That bookstore shelf space is disappearing is a reality that nobody denies. It makes sense that there are people trying to figure out how to arrest the decline. There has been some recent cheerleading about the “growth” of indie bookstores, but the hard reality is that they’re expanding shelf space more slowly than chains are shrinking it. No publisher today can make a living selling books just through brick-and-mortar bookstores. For straight text reading, it is rapidly becoming an ancillary channel, a special market. Illustrated book publishers, whose books don’t port so well to ebooks and whose printed books are more likely to be bought if they are seen and touched, are working “special” sales — those not made through outlets that primarily sell books — harder than ever. That means they’re trying to put books into retail stores that aren’t primarily bookstores.
A recent Publishing Perspectives brings us an article by a small publisher envisioning an expanded market for selling books through libraries. Deborah Emin of Sullivan Street Press imagines a world where libraries become book retailers liberated from the normal retailer’s concerns about “exorbitant rent and the dealings with landlords who can terminate a lease renewal at will”. But what really caught my attention was this statement:
What if bookstores could invest in what bookstores are best at — filling their shelves with books and taking chances on new authors rather than being concerned that their stock won’t move fast enough and they are wasting valuable space trying to sell what is more difficult to sell but that they know can be sold?
This stopped me because, in fact, I have precisely the opposite take on the problem. What I see is that the cost of buying books, and the impossiblity of doing it “right” based on the sales and inventory data of a single store, is really the biggest barrier to profitable bookselling, even more of a challenge than the cost of the space.
One big component of the problem, in a nutshell, is that most books don’t sell enough copies to have a “sales rate” in any one store. Consider a little quick retail math. A store that does $1.2 million in sales a year ($100,000 a month) is selling 5 to 10 thousand books a month. Call it eight thousand. The chances are that store’s eight thousand sales will be more than 7,500 “ones”, with the balance made up mostly of “twos”, with a handful of titles — in the neighborhood of a dozen — that sell three or more. If the store turns its stock 4 times a year (which would be a very good performance), it is sitting on about 25,000 books at a time, also mostly “ones”, so let’s say they have 22,000 titles. So in the average month, 2/3 of their titles sell zero and more than 90 percent sell no more than one.
In the following month, the 7,500 titles that sell one will largely change.
There is no mathematician in the world that can make meaningful predictions for what any particular title will sell in a subsequent month with data like that. And there is no mathematician in the world that can tell you how the hundreds of thousands of titles not in the store would have done if they had been there, based on the store’s data on those titles (which is zilch).
In the past decade, indie stores have gotten some real help getting some indications about sales outside their four walls. Ingram ranks titles across a much broader universe. The store system provider Above the Treeline provides some title-level visibility across their client base. That’s a lot better than nothing, but the data is not provided in a form that would enable any automated use of it for reordering.
And that points to the second, and larger, component of the problem: automating the ordering. The human attention it takes to make the stocking decisions for a bookstore has not really been scaled. B. Dalton Booksellers, which was bought by, absorbed into, and then discarded by Barnes & Noble, pioneered automated models in the 1970s, the first real computer-assisted inventory management in bookstores. A buyer would set an inventory level and reorder point for a book in a store (“setting the model” or “modeling the title”) and the computer would take over from there, automatically reordering when inventory fell to or below the reorder point. This capability made Dalton grow faster than Walden, its chief competitor, which didn’t have this ability to keep backlist in the stores without buyer or store manager intervention. The shortcoming of the model system, of course, is that a buyer has to put it on, take it off, or change it. So we have a manual requirement to manage the automation.
When you think about the sheer number of store-title model combinations in a chain of hundreds of stores with hundreds, if not thousands, of modeled titles per store, that’s no trivial task.
Unfortunately, the art or science or technology (or all three) of inventory management for books in stores hasn’t progressed a whole lot since then. Barnes & Noble built a great internal supply chain with warehouses that could resupply its stores very quickly and that improved the efficiency of the models. But an unnoticed and uncommented upon current reality is that internal supply chain will be hard to sustain and increasingly costly as the base of stores and sales it serves diminishes in size.
My father recognized this problem sixty years ago and created the Doubleday Merchandising Plan to solve it. That plan provided vendor-managed inventory for the stores. The reps walked out with an inventory count rather than an order. It was posted (manually) to a ledger by a roomful of workers at Doubleday’s home office, and an order was then created and sent to the store which had agreed in advance to accept it. Sales exploded, cost of sales shrank, and this program propelled Doubleday into the top echelon of book publishers. Leonard Shatzkin’s system was not automated, but it was a lot faster and more efficient than the store’s own efforts, particularly in those days when there was no computer assistance to track the inventory.
As stores gained the ability to track inventory through the 1980s, and were further assisted by a wholesale network led by Ingram that could restock them quickly, improved inventory management sharply increased bookstore profitability and the bookstore network grew. But with bookstores now heavily invested in systems to help them order more efficiently, the need for and receptiveness to publisher management of inventory declined.
But stores that don’t normally buy books and which can’t make the investments in book-oriented inventory tracking and buyers with the huge amounts of special knowledge that book buyers have still needed the help. Nearly two decades ago, I helped a client build an automated stocking system that could manage inventory on thousands of titles in thousands of stores with very little human intervention. It has run successfully to this day and is used to stock books in three of the largest chains in the country.
We used a pretty simple logic to build this system, limited as we were by what computers could do in 2000. The system calculates stock turn by title across the chain and then ranks the books by that metric. Then each store gets the highest-ranked books it doesn’t already have each week to replace the books it has sold. This automated system is crude, but extremely effective.
Of course, persuading a bookstore to accept a publisher’s or wholesaler’s decisions about what titles to stock would be a very heavy lift. But as the retail book market shifts from dedicated bookstores to shelf-space-for-books in retailers with other specialties, it becomes easier for publishers or distributors to find shelf space that can be stocked on that basis.
Since I am now working on a more modern version of what we designed in 2000, it is easy to see that much more sophisticated ranking systems and stocking rules can be managed in an automated way than was possible then.
Changing the paradigm by which books find their way to store shelves is a way to meaningfully improve the efficiency of book sales in brick-and-mortar stores. Coupling it with true consignment terms (which sale-and-return is not) can make book sales viable for stores at lower discounts, which could meaningfully improve publishers’ margins.
There’s plenty of rent being paid for space books would sell well in. The problem is the cost of putting the right books into those spaces. We won’t get there presenting, ordering, and fulfilling title-by-title as we’ve always done. That’s the first place to look for a better answer. Reducing or eliminating rent would be helpful in the short run, probably not sustainable in the long run, and it would sidestep the real challenge of retail: presenting the most saleable possible mix to the consumers who will shop from it every single day.