Hal O'Brien

This is a draft of a review I just wrote for Amazon. It's a draft and not the final submission because I tweaked the review in Amazon's own edit box, and forgot that now it's gone to their never-never land of "internal review".

---

Jakob Nielsen has made a career of overriding data found in the lab with quirky, idiosyncratic interpretations of his own. In an earlier book, he actually condemned empirical research on usability because "real users can be difficult or expensive to recruit," and that therefore the best method to establish usability is to "combine empirical results and inspections" -- inspections that just happened to be conducted and interpreted by himself, of course. (Usability Inspection Methods, p. 2, 1994)

This wouldn't be so bad if he interpreted data well. Instead, he has a pronounced tendency in this book to either refer to unsourced material, or to come up with some truly bizarre tangents.

For example, one of his most widely influential ideas from this book is the idea of "chunking" text. The premise is that users don't like to read on the Web -- which is inexplicable in itself, since there are only two activities one can do well on the Web, read text, and write text -- so they're averse to using the Scroll key. Therefore, one should break up long stretches of text when writing for the Web into multiple pages.

Why does he think users don't like reading online? "Research has shown that reading from computer screens is 25 percent slower than reading from paper. Even users who don't know about this human-factors research usually say they feel unpleasant when reading online text." (p. 101) Whose research? What methodologies did they use? When was the research done? Was the comparison between similarly themed texts (i.e., one may well read more slowly when reading difficult material like science texts, and more quickly when reading for leisure)? What percentage of users overall had these "unpleasant feelings" reading online -- 1%? 10%? 70%?

What do you want for today's Amazon price of $31.50? Information?

So his aversion to reading online text is unsupported, and would appear to be refuted by ever-increasing Web use -- what about scrolling?

Well, one could look at the index -- and see how badly it's done, with a scant few references to "reading" and "scrolling", despite both of them being frequently used terms throughout the book.

But on p. 112 we find: "In the usability studies I did of early web users in 1994 and 1995, few users ever scrolled. Maybe 10 percent or so of the users would scroll beyond the information that was visible in the window when the page came up. The only exception from this finding was users who had arrived at a destination page with an article that they found interesting or important to their work."

Fiction writer Theodore Sturgeon had a maxim: "90% of everything is crud." Rather than the obvious conclusion one could take from these data -- that 90% of Web writing is neither interesting nor important, and that therefore one should recommend to Web writers to make their writing better -- Nielsen decides the culprit is scrolling. Yup, if only Aunt Ethel would "chunk" her 30,000 word treatise on the antics of her cat Fluffy, it would suddenly become useful and important, and users would read more of it. Honest. Truly. Jakob says so. He has data from eight years ago to back it up.

Even all this wouldn't be so bad, if it wasn't for the fact that chunking directly contradicts one of Nielsen's other dicta -- keep the site simple, and easy for the user to access.

Here's an example Amazon users may find familiar: I have a number of Amazon friends who have written reviews. After reading their reviews on their personal listings, I would like to vote favorably for them. But because their reviews sometimes took place a long time ago, they're often buried pages and pages away from the book's main listing, because the reviews have been chunked ten-at-a-time. True to form of one of Nielsen's accurate observations -- that web users are an impatient lot -- the necessity of having to load all those pages in order to scan for my friends diminishes my enthusiasm.

Which is a shame, as some of my friends have written fine reviews.

Another contradiction: on p. 385 we're shown a screen capture of an early page from Cosmopolitan magazine's web site. The caption tells us: "The Web is not print. And a home page is not a magazine cover... A magazine must be arresting above all... On the Web, the user is already at the site and has already chosen to do business with the site." But on p. 112 we're told, "...each hypertext page should be written according to the 'inverted pyramid' principle that is commonly taught in journalism schools." Why? Because users can leave a site at any time, and they should be able to take away as much of the web site author's ideas as possible. You know, much like the way newspapers and magazines work. Not that the web is print, or anything. Or that the web is, to use a term not in the index but frequently used by Nielsen, an "attention economy".

I will concede the book has a certain entertainment value for one critically minded -- let's play Find The Inconsistency! -- but one would do better to seek clarity, empirical results, and wisdom elsewhere.