Algorithm, Anyone?

Despite being ritually deemed “the bane of the publishing industry” and “one of the costliest aspects of the business,” the problem of book returns remains an estimated $7 billion thorn in the industry’s side. Last year, BISG figures show, the average adult trade hardcover return rate was 37.5%, up 3% from the year before. Publishing’s response? “There is no statistical measure” for forecasting frontlist book demand at the world’s largest publishing house, says a knowledgeable source, while “gut-feel” remains the crude weapon of choice to battle supply-chain surfeit.

If Barnes & Noble has its way, action on ramping down returns may get shoved to the front burner. Two years ago, B&N went live with an inventory control system from i2 Technologies, which has introduced “very sophisticated statistical forecasting algorithms” to gauge title-by-title demand, says Anant Mahale, i2’s Program Director in Consulting. Replacing what was described as an ad hoc system, B&N’s new program tracks about 100,000 titles that contribute 80% of the bookseller’s business. Based on past order history, seasonality models, and other algorithms, the system also takes into account the vagaries of supplier lead-times — in effect, studying wholesaler or publisher shipping errors and creating a special model for each vendor — to predict how much inventory B&N needs to keep in stock. Mahale says the system, now operating at two distribution centers, has helped B&N slash its inventory levels yet keep customers happy. B&N executives have confirmed that the chain may well cut inventory levels 30–40%, saving $4 million annually, on top of a one-time $13 million inventory reduction.

To be sure, other vendors prowl the supply-chain space — among them Manugistics, which has been working with Scholastic, as well as TMS and its Bookmaster system, whose US clients include Columbia University Press — and there’s always the data provided by Bookscan to help monitor sales. But algorithm advocates say the industry’s holy grail — and a strategy B&N has “expressed a lot of interest” in — may lie in what’s known as a “collaboration solution,” whereby a distributor and publisher work off of the same forecast, so that publishers have advanced visibility into the distributor’s needs.

Mike Shatzkin’s Idea Logical Company, on the other hand, has worked with nine publishers this year on an analysis of one chain’s sales data. He thinks the returns problem is due to too many decisions, and not enough bandwidth to make them. “Barnes & Noble and Borders are trying to manage 100 million stock levels with a minimal level of automation,” he says. “What can be done at the distribution center level is important, and can make a critical difference to fill rates and inventory carrying costs. But high returns come principally from the impossibility of making timely decisions on a title-by-title, store-by-store basis.”

At B&N, anyway, the proof’s in the profits. The bookseller cited “improved margins” and “cost controls” in announcing its third-quarter profit last month. And a filing from trumpeted higher gross profit and gross margin due to “an increase in the Company’s internal fulfillment rate” as well as “more efficiency in fulfillment and customer service operations.”