple of other strategies Microsoft could pursue.

Next, it should stop charging for licenses for Windows Phone and Windows Tablet.  By my own calculations, Microsoft likely makes under $1 billion from Windows Phone licensing annually, the vast majority of which comes from Nokia anyway. By ending license fees for Windows Phone, Microsoft would remove one of the competitive disadvantages every Windows Phone vendor labors under, allowing them (including Nokia) to either reduce prices or increase margins on their device sales. This doesn’t solve the other fundamental problems with Windows Phone, but it should at least accelerate growth. Microsoft ought to capture far more than $1 billion annually from growth in the hardware market.

Making Windows Tablet free would solve some of the same problems for Microsoft and its OEMs in the tablet space as it would in the smartphone space. It would reduce the cost of making the more consumer-centric Windows tablets, allowing them to be more price-competitive, potentially stimulating demand for the Surface among other tablets, which would also benefit Microsoft on the hardware side. Microsoft could then sell Office RT as an add-on for these devices (or continue to bundle it for free as a unique selling point), but should recognize that the primary reason most people buy tablets is not to do work on them.

The corollary here is that Microsoft needs to do a much better job selling its consumer services, the other half of its Devices and Services strategy. Microsoft now has competitive Music and Video stores, which are available across Xbox, Windows 8 and Windows Phone, but which it does very little to promote. These devices can absolutely be viable entertainment devices, whether by using Microsoft’s own services or the Netflix, Hulu and other entertainment apps available on them. But Microsoft says very little about these in its marketing of any of its devices, whether Windows Phone, tablets or PCs. Apple has built up an installed base of 500 million iOS devices by most estimates, and the associated content and application stores now generate almost $10 billion in net revenue for Apple annually. Microsoft’s massive base of Windows PCs combined with a stronger base of Windows-based smartphones and tablets could come to rival the size of this services and content business quickly, on top of the existing $7-8 billion annual revenue from the Xbox platform. Together, this revenue stream could easily outweigh the lost revenues from Windows RT licensing, which must be minimal today.

The desktop/laptop version of Windows 8 should continue more or less as it is. There’s nothing fundamentally wrong with it, especially in its desktop UI version, but Microsoft has caused itself huge heartache by imposing a user interface designed for touch screens on tens of millions of non-touch devices. The Metro interface is fine on smartphones and tablets, but both unfamiliar and poorly suited to a traditional trackpad/mouse and keyboard device. As such, it should make the Metro UI an optional overlay for those users who see value in it (probably mostly touchscreen users) and allow other users to banish it entirely and revert to the desktop interface. Another lesson Microsoft can learn from Apple is that, although it has driven significant convergence between the desktop and touch versions of its operating systems, it’s all been in the systems and infrastructure rather than in the UI, and there are good reasons for that.

To sum up, if Microsoft is to become a Devices and Services company on the consumer side and not just on the enterprise side, it needs to do what it can to stimulate the virtuous circle that exists between the two: compelling services drive device purchasing, and device purchasing drives usage of services, which in turn creates greater loyalty around the ecosystem. Today, Microsoft is so obsessed with driving Windows and Office revenues that it often takes steps which are either irrelevant to, or at worst, counter-productive to maximizing device sales and installed base and the usage of services. Taking the steps outlined above would stimulate device sales by lowering costs, incentivize OEMs who are on the fence about making Windows smartphones and tablets to give it another go, and at the same time foster Microsoft’s own burgeoning hardware business. On top of it all, it would finally give Microsoft a shot at building loyalty around truly consumer-centric services such as content consumption rather than insisting that Office and Windows are the only things around which it wants to build loyalty in the consumer market.

Facebook and Twitter’s growth challenge

Facebook and Twitter are at different points in their histories: Facebook just celebrated its 10th anniversary, generates profits each quarter, and has become a dominant force in global social networking, while Twitter has a fraction of Facebook’s users, loses money and seems to suffer from a crisis of identity about its role in the world. But the two companies share a common path to growth, and it’s instructive to look at where each of them is on that path.

Fundamentally, both companies need to do three things: grow users, get their users more engaged, and then find ways to monetize that engagement. Twitter actually makes this pretty explicit, but both companies provide metrics that allow us to measure how they’re doing on each of these three strategic objectives.

User growth has peaked at both companies

First, user growth. The chart below shows the growth in the number of monthly active users (MAUs) year on year at both companies, as reported by each of them.

Twitter and Facebook growth in MAUsWhat’s clear is that both have peaked in terms of user growth. Facebook peaked way back in early 2011 at around 260 million new users year on year, whereas Twitter seems to have peaked in 2013 at around 70 million new monthly users year on year.  Since user growth is the major lever for overall growth, this is notable in its own right, but it’s particularly worrying for Twitter at 232 million monthly active users than it is for Facebook at a billion more. To be clear, both are still adding subscribers at a decent clip, but the pace of growth has slowed at both, suggesting that if that growth happens in an S-curve as it typically does, both are already in the top half of the S. Twitter appears to recognize this, and on the recent earnings call CEO Dick Costolo talked about the challenge of shifting from a world where growth just “happened to” Twitter to one where the company actively has to seek growth. But recognition that something has to be done is not the same thing as knowing what to do about it. Unless Costolo and his colleagues can figure this out, Twitter’s future growth will necessarily be constrained and it’s doubtful it will ever come close to Facebook’s scale.

As context for Twitter’s current scale, here’s a chart that shows user numbers for various other popular services:

User accounts for major servicesTwitter is shown in light blue, and although it’s recently passed Amazon’s number of active user accounts, it’s far behind Apple’s iTunes and iCloud user numbers, and just barely ahead of the number of Dropbox accounts, which is also growing more rapidly. If Twitter is going to be a mass-market phenomenon globally it needs to both grow significantly more quickly and achieve significantly greater scale.


Twitter talks explicitly about engagement, and uses timeline views per MAU as its chief measure of engagement. Facebook doesn’t talk about it in quite the same way, but it does provide regular reporting on its daily active users, which are a good proxy for the proportion of users who are actively engaged.  The chart below shows these engagement metrics for Twitter and Facebook. The first chart shows timeline views per MAU for Twitter, while the second shows the percentage of MAUs who are also daily active users at Facebook.

Engagement measures for Twitter and FacebookThe good news for both companies is that these metrics are very much heading in the right direction. The exception is the most recent quarter for Twitter, where timeline views per MAU actually fell, but this was the result of tweaks to the Twitter mobile apps, which led to less going back-and-forth to and from the timeline, depressing overall numbers[ref]See Dick Costolo’s remarks on the earnings call[/ref]. So it’s a one-time downward adjustment of the curve rather than a sign of a longer-term downward trend.

The other thing that’s worth noting is the variation by region. For both companies, engagement is significantly higher in the US than for their other regions. Facebook actually has a very balanced business in terms of users across the several regions it reports, with between 200 and 400 million in each of its four territories. Twitter, on the other hand, still has a very US-skewed user base, with 179 million domestic users and only 53 million overseas users, which only adds to the concerns about its slowing growth: it’s clearly struggling to achieve anything like its US penetration levels outside the US. However, both companies are successfully increasing engagement both domestically and abroad, to the extent that engagement numbers in overseas regions today are close to engagement levels from earlier periods in the US. This suggests that they are successfully matching US user behavior with overseas user behavior, just with a time delay of one to several years.

Overall, though, both companies are successfully increasing engagement, which means that there is at least one driver of growth even as user growth slows. However, Facebook’s engagement growth is slowing in the US, its dominant region, which is likely a sign that it is reaching saturation among its existing base, so it’s not all rosy.


Both companies’ revenue is directly tied to advertising, which is another way of saying that they both have to find ways of monetizing the user growth and related growth in engagement in order to grow revenues. So the third lever for growth is increasing the amount of revenue per user, which is shown in the next chart:

Ad revenue per MAU for Facebook and TwitterThe first obvious thing is that Facebook’s ad revenue per user is much higher, which of course reflects its greater maturity and experience in providing advertising to users. But both are rising in a healthy way, which is helping to drive overall revenue growth when combined with user growth and user engagement. Interestingly, though, both companies do far better at monetizing their US users than they do international users:

Regional ad revenue per MAU for Facebook and TwitterThis is particularly dramatic for Twitter, which generated just 23 cents per MAU in the fourth quarter outside of the US, compared with over two dollars per MAU in the US. Facebook sees a greater spread between its various regions, with the US generating over $5 per quarter per MAU, Europe at $2.36 in Q4, and its other regions at under a dollar per quarter per MAU. Again, the good news though is that each company is successfully growing revenue per user.

The other thing worth thinking about in this context is what’s possible longer-term. Unlike other forms of revenue generation, advertising has a natural saturation point at which it becomes off-putting to users, which puts a cap on revenues per user. The revenues per user in the US of Google and Yahoo, long-established US-based online advertising businesses, are vastly different, and show the range of what’s possible. For this exercise, I’ve used Comscore’s numbers for desktop unique users for both Yahoo and Google sites as a good proxy for their number of US users. It obviously excludes mobile users, but my guess would be that neither number would rise dramatically if mobile-only users were factored in since the total numbers are both so close to the size of the US online population. The chart below shows US revenue per US user (using that Comscore data) over the last few quarters for Google and Yahoo[ref]Yahoo hasn’t reported its US revenue numbers for Q4 2013 yet[/ref]:

US revenue per US user for Google and YahooGoogle clearly generates enormously more from its US users than Yahoo does at over $30 per quarter. Yahoo, on the other hand, generates just over $4 per quarter from its US users. Facebook has already passed that amount, and Twitter is about halfway there at this point. But neither is anywhere near catching Google yet. This obviously tells us a lot about Yahoo’s struggles too: applying the same growth strategy framework to Yahoo suggests that it’s maxed out user numbers in the US and it’s engagement and monetization where it needs to make significant progress. The big challenge here is that Google has long since hit upon the holy grail in online advertising: search. Search is unique in that it pairs advertisers with users who have expressed an explicit current interest in the item being advertised, resulting in very high hit rates. By contrast, Facebook and Twitter can only hope to derive enough about their users’ general interests to offer up profile-based targeting, which is inherently less effective (though it may help explain Facebook’s interest in graph search and other forms of search). See this separate post on advertising business models for more on this.

The other thing that’s worth noting is that Google still generates a large proportion of its overall revenues from the US, even though its overall user base is obviously much larger in the rest of the world. This suggests that even mature online advertising businesses will always see much higher average revenues per user in the US than in the rest of the world, and it’s not just a question of a lag as with engagement. This is another important factor for Facebook in particular to bear in mind as its growth in the US slows to a crawl.


So where do we stand in evaluating Facebook and Twitter’s growth challenges? Both companies have stopped growing their user bases as fast as they once did, though Facebook has hit that point with about a billion more users than Twitter and is still adding about a quarter billion each year. Twitter’s slowing growth is worrying and should be a major priority for the management to fix. However, both companies are successfully finding ways to get their users to spend more time on their services, which combined with increased efforts to develop effective advertising products has led to significant growth in revenue per user. Both companies could continue to grow revenue significantly in the coming years even with slowing user growth simply by growing engagement and revenue per user, though at some point both companies will hit up against the natural limits imposed by advertising based business models.