Distilling market noise into market sense

VisionMobile is the leading research company in the app economy. Our Developer Economics research program tracks developer experiences across platforms, revenues, apps, tools, APIs, segments and regions, via the largest, most global developer surveys.

The Dead Platform Graveyard: Lessons Learned

[Besides the iOS and Android platforms grabbing the industry headlines, there is an abundance of (over 25) platforms that didn’t make it. Managing Director Andreas Constantinou recounts the graveyard of dead platforms and exactly what it takes to build a successful platform today.]

VisionMobile - Dead mobile platforms

2011 turned out to be open hunting season for mobile platforms, with the MeeGo, webOS and LiMo projects coming to an end.

MeeGo, webOS and LiMo, together with Windows Mobile and Symbian are just the tip of the dead OS iceberg. The last 10 years have seen numerous companies launch operating systems or platforms for mobile devices, most of which have been fallen under the media radar.

A brief history of dead platforms

The table below lists all known mobile platforms that have died or are a ‘zombie’ (semi-dead) state – that’s all 26 of them, from Access Linux Platform to Windows Mobile. We‘ve also researched the birth and death date of each platform.

Most of these platforms have been designed as software platforms, that is, aimed at reducing costs and time-to-market for handset makers, aka OEMs. Most of the platforms were provided under a software licensing model were monetized via add-on services (e.g. IXI and Danger) or kept for in-house use (e.g. Nokia GEOS). Only post-2007 did we see applications platforms, i.e. those designed to target primarily developers and offered under a zero-royalty model. For the differences between software and applications platform see our earlier post on Platforms 101 and why not all mobile platforms are created equal.

Why are 25+ platforms dead?

In the last decade, software platforms have failed for a combination of reasons.

Cost of ownership. The cost of creating a mobile software platform should not be underestimated. Symbian was quoted as having cost over $700 million of development cost. Even for lighter platforms, a vendor is looking at a ballpark of $100 million cost over 2-3 years of initial development, plus the incremental integration cost with each new hardware platform and the long-term R&D cost to maintain the platform to a competitive level.

Conflicting revenue model. Prior to the zero-royalty model introduced by Android, all software platforms were monetized through per-unit licensing in the order of $5 to $15 per unit. This obviously represented significant costs for the OEM and also competed with bundled (free) software stacks from chipset vendors like Texas Instruments’ BMI, Qualcomm BREW, Mediatek (HOpen) and Infineon RedArrow. That was of course before the mass arrival of smartphones and the abandonment of the royalty model.

Lack of network effects. Even though Microsoft had pioneered the two-sided software platform strategy with Windows since 1995, the benefits of network effects in mobile platforms were not properly understood until the launch of the Apple App Store in 2008. It was Apple that proved how network effects – the positive feedback loop between app developers and users – can lead to enormous demand-side economies of scale. It was the power of well-oiled network effects that made Nokia realize that “it had to go to developers” (and not wait for developers to go to Nokia) before eventually losing the Symbian battle against Android and iOS.

High adoption barriers. For a handset maker, adopting a new platform is a painstaking, multi-year process. HTC is rumored to have been working with Android since 2005 and with Windows Mobile since 2000, 2-3 years before it produced the first G1 and SPV models, respectively.  In addition, handset makers are very risk-averse (they have tough customer commitments to keep up to) and so have in most cases preferred to stick with their internal spaghetti platforms rather than take the risk of adopting a new one.

The ingredients of a successful platform

There are a handful of remaining software platforms today. Besides the usual suspects Android, iOS, Windows Phone and Bada, we should also consider BREW MP (still surviving), Trolltech’s Qt (an API framework acquired by Nokia in 2008 and rumoured to be soon reappearing on Nokia’s Series 40 handsets) and Smarterphone, a niche ‘smart’ operating system for feature phones recently acquired by Nokia.

The chart above makes it clear what is the success factor of modern platforms. Firstly, software DNA, that is a company with resources, processes and values routed in the PC or Internet world where developers, not OEMs are the platform’s primary customer. Secondly, a successful platform vendor needs to have large pockets due to the billions of dollars in investment needed to build a stable and advanced software foundation, while attracting developers to the platform. Note that the bubble size in the chart shows last relative size of 4 quarters of vendor revenues.

But the secret sauce is neither in DNA or money; it’s hidden in Stephen Elop’s famous burning platform memo: ”Our competitors aren’t taking our market share with devices; they are taking our market share with an entire ecosystem”

The secret sauce behind the success of iOS and Android is how thanks to network effects (the closed loop driving users to developers and developers to users) platforms have managed to generate billions of external investment, both in the form of developer investments (time/effort) and operator investments (subsidies).

It’s network effects that have created near-insurmountable barriers to entry for Microsoft who despite boasting 75,000 WP7 developers achieved only one million sales of its flagship Lumia model from its strategic partner Nokia.

And whatever Bada, Tizen or any other alphabet-soup-chef tries to conjure up, they should never forget that you can’t buy developer love. You can only plant the seeds. That’s why Facebook Platform is following exactly the right strategy: take a vibrant developer community and offer it a new addressable market.

– Andreas
follow me on Twitter for more: @andreascon

  • Ben Hookway

    Excellent analysis Andreas. A few names there from feature phone world that I'd forgotten about.

    Being now active in the OTT/SmartTV space, it always amazes me how noting is ever new. The TV industry is going through exactly the same process again. Which platform will we demand of our TV's in years?

    • Thanks Ben. BTW how much overlap is there between TV platforms and mobile platforms beyond Android and iOS?

      • Ben Hookway

        Sorry for the late reply. I would say almost zero. Think of the SmartTV space as feature phone world five or so years ago but with 10x less volume. Everyone has their own SDK (more than one in some cases) for their platform which means there must be 12+ different development environments out there.

  • I couldn't help wondering how you'd classify RIM and BlackBerry OS 1-7, OS 10. It's clear we certainly don't need a new mobile OS, but are their pockets deep enough? Does anybody care?

  • A long view. Which points out that even today's giants will one day be dead dots on your chart. Is Android fragmentation a path to the disappearance of Android as a 'main trunk' or a path to its dominance?

  • Dave Balmer

    Um, one correction. webOS is not dead, but it is going open source. HP just announced the roadmap for 2012 last week and due to the "fire sale" last August, has a decent installed base for tablet. Besides, the entire premise of webOS is using web standards to build apps, which is perfectly in line with "the web". Also, you have its "birth" in 1999? Are you confusing with something else?

    • Sorry for the typo. WebOS birth date should have been 2009. We would class it as a zombie or orphaned platform. There are 200K+ open source projects out there, but those that survive are those that have commercial sponsors supporting them, whose core business is in someway enhanced through that software. Given that HP has given up on WebOS and 'donated' the software, there needs to be another sponsor that revives the community interest and productises WebOS.

      • I agree Andreas. By not announcing new webOS devices and simply open sourcing the platform, Meg Whitman, like Pontius Pilate, "washed her hands" from the webOS debacle. It's a way of killing the platform without directly taking credit for the decision to kill it.

  • Jonas Hermansson

    Hmm I would not attribute iOS success to Apple's focus on developers. Their focus on solving real end user problems is (in my opionion) more important for their success. The promise of the iPhone was more "internet in your pocket" to begin with. Apps came later and are now, as you describe, very important.

    • Well, that's the fundamental difference between Apple and Nokia. Both had in-house OSes, good industrial design and strong supply-side economies of scale. Apple realised that developers were going to be the primary driving force for growth long term, but Nokia didn't, at least not early enough to be able to turn the boat around.

      • What amazes me about Nokia is that they clung on to Symbian which was not really an easy platform for developers to work with and even more difficult to monetize. History has repeated itself just like the battle between Microsoft Windows 3.0 and Geoworks. Geoworks worked on 80286, took only 1K of memory and was integrated with AOL to make it one click to get to the Internet. But Geoworks went the way of the dodo because it did not support an external development community, Microsoft did. Apple has made it easier, but Google has beat Apple at its own game which is why the adoption rate has been much faster for Android, but the difference is that Apple has a closed eco-system and a very loyal fan base. Android, not so much and Microsoft burnt their developer base with Mobile 5 & 6 so bad, that most have not returned or recovered.

      • Sander van der Wal

        Nokia had given up on developers a year or so before the first iPhone, after years and years of trying to make apps a succes. And the consensus at that time was that there was no money in native apps. What Apple did different was to go all the way to make things perfect. Just compare Nokia Software Market to the App Store.

  • Kyle

    'Dead Platform Graveyard' is redundant. It's Dead Platforms or a Platform Graveyard.

  • Dev Khare

    The mobile web is a platform by your definition even though it is not owned by anybody. It has:
    – low adoption barriers
    – big network effects
    – zero cost of ownership
    – biggest ecosystem
    – (but) conflicting revenue model

    • JrSnow


    • Well.. mobile web is a technology (HTML, JavaScript, CSS) and a set of browsers, but no distribution (major app store), monetisation (beyond ads) or retailing (storefront) element.

  • Guilhem Ensuque

    Great piece Andreas,

    I would add a big driver for why OEM-focused platforms have died: industry consolidation

    Flash-back to circa 2004 when the likes of Sendo, Siemens … and many others were still around. Each shared a small piece of the overall feature phone market pie. In that context, the variable costs of licensing a platform on a royalty-based basis from a 3rd-party were more favorable than the fixed costs of building your own platform in-house (except for Nokia with S40 at the time).

    Fast forward to today, all small phone OEMs have disappeared leaving only a couple juggernauts: Nokia, Samsung (and to a lesser extent LG) with a large share of the pie each (+ a myriad of Asian OEMs relying on Mediatek). In this context, the fixed cost of having an internal platform R&D makes more sense from a financial point of view.

  • Andreas,

    Great insights. I think that revenue model is moving from royalty to patents – these OS's are definitely not free.
    * Google may give Android for free, but then Microsoft comes and scoops up their share.
    * iOS costs are simply wrapped in the device's cost, as it is an inhouse platform.
    * Windows Phone is probably going to be based on royalty costs, and I assume these won't be that much different than the patent fees they take from Android phone vendors.

    • Hi Tsahi – I think that patents in the anti-Android camp are not used in lieu of royalty. They are used as a way to either stop shipment of Android handsets (Apple) or to lower the OEM switching cost to Windows Phone 7 (Microsoft). Otherwise, today in mobile asin desktop, software platforms are complements that are used to drive the core business of the platform owner – ads (Google) or consumer electronics (Apple).

  • Guest

    Hi folks,
    I don't agree to the analysis, because the influence of cost on the device makers seem underestimated to me. The real story is that smartphones have been invented to support the major business in the Internet, which is advertising. So the cost for Android is no longer paid through royalties but through "ads by Google". So technology to get better profiles of the users (e.g. by gathering GPS data) and to have better (adblock free) APIs to submit ads to the user are the key success factors for a smartphone OS. So we all pay the cost for Android in the supermarket or our favorate pizzeria. It is no longer paid by HTC or Samsung.

    • Agreed – the business model for Android is to drive Android's core business which is driving eyeballs to Google properties.

  • marees

    hey, how come Maemo is not mentioned (either as living or dead!) ? N9(Harmattan) was actually QT on top of Maemo and not at all Meego. That is why Elop pronounced meego as dead. Nokia has never released a single proper meego phone. It has all been only maemo phones

    • Good point. We 'll add Maemo as "zombie" in the next revision.

  • Aron

    No Qtopia?

    • Very good point. Qtopia was the apps suite that died soon after the Nokia Trolltech acquisition. We 'll add in the next revision of the chart.

  • I think it would be more useful (perhaps in a future post) to distill the meaning of "Software DNA." It is too broad a term and hides the key differences. Nokia were supporting developers for years before iPhone. Nokia had lots of software DNA, as did companies like Motorola. I would say that design competence is a key differentiator and factor here, as is focus. Apple had only ONE device. It didn't support developers from the get-go – we were given the Job's "Browser apps" vision. Many would say that the iOS APIs were clearly an afterthought. Above all else, they focussed on design and user experience. They were vehemently focussed on the end-user, not carriers, as customers.

    • Hi Paul – DNA is made up of resources, processes and values (as per Christensen's RPV framework). Nokia was supporting developers since the early Symbian days but was lacking processes and values that were closer to a software company. Motorola was much the same, with emphasis being on pre-load processes (typical of an OEM), not post-launch processes (typical of a software company).

  • Good article, except that it's not true that "network effects in mobile platforms were not properly understood until the launch of the Apple App Store". It was just that the importance of keeping development experience optimal was not understood enough before Apple.

  • Wonder if RIM (BlackBerry OS and QNX) should be on the "Software DNA" chart? Still alive and kicking with some market share. Also wonder how come iOS and Windows Phone bubbles are same size, when chart legend says "bubble size = relative vendor revenues for last 4 quarters"? Guess the keyword is "relative" and how to interpret that 🙂

  • What will be interesting to watch will be how OEMs look to leverage the Android ecosystem outside the control of Google. RIM is doing it and I think there will be others that will implement the Delvik Virtual Machines outside the entire Android stack. This strategy will allow OEMs the ability to control the experience on their devices wile running Android apps.

  • As this is a subject which is somewhat dear to my hear I would like to offer a small correction on the subject of "Nokia GEOS" (which originally got me from PCs into mobile development):

    While it is true that Nokia launched the Nokia 9000 Communicator in 1996 using the GEOS operating system, the platform actually had a Desktop heritage, being a variant of the PC/GEOS operating system developed by a company called Geoworks.

    More details here: http://en.wikipedia.org/wiki/GEOS_%2816-bit_opera

    In a sense, the fate of PC/Geos is in itself an interesting case study, because it was an operating system that started out with an amazing ability to perform well on low-resourced (1MB, 16-bit, Intel 8086-class) devices of the early 90's, using an idiosyncratic C dialect closer to Objective-C than to C++, but then failed to "grow up" to other platforms when ARM CPUs and protected mode x86 architectures started to rise at the end of the decade.

    At the same time, it is remarkable how Nokia failed to recognize the extremely flexible user interface technology in GEOS with the ability of a "descriptive" user interface to adapt to a wide range of screen sizes, themes and even input modalities. Instead, the Communicator used a tailor-made software layer on top of the original platform that locked down much of that flexibility, in a way that remarkably mirrors how the Avkon layer in Geos took away much of the original design in the Eik class hierarchy.

    So in a sense I am now watching the end of the Symbian saga as "2000 redux", when Nokia already once swapped Smartphone operating systems rather abruptly. On the other hand, with he attempts of Intel to get back into mobile devices with its Atoms, the pendulum now seems to swing back towards x86, and Objective-C is rivaling C++ for mobile development.

    Enough nostalgia for today. 🙂

    • Small correction: I meant "how the Avkon layer in EPOC took away" – otherwise the comparison to the Foam "wrapper" on GEOS makes no sense. 🙂

      One more detail – GEOS on Nokia phones was in fact an open operating system in the sense that an SDK existed from the very start, though at quite a high price at first, and only gradually being reduced to being effectively "free". The only thing that still needed to be purchased was the mandatory Borland C++ compiler, which ended up being really hard to get near the end of the Nokia 9110 lifecycle.

  • jon bostrom

    You seem to have forgotten JME the largest installed base of mobile devices in the world
    Jon Bostrom -father of mobile Java

  • Michal

    You coud write somewhere a shorter url obtained from bit.ly so I could write it into my Bookeen reader web browser.

  • The main thing that Apple got right IMHO was not the network affect but the fact each phone HAD to have a data plan with the phone. Nokia really dropped the ball on this one.

    With a data plan came the connection to the app store and the internet and this made a market where there was none before.

    I see it highly unlikely that Microsoft will survive in the smartphone market due to the lack of a unique selling point, at a minimum they should be targeting Enterprises very heavily as this is largely an untapped market. Even worse, Webkit is the defacto browser on mobiles and this omission will leave Microsoft as a niche player in the smartphone world where the web apps work on all major os's (Android, iOS, BB, Bada) but not/limited on WP

  • DesDizzy

    You seem to have left out the Newton OS from Apple

  • DesDizzy

    Sorry I forgot to add that I believe Apple also were start-up owners of Symbian

  • Dwarak

    Nice composition Andreas, according to me even the Blackberry may join this, if it continues with same strategy. In recent survey, most of the Blackberry developer has shifted to Android platform, as there is no requirement for blackberry app development.

    About Symbian with recent release of Anna and belle, there is still hope for survival.

  • Chris

    Hi guys. What about Sony-Ericsson? What OS did they run on the K750i, and is it dead? I reckon it must be dead because Sony stopped supporting it many years ago, but I stand to be corrected. What OS do they run on their current phones, and what chance has it got of surviving for, say, the next 10 years?
    Regards, Chris

  • Andreas

    Chris, I believe SE was using Symbian for it’s high-end devices and Enea OSE for its mid-end devices. Today OSE is still used in embedded apps.

    SE was one of the first OEMs to put its bets on Android – and yes, Android (specifically AOSP) will be around for the next 5 years, or until the basis of competition for mobile platforms changes.


Distilling market noise into market sense

What gets desktop developers out of bed in the morning?

desktop developer segmentation

Despite all the hype around the death of the PC and the enormous amount of media attention focused on mobile,…

Continue reading ...

A proven model for targeting IoT developers


  What if you could identify a handful of developer personas, or segments, each with a very distinct set of…

Continue reading ...

1000 skills: Amazon Alexa as a metaphor for the IoT developer community


Alexa is a centerpiece of Amazon’s Smart Home push, and quickly growing to become one of the most promising attempts…

Continue reading ...


Research on the app economy and developer ecosystems

Developer Segmentation 2014


The Developer Segmentation Q3 2014 report is the most sophisticated study of developer segments to date. The report delivers a…

Continue reading ...

App Profits and Costs


This research report examines the critical success factors for a profitable app, and how business and technology choices, such as…

Continue reading ...

Developer Segmentation 2013

Developer Segmentation 2013

The Developer Segmentation 2013 report delivers a needs-based segmentation model that actually works, with extensive profiling of the eight principle…

Continue reading ...