Monday, March 30, 2009
I've been needing to get this online into the FAQ section of the site for some time — there actually was one before my old database imploded last year — but it comes up enough that it's worth a permanent note, in addition to the material listed under "The Fine Print" here in various places on The Comics Chronicles. (Some of that got also eaten in a recent redesign; restoration is underway.)
Estimates based on Diamond's sales charts — and before that, Capital City's — have been appearing for years, frequently sparking discussions as to what the numbers actually are reporting. Publisher purchase orders, creator royalty statements, print runs — all manner of other information about comics sales regularly available to some in the industry may be found to never coincide with the Diamond tables. There's a reason: what Diamond is reporting, and what those sources represent, are different groupings of comic books. In fact, it is a calculation unique to a moment in time, and to Diamond Comic Distributors itself: Unless you have an adding machine in the Diamond warehouse clocking your own comics for an identical period of time, your mileage will vary.
For reference, here’s what is NOT in the Diamond reports — and thus, in the Comichron, and ICV2, and ComicBookPage estimates of Diamond’s reports:
1) Diamond’s sales to the U.K. These tend to be in the 10% range when it comes to units. Diamond sold approximately $36.8 million in comics, trades, and magazines in May 2008 in North America; my best guess on UK sales is another $3.8 million. These are reported disparately to some publishers, but may not be divided out in other available publisher information. A print run of a book that sold out in a single month, for example, will invariably be higher than the Diamond estimate because overseas copies are not in the Diamond totes.
2) Newsstand sales, if the publisher has them, as well as any sales through other distributors, direct sales, or subscriptions — again, if the publisher has them.
3) Trade paperbacks through Diamond's returnable bookstore program — or anyone else's.
4) Any comics shipped by Diamond outside the exact period being reported. This one sounds obvious, but it's very significant. Comics shipping in the final week of the month find their sales bifurcated to a greater degree than comics shipping in the early weeks of the month. All sales are still reported — and so the aggregate totals are unaffected — but a Week 4 or Week 5 title will have a smaller relative fraction of its total sales with its initial entry.
Before 2003, we could safely say that the charts underestimated sales on all comics because they only reported preorders. Since 2003 and Diamond’s move to final order reporting though, we do get reorders that ship within the same calendar month. This benefits Week 1 books more than any on the list, as a consequence.
It's one reason I place more emphasis on the aggregate figures when it comes to studying the historic health of the industry — and relatively less on the internal trendlines of individual titles. Month-to-month comparisons become complicated when a book ships at different times in the months being compared. It’s this, not just the issue number, that causes the decline you usually see when multiple issues of a weekly ship in the same month. It’s not necessarily that sales are trailing off — just that the earlier issues have had more weeks in which to gather reorders.
It's also not entirely clear that the "shipping month" for Diamond exactly corresponds with the calendar in all occasions; it appears to match a set group of shipping days from the warehouse, but holidays and other events might cause minor alterations to the reporting calendar, as we've seen from some books being reported that came from slightly outside the reporting period.
So excluding all the above, the tables Diamond releases are said to reflect comics shipped in the calendar month to Diamond's retail accounts in North America. How reliably are the estimates that observers have been calculating for years reporting the numbers behind the table? The answer depends on the year being discussed.
Before 2003, when the tables were reporting preorders, different analysts found varying figures depending on where they sourced their actual sales figures from. Different publishers received purchase orders on different days, and so there was seldom perfect agreement between sources as to what the "magic number" was. I addressed this issue beginning in 1996 by taking data from several publishers at once — until I had data from comics accounting for nearly a quarter of the list. That narrowed down the margin of error quite a bit. While I maintain that method produced more reliable results, it remains the case that before 2003, different analysts using different publisher sources would always find slightly different results.
After February 2003, however, Diamond shifted to reporting actual comics shipped, including reordered issues. At that point, inter-publisher differences with the data very nearly vanished. It appears that, at that point, Diamond streamlined the issuance of order data to many publishers so that it synched up better with the shipping period being reported. Independent observers using the reports that publishers received from Diamond immediately found that margins of error collapsed, with the order index numbers far better reflecting the numbers Diamond was sending different publishers for the month. My estimates, Milton Griepp's, and others converged, even though we're all using different publisher data to calculate the "magic number."
So since 2003, the estimates do appear to be reliably reporting the numbers behind the Order Index Numbers — the comics, Diamond says, that it shipped to North American accounts in the shipping month. They are based on data that Diamond has itself provided publishers, corresponding to the period indicated, and that the publishers have provided analysts. Those numbers are not the sum total of comics a publisher sells, or even sells through Diamond, or even sells through Diamond in a single month. No royalty statement, no boardroom report will ever match the numbers in the table. As it is the most transparent and regularly reported indicator available — it has wide circulation, and deservedly so. But it is not all of circulation, nor do they (or we) claim it to be. When it comes to reporting the health of the comics market, it is instead a clue — a very important one, but not always the whole story.