Comixology and DC/Marvel do not publish digital sales statistics (at least, not anywhere I can find). The only means we have of judging sales is through print sales, as reported by Diamond. Nobody else reports actual sales, as far as I know. And of course, Diamond reports sales to the shops, not purchases by the customers after the shops have the books. So if issues are being bought up by the shops in a massive fit of speculation, but then sit around in unsold piles because customers don't want them, there would be no way to discover that.
Undoubtedly, part of the disconnect between ComiXology pull list and sales numbers is that many ComiXology pull list users buy digital rather than print comics. However, there is no way to test it vs. digital sales, as those are not reported.
Still, one would expect that there should be relatively little statistical difference between print sales and digital sales for a given series. I have never been able to divine a good reason why, for example, digital readers would swarm all over, say, Action Comics, but print readers would skip it. In general, the books reported as top digital sellers by ComiXology (without sales numbers, just ranks) seem to rank about the same way as the print rankings.
Therefore, one would expect, and I certainly did, that a high-pull comic would sell high on the Diamond list, and a low-pull comic would sell low. There would not be a one-to-one correspondence, but I certainly expected some statistical correlation. Yet I didn't find any.
Maybe I just chose the wrong books. Since this can really only be done by typing all the numbers in by hand (there's no obvious way to DL large databases of this stuff), and I didn't have the desire to spend hours typing things in, I only checked with a few of the comics whose sales patterns were most familiar to me.
If someone has a way to DL large amounts of data and can point me to it (say, the pulls for all of August and the sales for all of August) I'd be happy to do a more extensive correlation and see if there really is a match.