An early millennial retrospective

Abstract

Under normal circumstances you might not expect to see a retrospective of the new millennium for at least a few more years (or perhaps a few hundred, depending on your enthusiasm for such things). However, given that this column started in 2000 and that web time is a fairly accelerated phenomenon, it seemed appropriate to pay a brief visit to bygone days for this final print edition of the Bulletin. Our journey begins in the year the internet industry collapse began. The term "dot bomb" may have been born in the 1990's but 2000 was its coming of age. Jakob Nielsen wrote an Alertbox in the middle of the year foretelling the death of web design, primarily from a usability perspective. In my first Bulletin article, Web evolution: Is HCI an endangered species , I looked at some of the implications of the dot.com demise and considered the future of web design. Have things changed? In general I believe they have. E-commerce sites in particular have discovered what works and what doesn't in terms of converting visits into sales. Happily, clear consistent navigation based on other successful sites is where the safe money is (and safe money is what counts at the moment). The intervening years have also allowed the industry to mellow a little. What was originally a pitched battle between usability professionals and the design community has receded to the occasional skirmish as cooperation and understanding between protagonists continues to improve. Let's skip forward to the middle of 2001 where I asked one of the perennial questions of usability testing: How many users does it take to change a web site? Jared Spool and his colleagues at UIE had just presented a paper at CHI on their experience of testing with large numbers of users ("Five users is nowhere near enough"). The paper reported on a study that failed to find even half of a web site's predicted usability problems with 18 users, compared to Nielsen's recommended five users for discount usability testing. It is a little disconcerting that this question remains largely unresolved even now, especially considering that Nielsen's original recommendation was made in 1989 (later revised upwards in a 1993 paper with Thomas Landauer). In "How many users…" I suggested that the complexity of web pages was the culprit-how can you expect to find the majority of a web site's problems with just five users when …