[dot.gif] UI logo [dot.gif] The Webzine for Interaction Designers [SponsorStatesoft.png] uidesign.net Old News Papers Editorials Interviews Reviews Letters Book reviews Aug 99 Book Review Web Site Usability : A designer's guide Web Site Usability by Jared Spool et al, Morgan Kaufmann, 1999 ISBN 155860569X We just don't know! ..... Yet! A quick scan at Amazon will show you that books on web site design are hot this summer. Several of them are in the top 1000 best sellers. As the world rushes to get on-line, are the authors keeping up with the pace? This very recent book appears to offer enlightainment for designers seeking to get that "cool" rating for their oh so hot new site. Web Site Usability a designer's guide by Jared Spool, Tara Scanlon, Will Schroeder, Carolyn Snyder, Terri DeAngelo, Morgan Kaufmann, 1999 So yet another book that sounds like its aimed directly at people like me - designers who design web site and eCommerce applications. This book is written by the staff of User Interface Engineering, a usability consultancy on the east coast of the US. This book is actually a published report. It includes a coupon to receive three free issues of their magazine, "Eye for Design". The content here is the kind of stuff that appears on their and others websites and usually doesn't cost anything. Personally, I'm a big fan of hard copy, even if it costs a few dollars. You can read it on a plane! So is this little book worth a few of your hard earned dollars? First Impressions Its a thin book, in a big type font. You very quickly realise that its a report on some research activity. This in itself is no bad thing. It will be quick and easy to read. There are lots of screen shots, perhaps an average of one per page. There is about 100 pages of real content, then there is a short chapter describing each of the real websites used in the report. The first chapter hooks us in, with the management summary. Firstly, we are told what they tried to test - information retrieval - not browsing or surfing, and more significantly, not on-line transactions or eCommerce. We are introduced to the sites that were tested, namely, Disney, Inc., Fidelity, C|Net, Travelocity, Websaver, HP, Olympic and Edmund's. In summary they all faired poorly but some were better than others. Why? The rest of the book we are promised will give us some insight. Then there is the key point. If you take one thing from this review or from reading this book then it is this - Web Site design is counterintuitive in a counterintuitive way. It is countintuitive++! We are then hooked by the 5 key findings from the study. These are * Graphic Design Neither Helps Nor Hurts * Text Links are Vital * Navigation and Content are Inseparable * Information Retrieval is Different than Surfing * Web Sites aren't like software Point 3 about Navigation is worthy of some comment. The implication is that a "shell" type structure as can be seen at UIDesign.net and is linch pin of design advice from Darrell Sano, is aparently a complete red herring. As a result of this, I have included a short review of Sano's Designing Large Scale Websites in the August edition. We just don't know I entitled this review, "We just don't know" because this is the true theme of the book. There are lots of observations on the research findings but the truth is that the authors are very short of answers. So if you as a designer are looking for the answers then you won't find them here. What you will find is a lot of thought provoking anecdotal evidence. It will be up to you, to apply your own experience and knowledge as a designer to draw the appropriate conclusions. From time-to-time, you will also find the results are in opposition to results published elsewhere by other usability experts. For example, horizontal rules which were found to ease and improve information retrieval in another study had the opposite effect in the UIE study. Navigation The first of seven areas which the report studies in detail is site navigation. Unsurprisingly, the tests found that the User Base for a web site is diverse and there are many differing goals. Users experience with the site varies according to their goal and the amount of Domain Knowledge that they already have. One observation was that sites designed by teams who normally design "extranet" applications with defined user bases had trouble with the more general internet site for a wide audience. This was true particularly of Travelocity which is designed by the same firm which makes the software that travel agents use to book travel. They designed a site which needed too much Domain Knowledge. None of this was particularly startling. More interesting was the observation that Darrell Sano's advice regarding a "shell" structure may be wrong, and even more strange was that "User's don't form Mental Models of Sites". It was unclear whether this would hold true for eCommerce sites. The big implication of this, which was unstated, is that developing web sites will necessarily remain a very labour intensive excercise, and by implication patterns and re-use of designs are not possible unless the content is similar. Hmmm. I guess that we "just don't know yet!" The chapter carries on with a device by device analysis looking at frames (which do rather well) and Table of Contents and Navigation Bars and Hierarchical Maps and Site Maps. It all concludes with some observations on Users and superstitious behaviour. The key conclusion ought to have been that browsers offer poor visual affordance for navigation but like so many other "design" conclusions that might have been drawn in the book, it is missing. Links Chapter 3 takes a look at the specific navigation tool of Links. We get the observation that the usability of a link is directly affected by the predictability and the ability to differentiate it from other possible links. There are a few good points made about offering suitable explanatory text and avoiding obfuscation of links. More could have been said about building the User's trust - another conclusion that was missing. A few other good points are made, such as Bullet Points are visually good and clear but use a lot of space in HTML. Throughout my notes in the sidebar of the pages, I kept writing the words, "poor visual affordance" which was the answer to the questions being posed. Its such a pity that the book didn't answer these questions. The teaching value of it is damaged as a result. Search Engines Chapter 4 is all about in-site search engines. Someone had to say it and I'm glad it came from such an authority as the authors here. Search Engines are often terrible. Its time they got better. Chapter 5 looks at Comparative data. It ought to be obvious that the web is bad at this - browsers just aren't designed for it. Again another chapter full of observations on what is wrong but no advice and what to do to fix it. How do we get browsers to support comparing two pages side-by-side? Not every user realises that they can open multiple windows. Readability & Design Chapter 6 was fascinating. All about readability, the conclusion is that less readable pages are more usable. How can this be? You'd better read the book to find out. Again the conclusion is that much more research is needed and really we just don't know yet! The concept that readers "skim" more on the web is an interesting enough theory, though. Chapter 7 has one of the big revelations. Graphic Design doesn't matter! Well maybe it doesn't but surely it depends what you are testing for? Well yes it does and the book does admit that this rule really only applies to information retrieval. It made a good headline - all the same. Another astounding revelation is that download time doesn't matter. Huh? Chapter 8 gives us a brief overview on User Preference. Strangely? Users don't always prefer the sites which are easiest to use. They often preferred the sites with the prettiest graphics or most interesting content. The authors don't offer any insight on this. We are left to draw our own conclusions. Perhaps web sites are expected to entertain not just inform. How the tests were conducted Chapters 9 through 16 provide a brief overview and case study of each site which was tested. Chapter 17 is perhaps the most useful in the book. It details how the tests were conducted and how to repeat them. This is a little off topic for "designers" as the cover suggested target audience. Once again a book by usability experts provides its best value to usability practitioners. There is much to be gained the world over by repeating these tests for other sites and gathering more information. I am sure that the authors will be glad of such feedback as they work towards answering the many questions which the research has thrown up. Summary I titled the review, "We just don't know, yet!" and it is true that this book ends up asking many more questions and leaving so many loose ends. Despite my relatively mediocre rating, I would still encourage you to buy this book. If for no other reason, I want to encourage UIE to produce a second edition. A future edition must cover eCommerce sites and it must start to answer the unanswered questions. This book has wonderful production values. Its well laid out, there are many illustrations which are always in a relevant location to the reference text. Its a usable book with some nice graphics for explaining the results from the research. However, there is no index, no glossary and no references to other work. Maybe they just aren't necessary. This book will do nothing for inexperienced designers who will be unable to draw any conclusions from the observations. More experienced designers will be able to take something out of it, even if its just a few keys points. The book is most useful to other usability practitioners who can use it to replicate the UIE tests and also use the UIE results as benchmarks. They will be able to evaluate a client site and say, "well its not good news but you did score higher than Disney!" So not really a designers guide afterall. Pity. Recommendation 3/5 - a future edition of this book will undoubtedly score higher when it can answer more questions than it asks. At the moment it doesn't teach web site usability to designers as the title suggested. uidesign.net hosted by likk.net [dot.gif] Old News Papers Editorials Interviews Reviews Letters [dot.gif] Copyright uidesign.net, 1999 - 2003. The UI logo device and uidesign.net wordmark are trademarks of uidesign.net