One day we'll look back and say this was the end of the software platform
Ballmer has scribbled the epitaph of his own era
Comment I was once told by an editor not to waste too much time following Apple and other companies that sold hardware and software together.
He was reflecting the conventional wisdom in the late 1990s: any manufacturer that dared to "add value" by integrating its software into its hardware would be destroyed. Microsoft and Dell would carve up every market between them. It was as inevitable as the weather.
So much for conventional wisdom.
The truly astonishing thing about Microsoft’s acquisition of Nokia’s mobile phone business is the era of platform software licensing is over. It’s the end of an era, a model, even a business philosophy. And it was a fairly short era in which platform software could be commercially licensed, albeit for pennies but on an industry-wide scale. Microsoft is buying its way to becoming a "devices and services" company just like IBM, and today even Oracle makes hardware. What was recently considered a deadweight on a software company's business is now mandatory.
It’s worth remembering how short this software era has been. Commercially packaged software didn’t really exist until the late 1970s, but long before then business regulators were eyeing the integration of mechanical services and mechanical hardware.
Arriving at IBM in 1914, Thomas J Watson “had instilled a fundamental set of business practices … aimed at providing comprehensive services obtained from tightly integrated systems,” notes (pdf) technology historian Steve Usselman his history "Unbundling IBM". Can you distinguish the difference between that and Microsoft’s explanation yesterday?
According to Redmond:
A family of devices with integrated services that best empowers people and businesses … Devices help services and services help devices …
Within twenty years, Usselman recounts, IBM had suffered its first anti-competition judgement: in 1936 judges upheld a complaint that customers who leased IBM data processing machines needed to buy IBM punch cards. Bundling was considered poor form even in the pre-electronic era. And IBM responded with an argument Apple uses over its App Store today: it is not obliged to provide an open market, but wants to maintain quality so its own hardware functions well. Bundling wasn’t resolved and litigation would continue, more on than off, between the US Government and IBM for the next 50 years, until 1982. Regulators objected to IBM adding services such as programming services without an explicit price tag.
The platform licensing model is typically attributed to Bill Gates but it was actually pioneered by Gary Kildall of Digital Research Inc. He’d made one set of programming interfaces a standard across a bewildering range of microcomputers and licensed it for low cost: appreciating that it was a scale business. CP/M was the first "platform".
Ironically it was IBM who contributed to Gates' rise and Kildall’s demise by promoting a clone of Kildall’s CP/M, DOS, for its first IBM PC, and making CP/M a more expensive optional choice. While DOS was free and bundled in the PC box, CP/M was only available via a $240 voucher. There were no startups, in the modern sense. An IPO of a software company didn’t take place until 1978, and they didn’t pop up on NASDAQ until 1982.
The idea of promoting a standard across systems chimed with economic liberalisation, and the idea that increased competition and new markets would squeeze out costs in the supply chain, making products cheaper, and accelerate innovation. It’s mostly been proved correct – we don’t expect Vodafone to manufacture handsets and we don’t wait six months for a landline to be installed. But the Microsoft notion of applying PC economics to mobile, of merely being a component supplier, just hasn’t worked out.
Today people buy big name brand devices: Apple and Samsung have 90 per cent of the smartphone market. And Samsung’s dominance owes much to a giant advertising company giving away its component for nothing. Google takes a significant loss on Android not because it’s philanthropic, but to destroy other markets, and maintain its ad-brokerage and data trawling business. Microsoft has made more money from Android than Google, but Google doesn’t really care. As a result the quality of Android is exactly what you’d expect from a Java OS rushed to market by a large advertising company. For now, Android doesn’t have to be much better; and only a widespread consumer revolt (which is hard to envisage) will force either Samsung or Google (or both) to improve the rotten UX and performance with well-written native code, and better design.
Android is "good enough".
But now, by acquiring Nokia’s devices unit, Microsoft is licensing a major hardware brand and buying an expert hardware team – and as colleagues have pointed out, it’s doing so for peanuts.
Has Fester just saved Microsoft?
Perhaps it's too early to write a revisionist, positive history of Steve Ballmer as Microsoft CEO - after all, he's still in the hot seat and no successor has been chosen. But he's made an epic choice in determining that the era that made him so rich is now over. For all the mis-steps and goofy quotes, that's quite a brave thing to do.
The question is whether Microsoft can change its spots? It’s a company in transition, as the outgoing Ballmer reminds us at every opportunity, and it has had mixed results at selling consumer hardware. The great success is Xbox, which has been left alone and backed by billions over more than a decade. Being a games device spared Xbox from the dreaded Microsoft "integration" efforts that doomed almost every other promising initiative - it didn’t need to run Outlook. Elsewhere Microsoft has fatally compromised or snuffed out attractive hardware projects.
Microsoft doesn’t have a great record with its acquisitions and while Redmond has a ten year license on Nokia’s great brand and technology team, it’s hard to see Microsoft executing better than a standalone Nokia under Elop, which has been going great guns for the past year. The acquisition has been made to minimise risk – the risk of Nokia defecting from Windows Phone altogether. Redmond would do well to leave the team alone, and simply do what it did with Xbox – keep pumping in billions for the next few years.
Given that Microsoft professes to a curiously modest market share target of 15 per cent by 2018, it could be that's just what will happen. ®