 |
 |
Ken, Very interesting. Although I may be preaching to the choir on this list, I always think it's important to point out two nearly-inevitable things in this sort of design research: (1) the specificity of the results and (2) the causal questions unresolved. My guess is that Jared would agree that these are both important to keep in mind. >I know you like to hear about our latest findings. So, here are some >from our most recent study of 170 shopping expeditions on 13 apparel >and home goods sites: > >1) The more users used Search, the less they purchased. >2) Users waited, on average, three clicks into the site before > attempting to use Search. >3) The more users used Search, the worse they rated the site. >4) The more users used Search, the less they told us they'd > return to shop on the site again. >5) The more previous experience a user had with a site, the less > they used Search on that site. >6) Two of the study's best performing sites, the Gap and Old Navy, > have *no* Search feature. (And the users didn't notice.) In the next paragraph he points out the specificity problem: This study was specifically about apparel sites. Does the no search advice apply to non-apparel sites? Apparently not. The subsequent paragraph brings up some questions, though. >In fact, there are very few (read 'zero') good things we can say >about Search in this study. Search did better in previous studies, >where we tested sites selling CD's and videos, but with sites selling >apparel and home goods, Search seems to fail miserably. > >I know I'll get more angry mail from the search engine software >vendors for saying this, but it doesn't look good for Search. Maybe >the Gap and Old Navy have it right: don't waste your time on providing >Search. Presumably the Old Navy and Gap sites don't have search engines because their designers thought they didn't need them. I doubt that's just brilliant organization; those sites are simpler than the Macy's site they were compared with. (I'm guessing about Gap; unlike Old Navy, you can't even browse around the Gap site without accepting their cookies so I moved on.) Old Navy has Mens->Pants and you have your choices. Macy's has Mens-->Pants-->choice of dress/casual/denim and then many more choices. With more diverse merchandise and more offerings in many categories, it's not surprising that the site is more confusing. So is having a search engine a negative or is -needing- a search engine the negative? I suspect some of his observations about brand impression also have to do with people's preferences about smaller, well-defined merchandise sets vs. the classic "department store" approach to sales. I'd bet that his users impressions of the websites correspond closely to people's impressions of the sites' respective mall stores. As a general rule, favorable brand impression corresponds to coherent brand impression. Ken, I know that you posted this as a sample of the reports from Jared and the gang, not as an endorsements of a single piece of research. And I'm sure most readers of this list don't really care about the web and/or read research reports carefully. I am also not meaning to attack Jared Spool, Adventive, or even this particular study. I think it is interesting and, as Jared implies, is part of a general accumulation of knowledge, not an end all/be all. But there's a phenomenon worthy of a folklorist's attention where a piece of research seems to indicate that under x conditions, if y then z and soon z is widely believed to be absolutely true and generally applicable. I've seen it happen with marketing research, science (particularly environmental and health studies), and public health/policy issues. It seems to be fed by a combination of researcher/institutional self-promotion (witness Jared's mild overstating of the results), bad reporting, and our collective desire for easy answers. The research-based legend is a serious problem for design research and the acceptance of research by designers. I'm not challenging anyone to a debate or demanding a response; I just think it's worth mentioning fairly frequently since it's a real problem for all of us. Gunnar -- Gunnar Swanson Design Office 536 South Catalina Street Ventura CA 93001-3625 USA +1 805 667 2200 [log in to unmask] http://www.gunnarswanson.com