We had an A list of titles that sold quickly (the hugo & nebula winners, etc.) and we'd check those and order them as soon as we had a minimum order available. Then, once every two or three months, we'd do a stock check of the entire store, and we'd order based on that check.
Shelving and returns are important because shelving is the only real way to get to know your stock. It's not a lot of fun; it's fairly mechanical, and when there's simply not enough room, it's its own special kind of torture. There's no better way to know what you've got, and how quickly it's selling, though.
Enter computer inventory systems. Things are done differently now. Not only at our store, but across the board.
Now, we try to order what will sell within a month. Given the cost of retail space in Toronto (and possibly everywhere else), and given that we've never been located in a building that wasn't built at the turn of the century, we don't have a lot of back-room space for overstock (which is the stock that doesn't fit on the shelf). The space we do have is best devoted to new release titles, because those are the ones that will sell in the largest numbers out of the gate.
Since we've so little space, we don't carry a lot of inventory that we don't have to. This means that we rely on the publishers to warehouse their books. But so does everyone else.
What this means is that the initial order of a frontlist title is not the lead indicator that it used to be. There are titles for which the initial orders weren't nearly high enough, and books that can go into a second printing before the book hits the shelf the first time. How? Go back to the sales reps and the way the orders dictate print runs.
Factor into this the unwillingness of the chains, in recent years, to warehouse/inventory stock. I had mentioned that the print run will be based on rep orders plus a percentage, usually between 10-20, of that order. Part of the overage in percentage is decided based on optimism about how the title will sell, and what kind of reorders it will get (remembering that it's always better to print a larger run on a per unit basis); part of the overage is to cover the late orders that trickle in via catalogue mail-outs and telephone sales (if your store isn't a large enough account, you don't get reps to handle it; you get catalogues instead, and the timing of your orders is then, well, yours).
However, human error, and a change in orders can upset the books print number when it's already been set at the printers, which means that the initial run might be too small to fulfil the orders that come in. Which in turn means that another printing will have to be done. Does this mean the book is selling well? Not exactly; it means that the initial upfront order wasn't high enough, and that the buyers were placing the load of inventory and warehouse costs on the publishers, rather than on their stores. If the book sells quickly, they'll reorder -- but if the publisher didn't forecast this, the book is gone.
What this means, in a lopsided way, is that sell-through is a bit better over all, but that orders are down up front; that it does take longer to see the overall numbers of the book emerge (where before, you were simply waiting on returns, now you're waiting on everything).
Warehouses are not, as you might expect, free. They're not particularly cheap, either. I'd be willing to be that amazon.com makes more money, after costs are factored in, for its "but this used" sales than it does for its new books. Why? Because the used books are warehoused and handled and shipped by the seller, a person who has nothing to do with amazon.com.
A warehouse, if rented space, is costly; if you buy it, you still have to pay for basic utilities, taxes, equipment, supplies (boxes, peanuts, that stupid bubble wrap or paper), computer systems to handle inventory, etc., etc.,) and labour. Labour at a warehouse is usually more expensive than labour in a bookstore (which can often be had for a smidgen above minimum wage). Someone in theory goes and picks your order from the warehouse; if it's a backlist order, they may be picking 300 titles, two books at a time. Publishers of any size usually cover all shipping costs as well.
But because retail space is also expensive, and because inventory in a backroom often stays there (rather than making its way out to the shelves), it's more efficient in many ways for bookstores to force the publishers to carry and warehouse the stock for as long as possible; it also keeps bookstore costs down because the bookstore is charged for the books shipped, and given a certain amount of time to pay (this varies. In theory it's all net 30 days from the date of invoice, but in practice, the chains pay far later than that because, among other things, who is going to put a chain on hold and refuse to ship anything to them until they pay? Which is one of the reasons that the whole Canadian implosion occurred. If people want, I can try to speak more to that in a different post).
But back to topic, for a bookstore, leaving books in a publisher's warehouse until they're needed on your shelves is the most cost effective thing to do. Computer inventories, in theory, make this easier as well.
But it does mean that warehouses become more costly at the other end, which could affect things like which titles stay in print.
This is a funny business.