2/12/2010

on the Facebook redesigns...

Many many people hate change - they train themselves on a given system (good, or more often bad), and then resist even the tiniest little change with a religious fury bordering on lunacy. Now, this is perfectly understandable, and is a critical aspect to consider in any redesign process. When in doubt, go with the established conventions unless there is a strong reason for change (AND DO PROTOTYPE USER TESTING DAMNIT!!). This is one argument for not releasing a product until the interface is set (or keeping it in beta for a long time... Google...).

The problem is that the establish, and quite effective general design process is to iterate and tweak as you gain user data and feedback. The challenge is to find a balance between useful improvements to enhance the user experience for new and existing users (which will sometimes mean major revisions of key elements of a design) and keeping the interface familiar enough for existing users to remain comfortable. The highly variable factor here is what "familiar enough" means - for different user demographics this will mean very different things. To me for example, Facebook has never had a 'major' redesign - every revision to date has been a fairly straightforward tweak that has retained their existing design language and conventions, but clearly this opinion is not shared by all...

I want to simply dismiss those objecting to the redesigns as inflexible whiners who need to grow up. And it is certainly true that if they were in charge nothing would ever be improved and the product would die out as competitors and innovators passed it (Myspace... pretty much all of 'Old-Media'... Ebay... Republicans... and for that matter Democrats even... the US railroad industry post WWII until the 70s/80s... American automakers...).

The problem is, the single most important part of design is meeting the needs of your users in the 'best' manner possible. Usually this means in the most elegant manner, but there is the very strong argument that satisfying your existing users at the expense of a stagnant, gradually worsening UI makes sense. I think the key is understanding who your users are, and who you want to satisfy. Do you want to innovate and stay competitive for the early adopter market that will jump ship when something better comes along if you stagnate, or satisfy your growing population of stick-in-the-muds?

In theory you should be able to do both by designing an intuitive-enough UI for everyone, but in practice people have such diverse learning styles, skill levels, and usage needs that this is beyond impossible. So what's the answer?
--Compromise and make fairly minor incremental changes gradually? (I would argue that this is functionally what Facebook has done, although it would have definitely helped if they had made more frequent but smaller changes.)
--Have multiple interfaces for different users? (phase people over to the new one gradually, or have user-selectable interface choices - far more complicated and difficult)
--Stagnate and sacrifice some users for more of the mainstream ones? (I would argue that doing this will inevitably lead to failure since the mainstream will eventually follow the early adopters to something better when the innovators come along, but this is debatable.)
--Keep refining the design with the goal of satisfying most of the users and to hell with the whiners? (this is certainly the most emotionally-satisfying one, but probably not the wisest - it would be the appropriate solution for a younger product still early in development with strong competition, but for a dominant force like Facebook far more problematic.)

Most of this boils down to who you are trying to serve, something that is less clear now that Facebook has grown so far beyond it's original core. Facebook does have a terrible track record for bone-headedly implementing new features/designs 'suddenly' with what looks to be very little user-testing and being shocked by the inevitable backlash. Certainly they could do a better job of implementing redesigns, but I don't feel like they have made any UI changes that are radically-different and confusing (poorly-researched, perhaps, but always in keeping with the existing design language and conceptual interface model). I think this is important if they want to keep users like me invested in stalkerbook indefinitely, but if they do want to go after the larger, but debatably-valuable mainstream market then they should be more careful.

No comments:

Post a Comment