So, though I ended up in a creative field, I started with nuts and bolts (literally) and had the run of every store on Main Street. This has led to my becoming notorious for insisting, when someone makes claims about something, that I want the hard data: the numbers. And in the case of publishing, the money numbers. (It really annoys people, by the way, when they make an assertion and you say "What data support that? What numbers do you have?"
This may be a two-parter, thanks to LJ's 4300 character limit (grump again) but here goes.
First, let's talk numbers as they matter in the writing business. Publishers (of magazines or books) need to make money to survive. Publishing requires money for the entire process: staff for editing and copyediting, paper, typesetting, printing, cover and sometimes interior artwork, storage, shipping, advertising (including website design and maintenance) and (these days) payments to bookstore chains to get guaranteed "placement" in the bookstore (front of store tables, end of row displays, face-out display...) Publishers know, to the penny, how much income per page they need to cover costs...and thus how many copies a 300 page book needs to sell to cover its costs vs. a 350 page book, and how many more copies must be sold to cover the art costs of a book with a cover by an expensive artist, or interior artwork, and every other related expense. These things enter into their decisions to buy or not buy a given book (they won't buy one they don't think will tank), and how much to pay the writer in an advance. Two important things about the advance (no, three, and don't anyone mention the Spanish Inquisition...) 1) A book will break even for the publisher before it "earns out" for the writer and the writer starts getting royalties. This is the publisher's safety net. 2) Advances are not paid in full on signing...they're metered out between "signing" and "publication" with a variable number of payments along the way. 3) Most books do not "earn out"--the advance is all the writer gets.
"Numbers" in book-talk cover a range of things: the size of the initial print run (determined by pre-orders from bookstore chains for established writers, and by the editor's best guess for less-established ones), the raw sales figures and the shape of the sales curve (early peak, later peak, flat), and the sales/return ratio, known as the "sell-through." Since bookstores return books that do not sell within a given time period for full credit, they don't really care about sell-through.
Each of these numbers has different implications. Initial print run determines how many books can be sold in the first few weeks...maybe 5000 people would have bought the book, but if only 2000 are available, only 2000 will be sold. It takes time to reprint a book and ship it out...and by that time the 3000 people who would have bought it but couldn't may have spent their book budget on something else. It is common for initial printings to be small enough that the book would not earn out even if every copy sold. Writers often push for a larger initial printing., knowing that having their book in front of potential readers is the best way to make a sale (most people won't special order a book that's out of print, in order to nudge publishers into reprinting it) ..but if sales are low, that larger initial printing becomes a lower sell-through and lower net income from the book. Though the actual printing cost of a book is only a fraction of the total cost of production, printing more does mean a higher total cost--so more must be sold to cover the cost.
Publishers care most about sell-through, because sell-through is what makes or breaks profit. They also care about total sales, but high total sales with a miserable sell-through can still tank a book. Here's why. Say you ship 10,000 books that cost $5 to produce and you're selling them for $20 with a $10 discount to the bookstore (yes, publishers sell books to bookstores below the cover price. That's how bookstores make money, even with discounts.) The publisher looks to make $5/book. But half those puppies come back, sent home by the bookstore, which now wants $10 credit for each one it sends back. 5000 of the books sell, netting $25,000 for the publisher. They cost $25,000 to produce. That's not a good ratio but it's survivable. But say the sell-through is below 50%...say the book tanks and only 20% sell. The cost to produce those 10,000 books was $50,000...the return is only $10,000...the publisher is now in the hole, not somewhere they want to be.
Raw sales figures are what matter most to bookstores. If someone's first book sells 15,000 copies, and their next sells 26,000, the chain stores will see that writer as viable. With the reverse, they won't--and they won't order that writer's next book at all...thus deepsixing that writer's career. (Publishers won't publish what they know bookstores won't carry.) If the writer has one book with lower raw sales--even if previously that writer's sales were high--the chains won't order the same number of his/her next book. ANY drop in sales affects future sales...even a single point downward suggests to the beancounters in chain bookstores that the writer is becoming a has-been. Been there, experienced that with two different books, and it took years to regain publisher and bookstore support (and equivalent income.)