The interest of Windows 10 lies not in its different icons or the return of the “start” menu. Its deliberately slow launch has implications for everyone interested in organisational development and change management.
I’m not about to review Windows 10: there’s plenty to read about it elsewhere. I’ve been using it for months as a technical preview and I’m pleased that it’s now very stable. In the meantime it’s that “preview” process I want to comment on.
Windows 8 was a misstep, partly because the design was badly executed, partly because the future it anticipated hasn’t exactly come to pass. Windows 8 reflected a widespread belief that everything was going to move to touch interfaces, but here in 2015 it seems that tablet sales have peaked and there’s revived interest in desktop machines and laptops, or specifically hybrid machines combining touch and more traditional, mousey ways of working.
Windows 10 sensibly hedges its bets in this respect, and Microsoft has made it clear that while its operating system is no longer a “technical preview” it’s still very much a work in progress. It’s said there may be no more big milestone releases, but rather endless incremental refinement and improvement.
In doing this Microsoft is turning something fundamentally unsatisfactory about software on its head. It’s often said that when you buy a physical product like a car you expect it to work properly from the beginning rather than needing constant bug fixes. Software engineers it seemed couldn’t deliver that level of finish. There were and are many reasons, some of them good ones, and the story of Windows 8’s failure suggests it really was time to think about things differently.
IT has matured enormously in the last ten years, to the point where software business models that relied on regular paid updates are no longer viable; it’s become difficult for mainstream products to add compelling new features. Despite this maturity, other factors mean user behaviours and preferences are as unpredictable as ever. Indeed, experienced designers will always concede that you cannot anticipate exactly how users will shift their behaviour around a new piece of tech, or how technology and economics may collide to produce quite new behaviours (for example SMS started as a signalling technology for telecoms engineers).
Windows 10 has been set up in a way which acknowledges the continuing unpredictability of this environment. This is partly why it’s important that it can work across all computing areas (Microsoft’s weakness in phones remains an Achilles heel), partly why it has promised to maintain its very smart “Windows Insider” programme of millions of volunteer testers, willing to trade a bit of stability for early access to new features and refinements. The point is that the company no longer has to make big bets on its operating system design. It can try things out, see how people react, and develop its thinking from that reaction.
It’s possible that you can only be this blatantly experimental if you’re not charging people for the ride, and indeed for the moment Windows 10 is going to be freely available to most of its users. Its hand has been partly forced by its arch rival Apple, because Apple no longer charges for its OS upgrades (though Apple unlike Microsoft can rely on revenue from its hardware sales, as well as iTunes and its AppStore).
But there’s a bigger commercial and management point. An operating system is an extremely complex project. It’s going to have to work in all sorts of environments, in ways which its designers are unlikely to be able to anticipate. Addressing this unpredictability is explicitly part of Windows 10’s ambition. In the early days of the mass computing revolution it was understandable that developers should offer only what they could do, and demand that users adapt their ways of working around those limitations. But it’s not a great way of doing things, and in this sense Windows 10 marks a turning point.
This turning point goes deep. It’s generally true that most big IT projects fail. It’s also well understood that one of the fundamental reasons for this failure lies in what the sector calls “requirements capture”. Requirements capture is supposed to be the process that bridges the gulf between what people do or need, and what the system can be designed to deliver. This “capture” is generally still treated as a definitive process at the beginning of a project, setting its goals and enabling its planning. The problem is that most people are not very good at articulating what they do or want, because a lot of what they do is only tacitly understood. It’s only when they are confronted with a system that’s trying to force them to do something differently that they become aware of the importance of how they were doing things already.
The irony is that IT itself could and should have offered a way forward. Fifteen years ago I met Michael Schrage at MIT, who was talking even then about the power of rapid prototyping as a way of testing and refining innovation, and helping to ensure its acceptance (he wrote a book called Serious Play which explored these ideas in more detail). This was software prototyping, using spreadsheets and simulation to do things quickly and cheaply. It demands a different, iterative approach to system design and requirements capture, constantly testing and retesting ideas against the real world. Such an approach introduces complexity and is likely to be harder to manage, but then these things are never going to be as expensive as system failure.
I’m not suggesting that all software needs to be developed this way. For finite, well-defined tasks aimed at a known audience it might always be possible to say “here’s a different way of doing things and we think you’ll find it’s better”. Even here you’ll need to introduce your innovation properly, make sure it’s well-understood, its benefits clear, with adequate help and training on hand (you can assume as a default that people will resist change).
But for more ambitious projects that reach across different aspects of people’s working lives (or the totality of their lives), a modest and open approach is likely to yield dividends. It’s also likely to demand the recognition that managing change is not about going from A to B: it’s about developing a structural responsiveness to a constantly evolving environment.
In the meantime, I certainly won’t be going back to Windows 8, or 7. It has rough edges but Windows 10 is working pretty well. Now I’m just curious to see what happens next, and that’s a good place to be.