The End of Standardized Platforms

Historians of the technology industry observed a pattern that was predictable with regards to new computing platforms. In the early days of a computing segment, like mainframes, minis, and desktops, there was a great deal of platform fragmentation. These early computing systems often ran proprietary software and operating systems with little interoperability. Eventually, a standard emerged — Windows. Even though Macs stuck around, their market share stayed well below 3% for much of the build-out of the PC era. Here is a visual to show how platforms started fragmented and then standardized around Windows.

Screen Shot 2016-02-12 at 5.30.49 PM

What you are looking at is operating system share of devices sold annually. Of computing devices sold for most of the time period above, Microsoft was the standard. Fast forward to today and we see a slightly different picture emerge. Windows runs on PCs and Android and iOS run on mobile devices, giving us three primary operating systems occupying the bulk of computing devices sold each year.

Screen Shot 2016-02-12 at 5.31.40 PM

You can see as the pie went from several hundred million computing devices sold each year to now almost 2 billion computing devices sold annually (when you add up PCs, tablets, and smartphones), the pie has gotten much larger but the landscape has also changed. While Android has the largest chunk of the pie, they do not have the 97% share Microsoft once had. The size of the pie and the global diversity of the consumer market brought with it the opportunity for several computing platforms to exist simultaneously.

If we take a step back and look at the installed base, we see an even clearer picture of the diversity in computing platforms in use today as well as the size of the market.

Screen Shot 2016-02-12 at 4.50.01 PM

There are now well over three billion active computing devices in the world with five primary operating systems/computing platforms — Windows, OS X, iOS, Android, and AOSP (non-Google) Android running in China. The key point here is, with the addition of scale in mobile and the inclusion of the global consumer market, there is no single standard computing platform. The question then is, what will happen with things like Virtual/Augmented reality platforms or artificial intelligence platforms? Should we expect VR/AR or Artificial Intelligence to unify and one single platform emerge like during the enterprise PC days or will many different platforms exist as we see today in consumer computing?

I tend to lean toward the latter. While VR/AR will start off segmented with Oculus having a platform, Sony having a platform, Microsoft having a platform, Google having a platform and even Apple having a platform eventually, it may also stay segmented rather than consolidated.

The global consumer smartphone market has shown us it can sustain many platforms so perhaps whatever comes next will follow the same paradigm. As I’m observing with wearables, where the market is actually developing into a rich segmentation, perhaps VR/AR or artificial intelligence will do the same, adding new layers of computing platforms onto the existing ones rather than consolidating into a single one.

Published by

Ben Bajarin

Ben Bajarin is a Principal Analyst and the head of primary research at Creative Strategies, Inc - An industry analysis, market intelligence and research firm located in Silicon Valley. His primary focus is consumer technology and market trend research and he is responsible for studying over 30 countries. Full Bio

39 thoughts on “The End of Standardized Platforms”

  1. Interesting thank you.

    I hadn’t realized the Windows installed based is actually decreasing. That’s a surprise, and unsettling for MS.

    I’d argue that for platform talk Google-Android and AOSP-Android should probably be merged since most apps don’t use Google Play Services (mostly store/payment, location and game ladders) and can run as is on any flavor Android, that would point to a new dominant platform, though not as overwhelmingly dominant as Window used to be.

    2 other charts would be very interesting:

    – ecosystems’ share of revenues: hardware + software + peripherals + ads + services.
    – ecosystems’ share of time spent (especially on the consumer side).

    1. When you dig relly hard into the Android platform data, you could argue those who truly use the platform and are engaged is around the same size of the iOS base. So I could make a pretty compelling argument that while Android has the dominant share… it is not the dominant computing platform….

      I separate AOSP because it is a beast unto itslef in China. There is an entire non-Google ecosystem built up that makes it pretty much its own platform.. Now of course one could argue if I should include WeChat since and Facebook in this as well..

      1. I’m not clear what “engagement” means. Time spent ? Money being spent / spent in total beyond phone (or including device) ? Interest in the device or ecosystem beyond its functional value ? Inability / unwillingness to switch ie Brand / ecosystem loyalty + lock-in ?

        And I’m not sure why it matters how “engaged” we are, as long as we spend our money and our time on the stuff / in the ecosystem. I’ve even heard of engagements ruining relationships ;-p

        1. Engagement being usage so time, downloads, internet usage, etc.. One thing that really gets missed is how the vast majority of the lower-end user, the kind Android has the largest share of, is still basically using their android smartphone as a feature phone or just to get on Facebook, Whatsapp, etc..

          Android has OS share, but is a runtime for other platforms who usurp engagement. This is the main struggle with Google’s revenue model for this platform and why so many wall st. investors are more bullish on Facebook than Google.

          Android has helped make it so the world can get online affordably and that is fantastic for humanity. At a business level discussion it is a different issue.

        2. “1- I’m not clear what “engagement” means.”

          Well, according to Ben’s data it means Android isn’t as dominant as you had assumed. I’m looking forward to the new metrics you’ll put forth to prove that Android really is dominant after all.

          1. I’m not the one making up alternative metrics when the standard ones don’t fit the narrative I want ^^ Share to high ? Let’s switch to “engagement”, whatever that mans. I’ve nerver heard of “engagement” for cars, clothes, PCs…

          2. “I’ve never heard of “engagement” for cars, clothes, PCs”

            Didn’t you claim to be a business major on some other thread? For a business major you don’t seem to have much business acumen.

            You’ve really never heard of CE in business or strategic marketing, sales? Part of my work is business consulting, we look at engagement *all the time*, pretty much in the same context Ben is talking about. Why? Because it’s *important* for *business*. Of course we’re aware that looking at something like total market share is possibly the dumbest way to get good answers or shape a strategy.

            What’s next, you don’t know what segmentation is? It would seem that you don’t, the way you carry on about total market share.

          3. Hey, while you’re here, what about the made vs assembled discussion of the previous thread ? You made a fuss about the difference, now that the difference is acknowledged, how does it change any of the points in that discussion ?

          4. Pointing out the difference between “assembled in” and “made in” isn’t making a fuss, it’s simply pointing out the difference. As a business major you should have known the difference to begin with, and you shouldn’t have to ask how it changes the discussion (the impact on a business strategy is tremendous between the two).

            Let’s summarize, you didn’t know the Moto wasn’t actually made in the USA (until I told you the difference between assembled in and made in), you’ve never heard of engagement re: business analysis, and you don’t seem to know what segmentation is. Yikes. Remind me never to hire you for anything related to business decisions. As always I am ever the gentleman, please have the last word.

          5. So, your point about assembled vs made regarding “the jobs are abroad” is …. ?

            Also, I as I said before, I do know the difference between assembled vs made, I was using the headlines ( https://www.google.fr/search?q=moto+made+in+the+us&ie=utf-8&oe=utf-8&gws_rd=cr&ei=d2rOVpCtBMPSU6nBtMgB#q=moto+made+in+the+us ), which indeed are misleading and you were right on that point. Not on my understanding of it, but of the fact it should be “assembled” not “made”. Can we mobe back to the real issue now ? Or do you want to keep advertising your small and irrelevant victory (you don’t get many or … ?)

            You just randomly posited I don’t know what segmentation is and now are deriding me for it ? You’re a crazy jackass too, I guess ;-p

            As for “engagement”, it seems to be a catch all of brand loyalty and time spent online with, which is weird, ie off line stuff doesn’t count ? I’m sure you have no thoughts on that though…

          6. Let’s repeat: “Engagement being usage so time, downloads, internet usage, etc.” Where does Ben say that engagement is “brand loyalty”? Where does Ben say that engagement is only “time spent online”? Why wouldn’t downloads include apps or content that are used offline and/or online?

            Seriously, please try to understand what is written, and stop making up strawmen.

          7. And stop taking everything at face value. “engagement” is a word for which we have no precise definition, and no measure. Sorry for taking it with a huge grain of salt.

            Edit: to me, customer engagement is people being interested in a brand, or maybe more generally topic. It’s obviously being used with another meaning here, something like “compound usage”, but I’m not sure exactly, nor how that is measured.

          8. “”engagement” is a word for which we have no precise definition, and no measure.”
            Ben responded to you and clearly defined the way he used engagement in his charts, of which you accused him of “making up alternative metrics when the standard ones don’t fit the narrative”. So why don’t you take his answer at face value, instead of making up random stuff about the word?

            “And stop taking everything at face value.”
            I guess you want us to take everything you say as having no value.

            “Sorry for taking it with a huge grain of salt.”
            You can be skeptical of the value of the metric, but there’s no reason to make up a different definition for it. Or do you not know what “grain of salt” means?

          9. Ben’s definition is “usage so time, downloads, internet usage, etc.. “. Then there’s the caveat that this usage doesn’t count as engagement if it is used on other platforms, which I take it to mean other apps. But maybe not all apps ? When is an app a platform ?

            That begs the question, how do you measure that and feed it back into the original downloads, internet usage… measure of engagement ?

            Even with all those fantastic Android homepage widgets, I do spend most of my time within apps, so engagement is the 5% of time I spend on the home screen ? The 10% of the time I spend in Google apps incl. homescreen ? The 30% of the time I spend there if you add web browsing with a browser ? the 80% of the time I spend there if you add web browsing within my RSS reader ? When I’m streaming a show from my NAS, am I engaged ?

            You seem to know the answer, and this is an actual question.

          10. Ah, a response that we should take at face value and discuss. That’s much much better.

            The only distinction Ben makes in the installed base chart is between Android and AOSP. He said the distinction is that AOSP-phone users (mostly in China) primarily engage with a non-Google ecosystem.

            He also wondered if he should’ve broken out Facebook and WeChat, thus taking installed base share away from all smartphone OSes, including iOS. If he’d done this, I’d assume the Facebook or WeChat ecosystems would include the smartphone users that primarily stay in those ecosystems when using their smartphones.

          11. Responding to your edit: I interpret engagement to be a business measure of how much a user uses the underlying differentiated OS/ecosystem. I think you’d frame it as how locked in the user is.

            So I would think it covers all the ways an ecosystem is differentiated and monetized by the ecosystem owner. On Android, it would include using Google apps and services (both online or offline, widgets, Google Play Store for content sales, etc). On iOS, it would include all the Apple apps and services (including iMessage, iTunes/App Store, iCloud, Pay, etc).

            If a user is fully engaged with Facebook for newsfeed, messaging/contacts, games, sales, etc, or with WeChat, and ignores the differentiated OS-related ecosystem, that user has much less business value to the OS owner and its device partners (because he/she is much more likely to switch to any phone that has Facebook/WeChat regardless of OS/ecosystem).

            Engagement could also include usage of third-party apps unique to one ecosystem, but this is more fluid and thus harder to track. It could also include usage of offline benefits, such as Apple Store/Genius Bar/workshop visits. Ben would have to let us know if these related items are included.

            Bottom line is I trust Ben knows what he is doing with regard to engagement, and it’s not some flimsy, vague, handwaving metric.

          12. That’s clearer, thanks. It does make engagement a rather vague mish-mash of usage (especially ecosystem/OS basic apps usage), monetization, brand/feature satisfaction, and lock-in.

            I wouldn’t necessarily call engagement lock-in: engagement is getting in, lock-in is being unable to get out. For example Hangouts and Skype have no ecosystem lock-in because they’re cross-ecosystem, iMessage has ecosystem lock-in. That’s independent of their respective engagements.

            The devil is often in the details. For example,
            – if usage is a part of engagement, how do you account for offline usage ? On the one hand, it diminishes online activity, an easily measurable usage hence engagement metric. On the other hand, it does increase usage, and probably increases ecosystem engagement (on the Android side, because SD cards + easy NAS access + file manager + wide format support, I’m sure there are equivalent differentiators on the iOS side). In the end… does it get measured for engagement, how ?
            – if online usage is counted, how does Web usage count ? Web is the crossest of cross-platform “apps”. Does it count as engagement for a specific ecosystem ? Is Web its own platform ? If I go to Google from an iPhone, is that iPhone engagement or Google engagement ? Same issue with Facebook, Youtube… plus is FB over the web counted differently as app-based FB ?
            – etc…

            It’s not so much a question of trust as a question of communication. It’s hard to discuss about a metric that’s obviously complicated, conceptually not described, and concretely not detailed.

          13. I’m not exactly sure what the issue is. I’m using a pretty standard definition for engagement here and one that is well understood. Even with my talks with Google they fully understand this so the engagement issue is well known even from the horses mouth.

            I’ve used the analogy that Android is essentially a BIOS and for the largest part of its user base that BIOS is enabling value for others than Google. This is why the street leans toward a negative view as they struggle whether Google can really beat Facebook in the long run.

            At the heart of the matter is whether Android is translating into value for Google or Google assets given their business model. From my discussions with them I get the sense Android is quite the burden resource wise and questions have come up whether it is worth the burden or if they should more broadly open source it or get it in the hands of someone else, etc.

            There is a user base of value to Google it just isn’t that big. It is no secret their most valuable user base given their business model is the 600m or so iPhone users. Android has enabled Facebook to usurp value and while it is a underlying connectivity runtime, this issue of what value they get from it remains a central point.

          14. “I’m not exactly sure what the issue is.”

            It would seem the issue for obarthelemy is that reality has a pro-Apple bias. Hence the repeated attempts by obarthelemy to obfuscate and cast doubt on your analysis.

          15. It’s not you. I believe the person having issue is being intentionally obtuse to spur arguments to protect a platform that needs no defensive behavior.

          16. The issue is that “engagement” can have several meanings.

            – In the general marketing sense, it means your target market cares about your brand and products. There’s a bit of that here, but mostly, that’s not the topic.

            – The way you phrase it now, it’s a synonym for monetization. It’s clear that Google has challenges here, because they make no money on the hardware, no money on the OS, no direct money on their services… only ads, and their few destination sites (youtube, gmail, … that’s it ?) that are enabled by having more smartphones in the wild. How Google could make more money out of Android would be an interesting topic to discuss.

            – Engagement and monetization kind of dovetail with usage, and it’s clear Android is skewed to low usage because if I’m not using it much, I’m probably buying low-end hence Android. There’s a question about how usage is measured too, I think Android comparatively favours offline usage, both because it’s cheaper Android, and because Android makes it easier. That usage is harder to measure, maybe underestimated ?

            – I think Engagement can also be misconstrued as meaning user fidelity, as in you say “low engagement” meaning “low monetization”, readers understand “low fidelity”, whereas I’d argue it’s probably the opposite: users with low monetization are probably quite happy with Android so there is high fidelity/engagement.

          17. “Seriously, please try to understand what is written…”

            Thank you for your humor, Mr. Jones.

            The Black Knight has so many layers of cognitive dissonance between himself and reality–what you ask is the impossible. You are requesting the band to play a song it doesn’t know.

          18. Previously, you wrote:
            “- Share too high ? Let’s switch to “app sales”
            – App sales too high, let’s switch to “engagement”, whatever that means.”

            Then you write:

            “As for “engagement”, it seems to be a catch all of brand loyalty and time spent online with, which is weird, ie off line stuff doesn’t count ? I’m sure you have no thoughts on that though…”

            It seems you want to switch the measure of “engagement” from the parts there is data for (data that is mostly gathered via various methods of monitoring online activity of different sorts), to a definition that centers on offline engagement.

            Presumably you want to do this, because you have the idea that Android users are not so much engaged with Google searches, surfing, Fb and WhatsApp, or buying stuff, and other “merely trivial” activities; the idea that these measures of engagement center around “brand loyalty”, and you are happy to leave that to iPhone users (as long as you believe that those things define most of the iPhone engagement), so much as Android users are engaged with other, less measurable forms of engagement, forms that you deem more important.

            You seem to think that Android users are spending all their time doing extremely useful and interesting stuff within apps or offline, as they “engage” with their Android phones in some meaningful ways that are not reflected in any of Ben’s data; and you seem to think that iPhone users do far less of this meaningful engagement, and are trying to deflect attention from it or something.

            So, let’s think about all that for a second:
            1) iPhone users are doing all the interesting, useful, in-app, offline stuff, too.
            2) Google and other ways to track data and online engagement has an even better grasp of what is happening with Android engagement than with iPhone engagement. So, if iPhone total engagement that can be discerned through online data gathering is measured right up there with total Android engagement, despite Android’s higher market share, then the true picture might be even more in favour of higher iPhone engagement.

          19. I’m assuming with that definition of engagement, FB counts s a different platform, so not as Android engagement ? Unless you have info otherwise ?

            I’m not sure that definition of engagement has much to do with brand loyalty either, maybe with ecosystem engagement ?

            I’m not saying there’s more or less important engagement, I’m just trying to work out what it is. You all seem to think you know, yet your explanations are confused and contradictory ?

          20. I’m fine with calling engagement with FB engagement with “another platform” other than either Android or iOS, no matter which is used to access FB for that engagement. I don’t know how Ben is viewing it. Obviously, however you view it, you view it the same for both Android and iOS users.

            Let’s agree to discount FB on both Android and iOS. You’ll still find there is a lot of iOS engagement left to measure, and a lot of useful, offline iOS engagement that is harder to measure.

            Either way, according to Ben:
            “When you dig really hard into the Android platform data, you could argue those who truly use the platform and are engaged is around the same size of the iOS base. So I could make a pretty compelling argument that while Android has the dominant share… it is not the dominant computing platform….”

            You are saying your engagement with your Android phone is high despite what the data shows, because you are using it a lot for offline apps. But Ben specifically linked “data” and “those who truly use the platform and are engaged.”

            Anyway, there are other things besides FB that have an online component (and are therefore measurable by the data gatherers in some way). I might use Dropbox, or Evernote, or Office 365, or any number of things that may be offline at times, but sync when I am online. And you might use Google Apps.

            If through data it can be determined the aggregate use of all these things (including Google Apps) is greater on iOS than on Android, despite the larger Android user base, then what kind of definition of engagement are you going to appeal for a “switch” to? That is what several of us are asking you.

            We don’t have to prove the numbers that favour iOS do in fact favour iOS; you have to prove that there is some significant type of “engagement” that is being totally overlooked, but that, if it were considered, would show the opposite of Ben’s assertion: i.e., that Android is the dominant computing platform as well as the one with dominant share.

      2. I raised the AOSP point because
        – the apps are the same (even the small minority that do require Google Services typically use drop-in replacements and looks/feels/gets developed the same);
        – the UI is the same(custom launcher notwithstanding, you can swap that out),
        – the app userbases are conflated (I can Skype from my Google Play ‘droid to my cousin’s “360” droid. I couldn’t play against him until he installed Google’s Play Services though, but those were available in his OEM’s appstore: Android variants are not even mutually exclusive)
        So in the end I’d argue the ecosystem is the same, only with 4 appstores instead of one. I’ve been told translations are more work than writing the 4 variants, alternative appstores have APIs that are really really similar to Google’s for some reason ^^
        So I think the only people who care are the 4 AppStore/ecosystem wardens (Google, TenCent, 360, Baidu). To users and devs, both ecosystems (or the 4 ecosystems) are one and the same.

        I do think Facebook qualifies as an ecosystem. I think this thing needs 2 levels: a “system” level and a “content” level (with apps = content). Otherwise we’ll end up pitting WeChat against Linux and that makes little sense.

  2. I’d be interested to know the sources of the numbers behind these charts, and a discussion of just how much guesswork went into generating them. Installed base numbers, especially, are highly dependant on the platform owner revealing how many active devices exist, and what counts as an active device.

    For instance, Is the decline in the windows PC installed base due to actual junking of PCs, or is it an artifact of people who have PCs but the PC is gathering dust and hardly ever gets turned on anymore?

    Also, now that Apple has said they have hit the 1 billion active devices mark, it should be possible to evaluate just what percentage of devices they have ever sold are still in use — has someone done up a chart on that, and determined the degree of truth of the “iphones have an average service life of 2 years” rule of thumb that i have seen some analysts use?

    1. Not much guess work at all. If you know where to look as I do, there are more than enough data points to easily defend my model.

      The Microsoft PC one, yes, I’m looking more at actives than devices sitting on the desk gathering dust because they aren’t used. The data points I see point to active usage.

      Yes my iOS model includes a continuation of our revised estimates of life cycle. Again places I see provide the data we need plus our market research asking age of device, etc..

      Yes servers would be considered computing devices but I only focus on clients..

    2. The other point I would make is if I just showed you the consumer PC installed base, so take out the ones just used for work or sit on desks at offices of employment, the PC number would be even smaller.

  3. Hi Ben – Great chart … is ChromeOS in other ? Any idea what 2016 volumes were ?

    I may be simplifying but I see apple and google growing with MS in decline ?

  4. I think there is a key development that has allowed the proliferation in OSes in the mobile age… that is Cloud. In a sense, there is still only one dominant platform with the OSes merely providing a door into that platform. Without it, each OS would operate in isolation with little ability to interact across platforms. In that world, you probably would see consolidation as users move to the system with biggest network effects.

  5. Game consoles are multiple platforms. The dominance of Microsoft was more a function of how corporate culture would allow for personal computing to enter their space and not about the technology. Penetration into the corporate world through the blessing of IBM led to Microsoft’s dominance also in the growth of individuals buying PC’s mostly for work related functions. There was really no reason for non-business consumers to own a PC before the commercialization of the internet unless of course they used it as a game machine. People forget how in order to sell PC’s in the non-business non-game space one would have people talking about recipe lists and their book collections databases. Game consoles however continued to exist and given that most cutting edge work in programming often started in the games arena there is no reason to think that VR won’t also be multi-platform technology. I hope we never find ourselves again subjected to the monopoly of a software company like we were for so long under the Microsoft thumb.

Leave a Reply to benbajarin Cancel reply

Your email address will not be published. Required fields are marked *