Over on the CHI-Web mailing list – and you cannot consider yourself civilised unless you subscribe to CHI-Web – we discussed the deserved demise of boo.com. Since the CHI-Web list is made up of usability types, we recapitulated the blind-person-exploring-an-elephant phenomenon and laid blame for boo’s failure on usability.
Then Jared Spool piped in, saying that since no site his company had ever studied had enabled more than 42% of test subjects to complete typical tasks, usability is much less important than we think it is. Indeed, it has next to nothing to do with a site’s viability.
Thus was unleashed, however unwittingly, a battle of doctrines. One Fred Flintstone, hovering over a shoulder, exclaims “Boo died because of crap usability. Usability rules!”
A second Fred Flintstone, hovering over the other shoulder and resembling Jared Spool, counters with “Usability is barely relevant in determining who flourishes online – our subjects are successful less than half the time.”
Should we be looking at a multifactorial approach here?
Isn’t it likely that, as everywhere on the Web, usability and every other factor are all interlinked? A well-designed site with usability proven in repeated testing and iterative refinement can still fail over a 14.4 modem connection. You “did everything right,” but factors outside control worked against you.
If we think about factors we can control, though, shouldn’t it be obvious to seasoned Web designers and researchers that a successful site has to:
We take these factors all together at once and arrive at a gestalt. boo was slow, looked great, offered nothing of real relevance to genuine consumers (the wares were too twee and overpriced), was actively unusable, and told you to get lost if you used a Mac (and failed with any combination of graphics, Java, or Flash turned off).
Amazon is fast, looks OKish, offers many useful items and some content (nothing we particularly like), is quite usable, if only for repeat customers who have One-Click Ordering set up, has never failed to function in any of a zillion browsers we’ve used, and will let anyone browse and visit.
Now think of a couple of also-ran E$ sites and try applying the list above – sites you visit despite their annoyances. You will end up with a matrix of Dungeons & Dragons–style hit points that can help diagnose the successfulness of the site. (In fact, should we start a new meme – Web-site hit points?)
It’s sort of like love, isn’t it? Or getting a cat. We went to the Humane Society specifically looking for a small, short-haired female cat and went home with a long-haired Garfieldesque male orange tabby who was so huge he could reach all the way to the countertop.
Or, getting back to love, you may say you adore redheads, but suddenly a black person sweeps you off your feet. Or, over time, your black friend becomes something more.
Maybe Web sites are the same way. Maybe, viewed as designers/researchers, we have to stop slagging sites because they fail singularly on component X when it’s components T through Z that are all relevant. Maybe we shouldn’t look for love armed with a yes-or-no checklist.
And maybe – this is pretty well accepted – people put up with sites for the first confusing minutes and keep at it until they get what they want. This directly contradicts the nielsenism that people will instantly switch to another site if what you’re offering isn’t what they want or if it’s hard to use.
(There’s so much wrong with that declaration. Among other things, only superusers know what all the other options are to go to in the first place.)
Is it also true that, even with all the other factors usability specialists cannot control, we need to take care of usability anyway? Though we might not have the absolute diagnosis of why boo went under (“Usability rules!”), we should still keep working to make other sites usable (despite “Things work only 42% of the time”)?
We say this knowing it’s already the way to go. contenu.nu principal Joe Clark spent nearly 25 years working on disability issues and accessibility. Even in very large cities, you’re never serving more than a few people at any given time – by some objective measures, especially to conservatives, that represents failure. But not for those few people. Same with our hit points: The site may be slow (demerit!) and look like shite (demerit!), but by gar, most everyone who comes to the site can actually use it for its intended purpose. Because we knew usability was important despite the other problems and kept at it to make it work.
Usability Is Like Love. We can see the T-shirts already.
Posted on 2000-06-06