Speculating about the future is not easy. Traditional wisdom looks to technology industry history as our guide. The lessons from this history often shed great insight into the future. However, using this history has its challenges and requires context. But using history as our guide, I think it is very interesting to understand the role services will play in the future.
History as Our Guide
I was at a conference a few years ago and heard a presentation from an early executive at Compaq. He shared that when we look at the history of major segments, such as mainframes, mini computers, and desktops. we observe a familiar pattern. That pattern is related to how value evolves through a segment’s lifecycle. His point was that in the early days of a segment the value starts in hardware. During the hardware stage the innovation in hardware, components, form factors, etc., is the driving value proposition for the buyer. His next point was that once the hardware reaches some level of standardization the value then moves to software.
During the software stage all the value is found in software that takes advantage of the compatible build-up of hardware. In this stage, hardware is largely commoditized and all the value and opportunity is in software. He pointed out that while there was still some value left in the software cycle the next step is when that value shifts to services. Most software during this cycle is designed to take advantage of the services being offered. In essence software as a service becomes the longer-term sticky model.
To visualize this cycle as it was in mainframes, minis, and desktops, for example here is the slideshow showing how this happens. Cycle through each slide.
[slider id=’36323′ name=’Value_evolve’ size=’full’]
There is quite a bit of context within our history needed in looking at these slides but the core point is valid. For example, the history in the buildup of hardware may have played out very differently had Compaq’s Rod Canion caved and adopted IBM’s proprietary PS 2 bus architecture. Or perhaps if Gary Kildall would have called IBM back and they used his CP/M OS instead of MS-DOS. But those are stories for a different time.
Using this view of history as our guide, it would suggest that the value cycle related to personal computing is destined to end up in services. Hardware will be commodity and solely exist to run the bulk of compatible software which is just a gateway to a service.
This vision is interesting but I’m not sure it plays out this way. Let me offer two views I have on how this could play out.
In my first scenario we play out what could happen if the primary view of history, which I outlined above, prevails. This is the view most similar to that embodied today by Google’s Chromebooks and by Larry Ellison’s vision where all computing is cloud-based and hardware is purely thin-client. In this vision the Internet is the platform, not proprietary OS layers.
Here is a slide I created for this scenario using the value shift methodology.
In this view we are still in the ‘beginning’ or the ‘golden era.’ Everything we see happening today with the fragmented platforms, app stores, and services is ‘just the beginning.’ During this stage each platform is evolving on its own related to hardware, software, and services, but will eventually lead to the Internet becoming the platform. This is the pure cloud-based, browser-based computing vision.
What is interesting in this vision is that it is good for software developers but not so good for pure hardware companies. If all hardware simply needs to standardize around web standards, then developers no longer need to write for individual software platforms. Instead all they have to do is create cloud apps that will run on all hardware through a browser or web standards-based operating system.
This vision is at the center of the debate over native apps vs. web apps in our future. This vision seems good for developers because they truly have a universal platform to write software and deliver software to the largest base possible. But it will challenge pure hardware players and force them either more into the software game or services game.
This view, if it does happen, will not happen for quite some time (10 years, at least). I say that because this vision requires that a degree of hardware innovation be exhausted and we are nowhere close to that. It will also require a massive amount of infrastructure related to cloud and wireless networks that are nowhere near where they need to be to support this scenario.
Generally this view favors the strategy of companies like Google and Amazon. It is also the model driving Xaomi’s growth in China. The sheer possibility of this scenario is why companies such Apple, Microsoft, Samsung need to be aggressive about their services strategy.
In this scenario the future is a hybrid of the models described above. In this future, there will always be some value in hardware, there will be value in native apps but there will also be a plethora of quality web apps and of course increased value in services. In this view, however, platforms such as iOS, Android, Amazon’s platform, perhaps Windows Phone, or even something else all evolve in parallel down the hardware, software, and services value chain. This view favors a more proprietary ecosystem approach that wins with deeper integration across all three segments.
Thus it may look more like this for the specific players I am pointing out.
In this view the market will sustain a range of approaches and more importantly choice will exist throughout the value cycle. You could argue this scenario benefits the more vertical or integrated players. Google, Microsoft, Amazon, and Apple all have varying degrees of a vertical integration strategies. Each can tightly tune its hardware, its software, and its services to create a differentiated offering in the marketplace. We know from many segments that differentiation to mass market consumers is important.
This does not preclude the more cross-platform services such as Netflix or others from being successful. Only that those companies which offer services on all platforms run the risk of being out-integrated on certain platforms by the vertical player who seeks to offer competing services. The players with a services strategy such as Google, Amazon, Apple, Microsoft, and any else that could do it, have the ability to more tightly integrate their services with their hardware and software. Thus they win with an integrated approach. Integrated approaches are more convenient to mass market consumers, easier to use, and thus will be perceived as valuable (that is, worth paying for).
I can see both these scenarios playing out. In my mind the likelihood of each holds equal weight. Which basically means I’m not sure which will come to pass. If Scenario #1 is more modular, Scenario #2 is more integrated. The point that remains, however, is that any way you look at it we are still only just beginning.
21 thoughts on “Our Services Destiny”
The commodification of hardware presupposes that nobody cares about the hardware in question as long as it does what it has to do. This works for some kinds of hardware (refrigerators, washing machines, TV sets), but not at all for others (watches, cars, clothes).
The problem with going by computing history is that computers have until now been sold mainly to corporations, who don’t care about intangibles or brand cachet. Suddenly now computers (in the form of phones and tablets) are being sold mostly to individuals, who do very much care about intangibles and cachet.
If you think people are going to come to regard their computers like TV sets or appliances, then yes, a shift in value from hardware to services makes sense. But if they are going to regard their computers like watches or cars, then the lots of value will continue to remain in the hardware.
The next thought is that while some people care very much about their cars, others simply regard their cars as a means of transport — and this difference in attitude cuts across income. So some people will gravitate towards computers with brand cachet, while others will just buy the computer that does what they need it to do.
Which means, we might simultaneously see some computers being commodified with the value rising up the stack (arguably happening already with some android devices), and others continuing to have the bulk of their value in their hardware (with the hardware having luxury and brand icon status, ie, apple).
I think this is why I lean toward scenario 2 more than 1. However, I don’t think we can deny that the infrastructure and technology will exist to have pure browser based software that works and feels native. I’ve seen technology in a controlled environment already where I have seen what feels like desktop software (like Photoshop) fully running in the cloud.
So that is why I feel there is a hybrid but if we run this scenario out then what happens when all software is browser based and thus equal on all devices that simply support the web standards making this possible? Proprietary apps are no longer a differentiator therefore it moves to services, tighter integration, better overall experience, etc.
They key as you point out is to understand what will be found valuable by end customers in these scenarios. That is why given what we know of consumer markets it seems logical that to groups of people at least there is value in the hardware but the value has to be much more in other things, still tangible, but in other ways.
” However, I don’t think we can deny that the infrastructure and
technology will exist to have pure browser based software that works and
feels native. I’ve seen technology in a controlled environment already
where I have seen what feels like desktop software (like Photoshop)
fully running in the cloud”
Sure, apps for desktops might move to the cloud, but not for mobile devices. Phones and tablets are always going to be severely memory and CPU constrained compared to bigger computers (cf http://sealedabstract.com/rants/why-mobile-web-apps-are-slow/ ). Cram enough RAM and CPU into one to handle desktop-class web apps, and the battery life will become unacceptable compared to one that’s optimized for running native apps. Which means that the concentration of value *cannot* flow as far up the stack for phone devices as it can for non-mobile computers.
Isn’t the premise of cloud computing and browser based software that all the heavy lifting takes place in the cloud? In essence all the same constrains can exist in mobile as they are today and off load all the memory and process to the cloud.
Plus in 10yrs from now memory will be in a very different place.
Sorry but we disagree.
The cloud, by definition, has high latency. you have to wait for your data to upload and download. Using your example of a web-browser based photoshop clone — a lot of the image manipulation has to happen in the browser because if you’re uploading, recalculating, and then downloading everything every time the user moves layers around, the whole thing will feel hopelessly sluggish. So no, cloud computing means that a lot of computation has to be happening in the web browser, unless you postulate some future utopia in which everyone enjoys gigabit/s data transfer rates no matter where they are or what the reception is like.
The article I pointed to explains why it’s unlikely that browser web apps for mobile devices are going to ever be as capable as desktop apps. But aside from the arguments he makes there, even if Moore’s law continues to do its work (and all indications are that we’re going to hit an end to affordable improvements in silicon in the next 10-ish years), mobile devices are always going to be fighting trade offs. Every square inch of silicon devoted to RAM is space you aren’t devoting to a bigger battery. Every megahertz of CPU power has to be fed by the battery. Every ounce counts, every millimeter of thickness counts.
To be clear I’m not saying it WILL go this way only its one of several possible scenarios. Second, I absolutely see this from the semi-conductor view point. I started my career as a semi analyst and still engage heavily with all the vendors today. In fact it is several very large semi guys who showed me a controlled demo of desktop class software running in browser that felt as if it was native. I didn’t think this was possible until the demos I’ve seen as of late.
Second my scenario number 2 is more in line with your thinking. It is a hybrid. Some things will be possible or even better on mobile using a software and cloud services thin + heavy compute environment.
When you look at the semi roadmaps paired with the network topology roadmaps, it is hard to not believe that at least a hybrid model will exist.
Great article and I agree with this comment to a certain extent. I agree that this transition will be slightly different in that businesses were the primary customers for the mainframe, mini, and much of the desktop eras (i.e. a fairly common set of use cases). Now that computing power is much more ubiquitous, there are many more types of users to serve. This seems to explain why Scenario 2 seems more probable at least in the near term. Much of the current debate over web vs native and cloud vs local sounds very similar to previous client/server debates: It matters in terms of implementation at a single point in time but the platforms that endure will be those that are elastic. At present, this means a native mobile app that connects to a cloud service but is still usable as an independent client. Amazon, Google, and Apple all have relative strengths and weaknesses and until there is a clear winner (and there may never be just one), the resulting competition benefits everyone.
I would only take issue with the ‘brand cachet’ argument. It’s also about quality. People buy a Mercedes because a) they can afford it, b) the quality, c) it’s a Mercedes. Apple’s success has long been wrongly attributed to brand cachet. If we view the future through the same lens, our analysis will fail.
Brand cachet is a long lived phenomena. The Korean carmakers were perceived purveyors of junk for a long time. Their initial efforts were, then their quality improved to be as good as the rest, but it took a long time for that reputation to change. Likewise, some other makes that were long known for quality have stumbled recently, yet their reputation has been slow to tarnish.
Brand cachet is due to quality and/or other very real product attributes consumers value.
I get what you’re saying, but too many people don’t understand the subtlety and we end up with the myth that Apple is successful because of marketing, or because their products are status symbols. I would rather use the term ‘quality’ because that is more accurate.
One significant aspect of Web vs. native is stateless vs. stateful. A useful and battery efficient software must be stateful, ie. knowledge of user personal past history, to act correctly on current user inputs thru multiple sensors. 1) Does user feel comfortable and secure to store huge amount of personal data in the cloud for Web-base app? 2) Network and server are going to be faster. Does round trip time over network impact the responsiveness and smoothness of app?
number one will probably by for the business model and number two will be for us lab rats to beta the next ten years of “the computer” so business will be sure it works and they don’t get screwed and dead like the lab rats.
I think you could add apple in the three right columns of the desktop-market. Since 2006, the mac is a IBM-pc with different software and services.
It is not ‘open’ wins but Cheap wins as in everything.
Do you shop at thrift stores, buy food day old bakeries and only drive the cheapest car available? Do you go the cheapest medical facility you can?
Derp. One single person is statistically insignificant. (In other news, the sky is blue and water is wet.) Typical Silicon Valley narcissism.
Cheap wins is absolutely true for the majority. Do names like Walmart, McDonalds, and FOXCONN ring a bell?
Successful brands: Walmart, Mercedes, Carrefour, BMW, Target, Tiffany, Costco. The argument about whether cheap wins or not is silly. Cheap and luxury can both win, depending on the product and market.
Cheap does not win. Look around your home and look at what you’re wearing. Is it every the lowest cost item?
Initial Purchase Price or Total Cost of Ownership?
Macs were long known for lower TCO, while “cheap” Window boxes that required a lot of ongoing maintenance prevailed.
Any possible future will always be the beginning for something, that’s the beauty and the curse.
What is missing for me in these scenarios is a third dimension – level of technology.
I see the technology development as a spiral in where the service drives use cases and they in turn drive a new iteration of hardware. So every major player in mobile field has their own spiral. Time in this model is the length of the spiral. Spirals may intersect where one of the more successful iterations crosses the chasm.
I think in this environment the winner is the one who owns the whole vertical hardwaresoftwareservices and can iterate quickly. Apple for example has these characteristics, so I believe they are in a better position for the innovation.