Blockbuster book lays out the first 20 years of the Smartphone Wars
Symbian's David Wood bares all. Not for the faint hearted
Review David Wood, one of the founder executives of Symbian - and the one who saw it through to the bitter end - has written a book. A very big book.
Smartphones and beyond: Lessons from the remarkable rise and fall of Symbian tells the entire story from Symbian's conception, to world domination, to its rapid demise, and it must be one of the most candid and revealing books a technology executive has ever written.
It's currently No.1 in Amazon's mobile and wireless section.
Of course, it's popularly considered to be a story of failure. Although Symbian was in hundreds of millions phones, and for years powered the most-bleeding edge mobile tech, Android today is everything Symbian set out to be: creating a rich platform for modern smartphones and tablets, on which other industries built their services. And that's what makes this story interesting - much more interesting than if it had been a roaring success. The mobile industry today today is defined by what Symbian wasn't - or by what it failed to do well. For example, you'll have a much better understanding of Samsung from reading this book.
Although Americans today have a very different history of smartphones, in which Symbian doesn't figure, for the rest of the world it looked very different. Symbian dominated it for a very long time, either implicitly (no mass market phone shipped until four years after the venture's foundation) or in reality. With Nokia putting huge resources behind Series 60, Symbian's share of the smartphone business rarely dipped below 60 per cent. It was practically the only game in town for creating rich devices for many years - without it ever being a consumer brand, as Android is today.
The the sheer volume of detail in David Wood's book might overwhelm the casual reader. It's a linear narrative drawing on diary entries, business plans, memos, emails, and presentations; entire press releases and news stories are quoted (apparently without permission - naughty). The story is condensed into bullet points on almost every page. It needs an index to the hundreds of codenames (Symbian never refer to licensees or partners by their real name - in a kind of weird Secret Seven code). And some sections could have been radically condensed or snipped entirely without losing anything. It comes to well over 300,000 words, or 800 pages in regular typesetting (it's only an eBook right now).
But Wood reflects on many key decisions with completely fascinating digressions on counterfactual histories. "Could we have made a better decision - or done things differently?", he often asks. These counterfactuals probably help give us the most comprehensive account of the real dilemmas players in the mobile industry faced over the past two decades anyone has put to (e-) paper, as they pick out trends that are important but not obvious today. And Wood is so brutally candid and honest, nothing is spared.
Reading it, I was often left wondering how Symbian produced anything at all - let alone survived for so long. Such were the delays, political machinations of its shareholders and licensees, and painful mismatches between them all. Hyped as the most important company in the world - a hype it helped stoke, Wood confesses - it didn't power a mass market phone for four years. It over-promised then retrenched.
Symbian might have been the best choice for phone makers for years - but it could be a very painful choice too. Symbian was unable to deliver platform software on time for years - delays are the recurring motif of the book. It required device manufacturers to undertake painfully expensive integration, because the parts needed to make a Symbian phone didn't fit nicely - there were multiple UIs. The base ports (code for a specific hardware), Wood writes, were immature or buggy. There should have been nice, off-the-shelf reference implementations - Linux at the time had hundreds. All these things are taken for granted by manufacturers using Android or Windows Phone today. Even when Symbian got its defect count down and speeded up its delivery, Nokia couldn't use the platform release, as integrating its own half of the Symbian proposition was too difficult.
This created cost, and for years, the large costs of creating a Symbian smartphone were largely hidden.
Boy, were we wrong: confessions corner
You would expect the former exec of a colossal failure to self-justifying. Not so. In the "phoney war" between 1998 and 2002 when hardly anything shipped from the much-hyped venture, projects were being cancelled left, right and centre. "We had our handy, blame-defecting explanations" Wood writes.
Looking back at emails at another, particularly fraught period when Nokia - which had the biggest bet on Symbian - wanted some clarity from the company, Wood notes:
"I see only an intensely pragmatic attitude from the [Nokia] Calypso team, which elicited only faltering responses from the Symbian side."
Years later, as Symbian struggles with its eternal dilemma of trying to be a predictable software factory and yet be responsive to customers urgent requests for additions, Wood confesses: "Our approach.. almost guarantees that this misalignment will occur."
A cross-department "Matrix Team" approach was tried out, to cut through the bureaucratic turf-wars, but it became a bureaucratic exercise itself - and forgot to include customers.
At one point when Ericsson demands a vital piece of connectivity (OBEX) is supported, Symbian's chief of software development Kent Eriksson fumes: "we must convince them they don't need it. They just cannot come like this outside the requirements when we are already have full speed to implement what has been asked for". [sic]
We're not spared anything in this tale. The Employee Satisfaction results and exit interviews of key employees ("sausage factory"… "authority without responsibility") are related. The move to open source Symbian and unify it with the UI layers eventually takes place, but it's a flop: it only reminds everyone what a politically tense venture it was in the first place. Nokia distances itself from it. Licensees drop it, and developers who never liked it, fail to contribute substantially to the code base.
What comes through is how painfully respectful Symbian was to its partners - how hard it tried to do the right thing. If I had to sum up the Symbian story it would be, "they were always trying to do the right thing for people who were important, but ultimately didn't matter the most". Symbian were visionary and ethical in anticipating the needs of the manufacturers, the networks and the entire industry. But the device owner, the end-user, was removed the Symbian story less three years in, in early 2001. After that, when Symbian withdrew from creating UIs, the company knew what the punter needed, but had lost the power to influence it.
None of this prevented Symbian being a leader for years - although it had taken many painful years to get there. Symbian served a cosy club: in practice, you had to be a huge manufacturer or carrier to make a Symbian phone, as the "crown jewels" of the big phone players were closely guarded secrets - and integrating Symbian was expensive. (This makes the achievement of Sendo - a small British manufacturer that cranked out two Symbian phones - all the more noteable).
For these years, the huge costs of creating Symbian smartphones were largely hidden. Then capitalism intervened.
Things really changed when (as Charles Davies said here) packaged hardware began to appear, allowing newcomers, like Apple, to try their hand.
Apple also set a new standard for user interaction, by targeting a niche: the wealthy consumer, with very lavish phones. I remember Nokians being impressed that the first iPhone had a dedicated GPU - to drive the compositor - and quite flummoxed that the iPhone drew the UI animations at 60 frames per second. Quite unnecessary, the Nokia experts said. (Nokia could have made expensive prestige phones too, Wood points out, but instead targeted volume sales and cut costs, fatally skimping on the hardware).
The Apple effect raised the bar of expectations and Symbian phones couldn't fulfil that demand. Instead, Android did. With the packaged hardware, and higher specs, a much friendlier development environment could be used: Java. And with its clean slate design, Android was a whole world less pain for phone makers to integrate. The Android team could also add new features far quicker than older systems, which Wood describes as having "technological debt". (Future generations pay the bill).
"One of the key reasons Symbian licensees defected in such numbers in subsequent years, from Symbian to Android, is that licensees found it took them much less time and effort to create a smartphone using Android than using Symbian". There were further wrinkles - today, Microsoft makes more from Android via licensing than Google does - but not so many as to make the choice for manufacturers a difficult one.
There was to be no "EasyAPI", and licensees had to wrestle with an API (later called Series 60, or S60) that was hurriedly reverse-engineered from a single phone project, by engineers who'd never written an API before. Nokia management was told by its engineers that it could add Hildon's pen support to its existing S60 APIs in "two weeks" - but it didn't for four years.
Symbian had a decent ride at the top - four years in shipping PDAs, then eleven more years in shipping phones - and what killed it was Android. Java was at last viable on modern hardware, making writing applications much easier. The parsimonious advantage of Symbian no longer counted for as much. And Google offered a turnkey platform - download the code, Wood points out, and two weeks later you had a working prototype phone.
Symbian responded by going open source, a move whose benefits were vastly overrated, Wood concludes, after being one of the key instigators for it.
The What-Ifs make for some fascinating discussions. What if Symbian had broadened its shareholder base at the turn of the century beyond the major Tier 1 manufacturers? What if IBM and Sun had bought in? What if Symbian had then used this capital to fund UI development, and control the whole puzzle? It would not have been beholden to phone manufacturers, and would look much more like Android today. What if Palm had joined, as it once promised to do, and trailblazed bleeding-edge 3G products into the US market? What if Symbian had produced first rate CDMA support? What if Nokia had done all of its software development in the US? What if Nokia had taken over Symbian completely in 2004, as it wanted to? And later, what if Symbian hadn't been open sourced? Fourteen months after the announcement, only 12 per cent of the source code had actually been freed. The world had moved on. And so had Wood.
(Wood would rejoin the former Symbian consulting team, which had been spun out to Accenture, and work there until March 2013).
Wood also has some acute observations on Nokia's fate. He thinks Stephen Elop blew in with good ideas and a good diagnosis of the Nokia malaise. But he over-estimated the maturity of Windows Phone, then "Osborned" Symbian by effectively declaring it a dead-end. Microsoft then added to Nokia's woes with Apollo, the codename for moving Windows Phone to a new kernel and new middleware, which made the first Lumias dead-ends, too - they couldn't be upgraded to Windows Phone 8. That left Nokia with little to sell in 2012 as the smartphone market continued to explode, and Android became more mature and entrenched.
He also notes that Nokia staff found it convenient, after 2009 and the debacle of the buggy, underpowered N97 flagship, to blame Symbian for its own mistakes.
As with most What-Ifs, the dispassionate Wood is sympathetic to all sides. All decisions have their advantages and their costs, and it isn't clear what they are at the time. He even warns against certain kinds of counter-factual history, speculation where one product substitutes for another. We're reminded that Android, for a while, didn't look like the runaway success it became.
Wood also reminds us that even Nokia joined in trashing Symbian's reputation by the end, as kernel developer Dennis May points out, because much of the blame for the mess can be laid at its own dysfunctional software division.
There's little dispute who won, today: Android will ship in around a billion devices this year. Google backed a clever skunkworks with enormous resources, that were beyond Symbian, and didn't burden them with having to make a profit. And it's worth looking at the cost of victory and the failure of Symbian venture. There are huge costs not just to Google, but to the licensees, who pay more in royalties than they did to Symbian when it was a profit-making company. Billions have been accrued by lawyers in fees. Then there's the long-term cost to the consumer, of helping create a global data processing monopoly, and leverages that to destroy other markets.
Nothing comes for free - we can't always count the cost, though. That's a different kind of "debt". If the doomiest scenario comes about, and one company (Google) accrues most of the value of transactions that take place over the network, then our obsession with shiny things today will have a very high cost for future generations.
It's such a fascinating story there's a scope here for a much shorter "civilian edition" to be "productized" from the source material - I hope one appears.
If you can't wait, read our three-part history of Symbian (Part One) describes the formation of the venture; Part Two: the tense early battles that defined the industry, and former CTO Charles Davies' reflections here). There's also a piece on the "great lost platform", Hildon, that could have saved Nokia a lot of future grief. ®