The other day my paperback publisher told me they'd sold a total of n copies of Hild. On the other hand, for the same period Bookscan shows sales of 0.6 n. I have less exact figures for the hardcover but I think they're roughly comparable. This surprised me because reports I've seen indicate Bookscan captures 75-80% of points-of-sale.
If you factor in digital sales, which Bookscan doesn't report*, then the figure reflects less than half my market. I knew that Hild was doing well via channels that often don't report to Bookscan (smaller independents mostly) but, still, I was surprised.
In a year or so, when the final print and digital picture is clear to me, I'll revisit this, complete with pretty graphs. For now here's my back-of-the-envelope estimate: for Hild, Bookscan gets 46.5% of real-world sales.
* I'm sure it would love to, but Amazon owns a big chunk of the market (65% in the US and far more in the UK) and it won't share that data.