Iterative Design Change

"Linux will never make any meaningful headway into the desktop," argues a recent item in IT paper The Register. I can scarcely think of a duller subject than computer operating systems, but on reflection I thought that this piece and the comments which accompanied it had something interesting to say about product design.

The essence of the argument is that most ordinary computer users simply want to get their job done, are familiar with Windows, and do not care about the sort of performance or security issues which attract techies to alternatives. The discussion which followed it largely divided into those who agreed, and those who felt users could or should be educated into using other systems. I thought one particularly insightful comment was this:

I wonder how many of the problems Microsoft had with Vista and market acceptance are not so much the resource usage (which only really the techies care about, and you probably wouldn't really notice anything in day-to-day usage if you've just bought a new computer three times as powerful as your old one), but the "everyday" users clamouring,

"Aargh, how do I foo in this one!" "I can't find the bar anywhere!" "They've completely changed quux!" &etc. to "I want my XP back!"

Watching non-technical users being exposed to new systems is an education. Techies have a tendency to go, "oh, it's not there any more", grumble a bit about relentless UI fiddling, then go menu spelunking until they've found what they need, no more worries. Normal people, on the other hand, are more apt to freeze, stare blankly at the screen, and murmur, "I want the old version back."

It’s no secret that Windows Vista has performed disappointingly in the market and this could be one reason why. I avoided buying a Vista machine myself because I’d heard horror stories about needing a supercomputer to run it, but also because Windows XP does everything I need it to, and at this point I don’t particularly care if they never release another version ever again. However if Microsoft never tried to introduce innovations then they’d be a considerably less successful business and Windows would still look like this:



It is an issue familiar to most of us who have ever tried to develop a product with an existing customer base. On the one hand, you don’t want to alienate your users by changing things with which they are comfortable. On the other, you presumably have sound business reasons for making a change and users generally find it hard to imagine how things might be better unless you show them.

One neat attempt to address this dilemma was the recent transition to ‘the new Facebook’. At first users saw a link on the existing Facebook homepage inviting them to preview the new design. Then the new design was set as the default, with an option to switch back to the old version. Finally ‘the new Facebook’ became ‘the only Facebook’. This still caused some consternation amongst users – but a few months down the line, who remembers the ‘old’ Facebook?

The idea of design change by gradual evolution is hardly new and has long been used by the likes of Amazon and eBay. Unfortunately it isn’t an option open to Microsoft, they can’t release dozens of new versions of Windows with small changes from one to the next. With web based products there are still times when a total relaunch is called for, but generally speaking the ability to change a little and often is a unique advantage, and of course it is an approach ideally suited to iterative, agile development methods. You can alter the product a bit at a time, taking your existing users with you whilst making it more attractive to new ones, and react to feedback along the way. A traditional waterfall approach is much more likely to lead to a big all-in-one redesign, and the risk that your product will become the new Vista, or even the new New Coke.