Thoughts on design principles
Originally a comment on the post "effects of a design" on Paul Scriven's thoroughly excellent white__space (incidentally, Paul is pretty cute).
[The problem with web design] is that you're thinking of design as having only one purpose. Design is many things:
- Graphical design is the bit that gives customers their first impression. It's all-important for first-time visitors, but once they're customers they want it out of the way. You have to balance these.
- Usability design is massively important for first-time visitors; they need to be able to use the site instantly. But once they're long-time customers they will have learnt your interface, no matter how bad it is. Thus the poor designs of Amazon and Ebay matter not one bit -- they've hit critical mass, so it doesn't matter that they're crap. (Also why Windows is still being used)
- Functionality design, i.e. what functionality you include, matters a lot to repeat customers: being able to do a lot of things, things that you might not want to do so often. These rare actions are unimportant to new visitors and in fact presenting them with all options will probably confuse them, so that's another balance you have to strike. Ebay lives on because it's got everything and the kitchen sink, even if it is confusing.
- Usability efficiency: finally, for frequent visitors, usability comes into play again, but a different kind -- they don't care if it's not obvious what to do, once it's very quick to do it once you've learned how. Amazon's one-click solved this problem for them.
Incidentally, the three major desktop operating systems are all extremes of each school of design:
- Macs have historically been all about the new user: graphical and obvious, but laborious for a "power" user to get lots done (no longer the case in OS X, since it's got UNIX stuff underneath).
- Windows is all about functionality: everything and the kitchen sink is in there, *somewhere* -- just keep clicking around and it'll turn up eventually.
- Linux is the power user's operating system -- it's totally impossible to work out what to do from a standing start, but once you've got the hang of it you can be very efficient indeed.
And this is where operating systems are after more than 30 years; they haven't solved the balance between these design schools, they've just chosen a niche each, and are sitting there trying to take over the other two:
- Macs are (quite successfully!) wooing power users by adopting UNIX under the hood.
- Linux is furiously trying (with mixed results, so far) to produce a new-user friendly graphical front end.
- Windows is adding remote terminal services and other power-user-friendly features (also with mixed success, and I'm a windows user).
Websites as a concept have been around only about a third of that time, and very few individual websites have been around for more than 5 years. So it's no wonder we haven't found this balance yet -- maybe it's not even possible to find the balance, and you have to lean towards one extreme or another, depending on the types of users you typically attract (forum sites will want functionality, e-commerce sites usability, corporate sites graphical design).
So if your question is "what use is your pretty site if it doesn't improve my metrics" the answer is "none" -- if it's not improving your metrics, then you're not concentrating on the right type of design for the types of users you have (or, just possibly, your site is perfect. But that's not very likely, unless your site is extremely simple).
Update: one of the reasons RSS is so popular is because it effectively provides a power-user's interface to the generally quite graphical web: all the content, none of that pesky interface. Perhaps one should be developing several style sheets, or perhaps completely different interfaces, for different types of users?