Whether you think about it much or not, all products have users; the users, in turn, have expect­ations. Whether you look at a physical object or a website, similar principles and rules apply. A car steering wheel should turn the car in a corresponding direction. The position of a light switch gen­erally indi­cates whether the light is on or off, and often suggests what is actually cont­rolled. Inter­faces need to be built for and around their users—this is known as a ‘human-based interface’, rather than a ‘machine-based’ one—yet they often actively exclude them.

People don’t really like reading on computer screens. Text rendering is getting better, and so are displays, but that doesn’t change the fact that reading lots of content in a small space winds up awkward, at best. So why, then, does Amazon have so much text – and why isn’t it related to what most users need? Have you ever purchased something at the jewelry store? Do I really need to “Save up to 60% in the Bed & Bath White Sale”? Nope. Does anyone buy cars from Amazon? Musical instruments? Probably not, and cert­ainly not the majority of the store’s shoppers.

Just about everyone buys products at Amazon, but they’re almost all buying books or DVDs. I just read that a large majority of the computers sold on Amazon were Macs. Clue alert: that doesn’t mean lots of Macs are getting sold, it means hardly any computers are. Yet the front page of amazon.com is covered in text, mostly links to spe­cific stores and lots of recommendations. Why not em­phasize search – what most people do first anyway – while hiding the useless components and improving the recommendation engine?

On a more basic user interface level, what about links? Everyone loves pointing to Nielsen, who wants them to all be blue and underlined. I think we can all agree that his usability desire is neither realistic nor ever likely; links look different on every site. But what if links weren’t links? There’s a service called IntelliTXT used by some major websites to turn content keywords into advertising out­clicks. It’s utterly horrific, as you can see for yourself:

IntelliTXT makes users second-guess their own actions, and eliminates any barrier between content and advertising. The supposed “iTunes Music Store” link above, for ex­ample, links to some sketchy third-party download site. Users expect links to be related to the content on a meaningful level, rather than a superficial advertising off­shoot. If most Google users can’t tell the difference bet­ween ads and results, do you expect them to spot ad links in an article? This type of crafty advertising undermines user loyalty and is a usability disaster.

On a more positive note, sites that do seem meet most users’ expectations include google search and maps (now ‘local’), both of which are highly successful and quite in­tuitive. The 37signals web apps, like Basecamp and Writeboard, are also clean designs with carefully-crafted usability features. There are, of course, hundreds of great websites, but they’re far outnumbered and overshadowed by their less pleasing compatriots. Designers and marketers need to more thoroughly evaluate their target audience, rather than assuming their product is working well enough.