The New Yorker recently reported on Simon & Schuster’s efforts with MediaPredict to develop what would amount to the “collective judgment of readers to evaluate books proposals” by reading selections presented on a website. The reason why any bottom-line oriented publisher would attempt to discover a better way of determining what books will be commercially successful is obvious to anyone familiar with the publishing industry — something like seven out of every ten books published lose money. Needless to say, more than a few people responded with comments suggesting this “American Idol” approach would doom the publishing industry to institutionalized mediocrity.

As those of you who have read more than a few of my books know, I believe that, with a few well-cited exceptions, extremely popular works of art in any field tend not to be excellent, and many of the few that are both popular and excellent tend to be those from earlier historical periods that have been propagandized by well-established cultural and social institutions. This is the way it is and has always been… and it may well continue. In the publishing industry charges and countercharges have flown back and forth for years, on subjects such as editorial elitism, genre segregation, reviewer bias, critical prejudice against commercially successful authors… and on and on.

For all that, the publishing industry has managed a remarkable diversity in publication, and in the F&SF field, small and niche publishers have broadened that diversity, as have more recent internet publishers.

What bothers me more about the MediaPredict approach is that it substitutes the judgment of one small group — those who enjoy reading off electronic displays and have time to read online — for that of another smaller group — editors and agents. Since my work has been far more popular with readers than with editors and agents — with the notable exception of one long-time editor — I certainly have always questioned the collective judgment of editors and agents. Any competent editor or agent can certainly tell what kind of novels are selling. On the other hand, it takes a truly outstanding editor to determine what kind of book that isn’t currently being published will sell, and there are very few editors who can make an accurate judgment like that on a consistent basis.

But will the MediaPredict approach make any better judgment on the commercial potential of a book? I doubt it… and here’s why.

Both online readers and editors are largely self-selected groups, if self-selected for different reasons, and this reflects the larger problem I see with the MediaPredict approach. The self-selection criteria for being an online reader effectively eliminate large groups of individuals from the selection process. Even some twenty years into the wide-scale personal computing/cellphone/PDA age, the majority of novel readers doesn’t read and doesn’t want to read books off a screen… any kind of screen. It takes a certain mindset to enjoy doing this, and I suspect that mindset is different from non-screen-readers. MediaPredict might do quite well at determining what kind of books appeal to that audience, but that audience is currently a minority of readers– especially outside the F&SF and possibly the thriller fields.

Editors, for all their short-comings, and they do have many, are held responsible over time for the success of their selections, and editors who tend to have too many commercial failures generally have short careers. There’s not even that check over the MediaPredict approach, nor has anyone asked one other critical question: Do the “screen-readers” predict accurately not only who likes the books being previewed, but whether they represent actual buyers? In short, will those on-screen preferences translate accurately into bottom-line profits? Because, in the end, that’s how the industry measures success.