[originally posted 6/27/07]

Folks have asked my opinion about the article in the latest Comics Journal about whether the Diamond sales figures distort the market. In fact, I have had an expansive (and friendly) conversation about it with article author David Beard, in which we found many points of agreement about what the statistics say and do not say. Here's the basics of my take on it:

I wrote in CBG last year in a column on the Diamond charts that what they report, and the form that they are in, are largely artifacts of a different time. The truth is that this data was only ever intended for the use of comics retailers in making decisions about what to sell and in what quantities. Milton Griepp himself will tell you that -- and did, in CBG a few months back. There were two reasons Milton used an Order Index number rather than actual quantities: He didn't want to say exactly what Capital's sales were relative to other distributors (understandable) and the ratio made it easier for retailers to use them as they were intended: as a means of telling them when they should adjust their orders. With no Internet transmitting "buzz" on specific books, one of the only ways word on a Watchmen phenomenon reached the hinterlands (besides distributor sales reps telling retailers) was through the trade press and conventions. The charts helped to address that.

But while Capital, Diamond, and Heroes World (three that ran such charts) did put them into their consumer publications, the indexed columns were generally removed. It was never imagined that fans would be interested. My own data -- which I began calculating to address the lack of information in the 1995-96 Marvel/Diamond schism -- ran exclusively in Comics & Games Retailer  until 2005, when so many people were posting the data online that it really didn't make sense to be the final holdout. That's when the library of data on CBGXtra appeared -- and the editors moved a set of charts into the magazine.

The result is that, while I try to infuse many grains of salt when I report -- being specific almost to the point of reader boredom when it comes to clarifying that a particular measure is "combined dollar sales of Diamond's Top 300 Comic Books and Top 100 Trade Paperbacks" -- not everyone who talks about them is so specific, and even I will go to the shorthand of "industry sales" for headlines. So, yes, it is easy to overstate what the data is representing -- completely apart from overstating its significance!

On the other hand -- and this is where I diverge from the article's analysis -- I think different writers may differ on what we mean when we say "the industry." Since the original audience for my charts was the readership of Comics & Games Retailer, "the industry" meant the direct market -- and while I acknowledge that comics shops do have other sources, what they add to the picture tends to be less meaningful, numerically -- particularly when it comes to periodicals. Diamond is the exclusive reseller to comics shops of products from a well-known slate of publishers -- and few retailers have returnable accounts for those products on the side.

OK, so the Diamond information has gotten outside the direct-market box -- and now people are using it to speak of the whole market. How good a measure is it? For trade paperbacks, pretty bad -- as I readily acknowledge frequently. The Top 100 trades don't even represent half the dollars Diamond makes on trades in a given month -- and while I have the ability to find the whole Diamond number, it of course leaves out alternative distribution and most manga. And with trades, Diamond is one seller of several -- as the article points out. (You'll note that on Comichron, I don't include the trade lists at all, for that reason.)

On the other hand, as a measure of periodical sales, the Top 300 is a much better measure. Algebra tells us that the contribution of the items after #300 add only a few percent to the total -- and the returnable market is absolutely moribund. Sell-throughs on the newsstand are lower than 20% on whole lines -- and the contribution of the entire newsstand is probably finally down to the 10% range, just for those publishers that sell there. (Yes, individual mileage -- like Archie's -- varies. More about this in a moment.) Subscription copies, likewise, are a small fraction compared to Diamond -- if you could calculate them.

Which brings me to the practical aspect -- which is, does the error, that which is being missed by a measure, warrant the effort that would be required to correct it?

In the case of periodical sales, the answer is "probably not." The nature of returnable newsstand circulation means that getting that data on a monthly basis is logistically unlikely. (Yes, Ingram has monthly pay-on-scan reports for magazines -- I've seen them -- but that information is proprietary, and in any event merging the two data sets would be a challenge.) You'll never get subscription sales into this picture. And for the outliers, like Archie, who sell heavily into the newsstand market, it is well to ask: Who does not having this information available affect? The publisher, certainly, in the "horse race." But the direct-market retailer? Not so much. These sales aren't "in play" for the direct market, in the sense that no one but a newsstand retailer would be likely to sell, say, a $3.29 Archie digest effectively (and in fact, the digests are most of Archie's sales).

With trades, it's another matter. The fact that a book has strong sales outside the direct market is obviously meaningful to the "horse race" -- and it is also meaningful to retailers, depending, similarly to the Archie case, on what it is.
 
Unfortunately, there's no good way at present, logistically, for getting the bookstores' part of the picture. Bookscan reports are proprietary; when we see Bookscan lists with numbers hitting the web, someone has probably violated somebody's agreement with Nielsen by publishing the numbers they have. Obviously, we're all glad they did -- I thrive on leaked data -- but that's not a good mechanism for producing regular reports. You'd have to have some kind of an agreement with Nielsen -- certainly beyond the financial means of any of us private observers -- and then, you're dealing with a serious data reconciliation problem. Merging lists of 400 comics a month from multiple distributors was bad; dealing with the whole 5,000+ item backlist every month would be simply impossible, without Diamond pitching in by flagging ISBN-13 numbers on every single chart it releases.

So it all comes back to Diamond -- and while I have lobbied for a long time to get ever more data, the question that always must be answered is, indeed, what benefit does Diamond get from that extra effort? Will it increase direct-market retailers' orders (their bread-and-butter, compared to the less profitable bookstore market)? So far, the answer has been no.

Lastly -- the article delves a bit about sales to retailers versus sales by retailers. This is certainly a distinction that exists, but I would argue that it has grown less meaningful over the years. Both Diamond and I initiated sell-through surveys in 1994 -- but in the years of sales hardship that followed, retailers played their orders much closer to the vest. Sell-throughs are much higher in 2007 than they were in those days -- with what overage there is more often than not finding its way to eBay. (We can see that inventories have dropped just in the number of shops that no longer sell back issues on the floor.) The "phantom reader" -- the phenomenon that caused publishers in the early 1990s to perceive customers where they weren't -- has largely disappeared. I would say that the distributor sales are probably closer to actual sales now than they've ever been. And, again, the error introduced by not knowing the difference is so costly to correct, it's not worth it. A single POS system for thousands of one-store operations would seem not to be on the horizon. Though, again, I would sure like to see the numbers from it!

I guess that's the key, as far as I'm concerned. A lot of work has gone into these figures in the past, but I'm not so protective of that work that I don't want to see an improvement to the status quo. When Diamond has made more specific measures available that have added to the picture of units and dollars, I've happily worked them in -- even when they've caused apples-and-oranges problems of comparison with the previous data. I'm happy to deal with the historical comparison issues if the result is more complete information. I think all of us working with the figures feel the same way!