A couple of caveats.
First, there are some bestseller lists, like that published by the New York Times, whose formulation are considered trade secrets. In the cases in which I'm privy to how these lists are created, I can't divulge that information; in the cases in which I'm not, I can't tell you because I simply don't know.
Second, it's not the case that all bestseller lists are concocted by publishers'/booksellers' resident mad scientists bent on establishing the literary equivalent of world domination by listing and discounting titles by megabestselling authors—they do try their best to list the titles that are showing the strongest sales figures in the marketplace. However, not all lists are created equal, and some are truer to the numbers than others.
For example: let's say Barrel O' Books maintains a store-wide "Top Ten" bestseller list, and they're overstocked on a particular title that isn't quite making that list. They may swap out the #10 title for the overstocked title, or may grant individual stores limited discretion when displaying the list, meaning it may differ slightly from location to location. (Book sales are surprisingly regionally varied in nature.) It's not exactly underhanded, since the action of adding the title to the list (and applying the appropriate discount, if applicable) will probably bump that title onto the list numbers-wise in short order. It is, however, something of a self-fulfilling prophecy.
Some bestseller lists also factor projected sales into the process, or will base the lists (at least initially) off how many copies are acquired by retailers from the publisher, and not how many copies are actually moved through the register. This is why some lists will reflect the bestseller-dom of titles that have gone on sale only that day, or a day or two prior. Granted, daily and even hourly sales figures can be and are available in our increasingly digital world, but it's not always the case that bestseller lists are produced with these numbers in hand.
In short: bestseller lists are good indications of how well a book is doing in the market, but it's not always the case that a book that makes the list is selling better than a book that doesn't, or even that #9 is necessarily outselling #10 across the board. It's not an exact science, mes auteurs, and so I wouldn't treat it as one if I were you—neither as a consumer nor a producer of the written word.
