I’m not familiar with your work, so I took a look. I see almost everything you do is franchised property. That’s not an accurate measure of what the rest of use do, or experience

If almost your entire catalogue of work is the Star Wars franchise, of course your data would be different than ours, because you’re getting different trade market penetration than others are. The point being that there is no standard baseline. Your assumption is faulty, because our observation is wild fluctuations in the data. 95% off for one book with 70% off for the next.

A sample like yours, almost entirely limited to a licensed franchise, is not the same thing as an independent book of the sort Jim Ottaviani does about research scientists.

Your numbers are not going to be remotely like my numbers because your market is the Star Wars market.

Your experience is not a universal experience. That is the point. I am sure James Patterson isn’t having my experience, either.

Moreover, I made no comment on the weekly data. I’m not drawing conclusions based on that. If I have a book that I have two years of Bookscan data for, and Bookscan shows me 500 sales and I have 10,000 sales, then we have a problem.

What you guys are trying to do is make a point that there must be some nice round number you can slap on this discrepancy and say “that’s the percentage that is off”. And there isn’t. This varies so wildly from book to book it’s impossible to draw a conclusion about one book that can be moved to the next. Or the next publisher. Or the next author.

The point of this post is not directed at number crunchers who just love their trendy data, the point is the data can be so damned far off from one project to the next that it can literally kill a creator’s rep, and that there are some people in publishing who seem to worship this data.

Or, as Will Wieser Associate Publisher, and Marketing Director for the Portfolio, Sentinel, and Current imprints of Penguin Group put it when asked “What is the hardest thing about selling your 2nd book?” “The tyranny of Bookscan”.

I think that’s pretty clear.

UPDATE: This website closes threads to commentary after a couple of weeks, but Mr Miller had another comment he wanted me to post.

“I fully agree the Bookscan data has problems with individual titles — I was speaking more to the macro level, which is what I deal with on my Comichron research site. I actually don’t use Bookscan much in my circulation research for the very reason that it isn’t more open; it forces individual authors to spot-check on their own and compare notes. My intent was just to add some more data points to the thread.”

Very cool. Thanks so much!

One of the problems I have seen with the system is that they sometimes have trouble mapping the information to the correct title — especially when you get multiple editions in the system. With books before Author Central added the service, this happens quite a bit.

To the larger issue — well, having seen the big files that Brian Hibbs reports on, I can say that at least on the end-of-year data he’s been getting, the figures seem generally reflective of what they “should” be, which is 70% or so of the returnable-copy sales I see on royalty statements. (That’s what Bookscan claims to cover.) So I tend to trust that annual dataset more than the weekly data, which I agree does tend to seem very erratic.

I wish I knew the answer. But whatever the a la carte searches are serving up, there’s still a floor to how much the bottom line of Brian’s end-of year Bookscan dataset can be off by. If it’s low by half, then the book trade would be doing half a billion dollars in GNs annually, which (as great as that might be!) doesn’t square with other evidence I’ve seen.

Zero?

Ahem.

I’ll assume there’s some sort of error that will get fixed next week. Never in the life of Sandman has a Sandman GN not sold a copy. EDIT: I am hearing from other authors who all show no sales last week. It does appear to be a glitch.