Horace Dediu

Picking up the discussion of vendor share, I used Canalys estimates of platform shares to visualize the effect of Android on the smartphone market:

Not wishing to belabor the point, but Android and iOS are less than 2 years old as platforms. In that time frame these two johnny-come-latelys have taken 50.2% of the most competitive technology market on the planet.

Here is the share data:

Since the market is not a zero-sum game, I prefer the first chart, but this second view stresses the transition away from incumbent "others" / Windows Mobile toward Android in the licensed OS market.

The final chart shows the competition against smartphone non-consumption. I like this chart best of all as it shows the effect of Android in growing the entire pie for smartphones.

What would have happened without Android is anyone's guess but my bet is that the market would not have grown as quickly. Apple is producing iPhones at capacity and RIM is growing into new markets. Symbian's strongest markets also do not overlap (much) with Android.

However, the trend seems to be for consolidation of platforms (fragmentation within platforms notwithstanding). This trend, however, is over a very short time span. The smartphone industry started a decade ago and the time frame shown in these charts is a mere 3.5 years.

The question remains whether the headroom visible in this chart still allows for new entrants or whether the platforms we see today are the last we'll ever see in mobile computing.

I'm still willing to bet that there is room for innovation and for new platforms. Just like the two newcomers that came to conquer half the market emerged seven years after its inception, new contenders like Windows Phone, Bada, MeeGo, WebOS can still make an impact.

The conquerors came with new business models and a focus on computing not telephony. Can factor innovation and the ubiquity of mobile broadband keep this market open to newcomers?

The best way to get to the essence of any company is by evaluating its priorities. These priorities are like an unwritten constitution. The analog in theology is dogma which when codified becomes doctrine. In law it's common or case law.[1] In business, priorities are hard to discern and are usually only anecdotally observed.

At Apple the top priority is the product.

Sounds trivial, but very few companies place product first. Those who do tend to be producing creative works (e.g. movie or advertising studios, companies built around a creative process). Most companies place either production or distribution first.

Placing product first forces the bizarre behavior that Apple is well known for: being innovative and quixotic. It makes them foolish and hungry. Sometimes it even makes them catastrophically destructive to competitors.

---

Notes:

1. These definitions are from Wikipedia:

Constitution: A set of fundamental principles or established precedents according to which a state or other organization is governed. These rules together make up, i.e. constitute, what the entity is.

Dogma: The established belief or doctrine held by a religion, or by extension by some other group or organization. It is authoritative and not to be disputed, doubted, or diverged from, by the practitioner or believers.

Military Doctrine: an established procedure to a complex operation in warfare. The typical example is tactical doctrine in which a standard set of maneuvers, kinds of troops and weapons are employed as a default approach to a kind of attack.

Common law: also known precedent, is law developed by judges through decisions of courts and similar tribunals rather than through legislative statutes or executive branch action.

The smartphone market has grown threefold in the span of three years. However, as noted previously, the share of units running a licensed OS has not grown dramatically. The following chart shows the vendors' shares with the same brown/green dichotomy between licensed and unlicensed OS's.

In order to gain some insight into this licensed OS area (and who the "others" are) we need some visibility into which OSs are being licensed and what their shares are.

The following chart shows the share of the various licensed platforms[1]. (I blanked out the Integrated vendors.)

Before we dive in to understand the effect of Android, we have to understand the licensing of OSs in 2007.

The chart shows Windows Mobile and "Other" splitting the market. Windows Mobile is rather easy to understand. It was still growing as a platform and Ballmer was famously quoted in 2007 as saying he liked their strategy "a lot".

But "Other" is something of a mystery. What "Other" actually mostly meant was Symbian[2]. But not only the Symbian that Nokia was shipping and licensing to some vendors. At the time, Symbian also had a version called MOAP "Mobile Oriented Applications Platform" which was in use in Japan (somewhat analogous to the variants of Android now in use in China).

Fujitsu, Mitsubishi, Sony Ericsson and Sharp developed phones for NTT DoCoMo, using an interface developed specifically for DoCoMo's FOMA "Freedom of Mobile Access" network brand. This UI platform was based on the UI from earlier Fujitsu FOMA models[3]. The user could not install new C++ applications and therefore these were not platform devices, however the industry still included these units as "smartphones" .

So what we've witnessed in the last two to three years is the replacement of MOAP and Windows Mobile share with Android TAPAS and OMS.[4]

But, more significantly, when looking at the rise of Samsung and Motorola and HTC we were also witnessing the demise of Palm, Fujitsu, Mitsubishi and Sharp (plus a myriad of Windows Mobile licensees even larger than the Android army).

The significance here is that "other" in smartphones is still mostly about branded vendors. This is in stark contrast to the low-end of the phone market where “others” are quickly replacing branded vendors.

So we can conclude that the smartphone market is so far showing a distinct preference for branded vendors. The value of the brand is still evident. But ZTE and Huawei are quickly developing Android competencies. How long will it be before the branded vendors feel the pressure from the upstarts?

----

Notes:

1. The Other category includes WebOS which was not licensed during this time frame but is not material enough in terms of share.
2. A few million units running flavors of Linux were also on the market in 2007/2008.
3. FOMA was also implemented with a Linux OS and the same UI.
4. It should be noted that except for Windows Mobile, all the licensed mobile OSs including LiMO and Symbian are now open source. Android's success is not therefore due to being the only OS that is "open".

The smartphone market grew to about 100 million units last quarter. That's nearly double what it was a year earlier and triple what it was three years earlier, the year the iPhone made its debut.

Few markets grow this quickly, especially as this tripling happened during one of the worst recessions for a century. 100 million units a quarter is not a small number. The rate at which smartphones are growing makes clear the trajectory of where all phones are going.

As I've shown in profitability charts, vendors have been benefiting to differing degrees. The overall smartphone market with individual vendors is shown below:

I've used the color coding of shades of green for vendors which use their own operating systems and shades of brown for vendors which use a licensed operating system (licensees).

What remains interesting is how the share of licensed OS products has changed since 2007. I selected two snapshots in time: Q1 2008 and Q4 2010 and compared the shares that licensees were able to obtain.

The shares of licensees was then 36%, and is now 38%. Not a huge increase, but an increase nevertheless. In the last such analysis in Q3, the share of these vendors (which I also call "Modular") was actually lower than in early 2008.

RIM has held steady in terms of share. HTC has added a few points of share. Apple seems to have gained a healthy share which Nokia seems to have lost. Samsung and Motorola have certainly made a bigger mark in the last few quarters.

But I would caution that share comparisons when the market is growing at nearly 100% are not very meaningful. All vendors are increasing their sales, though a different rates. What would be more meaningful would be a "share of growth" measure where each vendor's increase in units is divided by the overall market's increase.

That analysis gives us the following chart. (I added Q1 2010 for comparison).

Samsung did very well, but surprisingly perhaps, so did Nokia, tying Apple for second place.

The licensees (whose pie slices have been exploded) added up to 52% of total growth (y/y) and the integrated vendors had 48%. A pretty even split.

In a recent post I pointed out that Apple's R&D was about 2.2% of sales in the last quarter. Bernstein took a look at the R&D for Nokia and presented a chart showing the difference between the mobile industry players in terms of total expenditure on R&D.

I took inspiration from that to plot the Devices R&D for both Nokia and Apple over the entire 2010 period. I also compared that with sales and computed the ratio between R&D and sales.

The result is shown in the chart on the left.

Bottom line: Nokia spent 10.2% of phone sales in 2010 on phone R&D while Apple spent 2.5%.

Bernstein goes on to argue that at least for Devices,

Nokia spent $3.9bn in R&D in 2010, almost 3x the average of its peers, 31% of the industry's R&D total spending, for an output that we can qualify as visibly disappointing.

To relate the $3.9 billion for Devices into head count, they estimate that Symbian projects employ 6,200 people; MeeGo and Qt 1,800; Services 1,800; and S40 1,800. Hardware headcount is assumed to be 4,700 and 900 more for fundamental research.

So Nokia's total software headcount adds up to 11,600 people. Nokia smartphone headcount adds up to about 8,000.

Applying a similar formula ($240k/employee[1]) to Apple's estimated iPhone R&D (from Bernstein's chart) yields a headcount of about 3,200 and a total Apple company R&D headcount of 8,200.

So Nokia employs about the same number of engineers[2] for its smartphone software platforms as Apple does for all its product lines[3].

In fact, Symbian alone may cost twice as much to develop than the iPhone (including the hardware).

---

Notes:

1. The cost per employee is obviously different between the companies as they operate in different countries. However there are enough sources of error in these estimates that they probably overwhelm the cost of living/tax differentials.
2. I know that engineers are not all that R&D headcount covers, but let's assume that the mix of overhead is similar for the two companies.
3. The total for Symbian + MeeGo is 8000 and the total for Apple overall is 8,200, however one can comfortably allocate 200 out of 900 researchers out of Nokia research to their mobile software effort.

People are lining up to call the market for mobile phones. Analysts and amateurs alike are connecting points on charts and predicting with confidence the future of mobile platforms. Consensus is forming that there is no future but a quiescent state. By the acclamation of pundits, the survivors are declared to be iOS and Android. They are also predictably arranged in a way similar to OS X and Windows. End of story.

Except for one thing.

3.5 years ago neither of these platforms existed.  In fact, it was only two and a half years ago, in mid 2008, that one of the finalists even became a platform with the launch of an app store. The other "winner" only launched in a handset later that year and had no significant volumes until a year ago. In other words, these suddenly predictable platforms have been in existence for less than the life span of one device that runs them.

To point out how extraordinary this claim is, if we look back to the history of computing, could we have declared the PC industry future cut and dried in 1978, two years after the founding of Apple? What about 1983, two years after the emergence of the IBM PC? What about 1986, two years after the Mac? Or maybe 1990, after Windows was born? Maybe the writing was on the wall in 1997, two years after Windows 95 caught up with the Mac's UX. That would be 21 years after Apple's founding. Some would argue that even today the story of desktop or portable computing is far from over.

Sure, times have changed. Things move much faster now. But markets are also much bigger and take a long time to penetrate. Mobile channels are bounded by provincialism and parochialism.

What's more, it's not like there haven't been platforms before these two upstarts came along. I've seen 9 mobile platforms in 9 years (PalmOS, Windows Mobile, Symbian, Java, WebOS, Windows Phone, iOS, Android, RIM Blackberry). New platforms are still emerging and evolving (e.g. MeeGo, Bada, LiMo, QNX and Tapas).

So the existential question is this: If you were to draw a time line, would a mark in early 2011 signify the end of mobile platform history? The point in time when unforeseeable change will cease to happen?

Let me offer a hint as to why not: every major platform battle has been sparked by an innovation built around a new input method. The keyboard integrated into a personal computer in the 1970s, the mouse in the 1980s, the stylus in the 1990s and the fingertip in the 2000s. Each of these input methods created new platforms (DOS/Apple ][, MacOS/Windows, PalmOS/Windows CE, iOS/Android), new ecosystems, new industries, new competitors and new platform battles designed to enrich forecasting pundits. I won't even try to go back to the changed in the half century of computing prior to the PC.

So here's my challenge to the prognosticators: If you are willing to draw a line in the calendar now and say that no input methods will ever emerge in mobile computing after this point in time I'll buy the thesis that it's over and we can count our chips.

If you think that out of the hundreds of user experience patents being filed, and hundreds of prototypes being slaved over in labs throughout the world, not a single new product will emerge with a new way of interacting with a computing device, then it's time to move on to something other than what can be declared a commodity market.

But if you believe that, as happened four years ago, a new input method, previously only in the realm of mockups and movie magic, suddenly was shown, in a working product, on stage, then this business still has a chance of staying interesting.

Prices provide accurate, independent signals about where, when, and how to create and deploy value-creating innovations. The mechanism of free markets signals what should be rewarded and what shouldn't. When comparing competitors, prices are the best indicators of differentiable positioning.

However, prices are sometimes anomalous and subject to transient market conditions. It's therefore important to observe pricing over a long time frame.

This chart shows the average selling prices for all phones sold by the eight publicly traded phone vendors (covering approximately 70% of the market) since mid-2007.

There is remarkable stability in the pricing of the competitors. One could argue that only Motorola and LG saw significant swings in price. (Apple's instability in 2007 was due to the revenue sharing deal for the 2G iPhone on AT&T).

Motorola pared down its portfolio (and its market share) and as a result has seen a doubling in ASP. LG had a rapid rise and rapid fall as its feature phone business boomed and busted.

But otherwise, pricing trends are subtle: down from Nokia and Samsung. Slight decrease for RIM and HTC and stable for Apple. In a future post I'll dive into the relationship between pricing power and share.

When I started Asymco.com I did not have a plan. I did not have a policy or vision statement. I wrote what I found interesting. I knew nothing of blogging or of web publishing as a business. Over the following nine months I learned a few things. I also learned a bit about how web publishing (and publishing in general) works.

My observations were brought into clear focus by this brief overview of the core values of web publishing summed up by this leaked document form AOL: "The AOL Way".

AOL tells its editors to decide what topics to cover based on four considerations: traffic potential, revenue potential, edit quality and turn-around time.

AOL asks its editors to decide whether to produce content based on "the profitability consideration."

The documents reveal that AOL is, when the story calls for it, willing to boost traffic by 5 to 10% with search ads and other "paid media."

AOL site leaders are expected to have eight ideas for packages that could generate at least $1 million in revenue on hand at all times.

In-house AOL staffers are expected to write five to 10 stories per day.

AOL knows its sites are too dependent on traffic from AOL.com, and it wants its editors to fix the problem by posting more frequently, with more emphasis on getting pageviews.

This codification of values has inspired me to put in writing my own priorities for Asymco:

• Learn by writing. Teach by listening.
• Improve. Move the intellectual ball forward.
• Illuminate topics which are bereft of analysis.
• Be notable. "The proliferation consideration." How likely is the idea to being widely re-published?
• Review. Encourage participation by reading all comments and reply to as many as possible. Police comments with zero tolerance.
• Repair. Declare and correct errors.
• Select. Publish only when the contribution is unique. Avoid redundancy, clutter and noise. Don't waste reader time.

Many have asked: What about the business model? I'm afraid there isn't one. I'm still naive enough to think that if I build a great product then everything else will take care of itself.

Today's charts show the amount of profit captured by the top eight public mobile phone companies in two different ways.

The first, as a bar chart:

The second shows the way they captured it by taking into consideration the number of devices sold.

Please click on the image for a full size view (1024 pixels wide).

Generally speaking, vendors which are focused on smartphones have higher pricing and higher profitability (pricing discussion will follow at a later time). However, some, like Motorola and Sony Ericsson are still trying to recover from the collapse of their non-smartphone business and have higher prices but not very high margins overall. LG does not seem far behind.

Samsung and Nokia profit per device is quite small but they "make it up in volume".

The areas above represent operating profits and are comparable. One wonders how the ratios between horizontal and vertical dimensions for these boxes will change over time.

Since the overall market will expand, the sum of all horizontal dimensions will increase. As the first chart shows, there is no law for the conservation of profits in the industry, so it's possible that the sum of all these areas will increase, but it's still hard to imagine it increasing by an order of magnitude.

Will the industry remain clustered into a tall-and-narrow group and a short-but-wide group? That's a question with many strategic implications..

In a recent tally of mobile OS market share estimates from Canalys, the market was organized by OS vendors. The number of devices sold by operating system vendor includes as many OS's as each vendor may supply.

For example, Microsoft licensed two different OS's during the period: Windows Mobile 6.5 and Windows Phone 7. Nokia also offered multiple versions of Symbian, which have varying degrees of inter-operability.

Google's offerings include (as per footnote) OMS and Tapas platforms in addition to Android.

Fragmentation is not a new subject as OS licensing models have been ripe with it for as long as they existed.

What I want to draw attention to is the concept of Google as "vendor."

A vendor in the traditional supply chain model is a supplier that sells to an up-stream value-adding stage of the value chain. The word itself has as its root the word "vend" which means to sell.

But we know Google does not sell Android. And it's not just that Google builds a product and ships it for free.

There are no contracts associated with the transaction of downloading the Android source. No contract means no obligations, no warranties, no terms and no conditions. The product is not indemnified against litigation. There are no licenses, no support, no copyrights and no patent license granted.

The contract-less nature of Android makes it easy to distribute. It's one of the key reasons why distribution can grow at 600% a year--there is no distribution in the classic sense. The real disruptive advantage for Android lies in the removal of distribution costs (not just development costs). Google themselves admit that they have no idea who implements the software and how it's used.

Which makes the classification of Google as a "vendor" a misnomer. The notion of 'vending' something implies a contract. What Google does is write code and place it on an FTP server.

It's a wonderful way to save on SG&A (there is no sales, general or administrative expense.) But as everyone knows, there is no such thing as a free lunch.

By avoiding any contractual engagement with the authors of the source code, implementors are subject to independent litigation for IP violations. This is why HTC and Motorola are targets for Apple and not Google.  Google is no more liable for writing Android than the author of a book on bomb making is liable for the bombs put into use [1].

Patent violation rests with implementation, not the writing of the idea itself.

Furthermore, Google cannot ensure Android users have any specific experience. They can't encourage upgrades, service models or any form of common UI. The experience, like the protection of property is the responsibility of the implementor.

Can Google change this approach with Android? Not as long as Android is open source. Can Google change Android from being open source? No, they don't own the source to Android. Nobody does.

---

Notes:

1. If the bomb making book was a copy of another, then the author is liable, which is the case being made by Oracle.

The mobile phone market is growing rapidly. It grew 18% in the last quarter. The smartphone market grew even faster, about 75%.

To grow share in this market means growing even faster than the market. What we can see from the following charts is how the iPhone as a product grew in share since it was launched 3.5 years ago:

Unit share growth slowed in the last quarter but the fourth quarter is not usually the most expansive for Apple. That is usually the third quarter which is the first full-availability launch quarter.

The iPhone ended the quarter with 17.25% smartphone share and 4.2% phone share. Share of revenues was about 22% and share of earnings was about 51%.

I still hold that 20% smartphone share is possible for the iPhone. As the smartphone market slowly becomes the entire phone market (see chart to the left) that share will be worth something.