The Shatzkin Files

Buying is a hard thing for bookstores to do effectively, and that becomes an increasingly important reality for publishers

Tweet about this on TwitterShare on LinkedInShare on FacebookShare on Google+Share on RedditShare on StumbleUponEmail this to someone

One of the most underappreciated realities of the book business is how hard it is for a retailer to manage an inventory of trade books.

This is an existential problem for a bookstore. A bookstore’s inventory is its biggest investment. The performance of inventory — how many times it “turns” in a year and how successfully the store manages to buy what it needs without wasting investment (tying up cash) and incurring margin-destroying revenueless-costs (return freight, probably added to inbound freight, plus wasted labor shelving, removing, packing, and shipping) — is, by far, the single biggest determinant of whether a store succeeds commercially or fails.

The fact is that a single store, almost regardless of the quality of its systems and its management, has to guess at what will be the right books to order most of the time. All advance order purchases (placed by a store long before the book is published) are based on speculation and guesswork. The publisher through a rep or a catalog states expectations, sometimes based on “comparable” titles (which were, of course, published previously at what might not have been a comparable time!) and sometimes based on the publisher’s hopes, all of which are connected to the publisher’s promotional promises. Usually stores — independents and chains — will be guided by those expectations because any other evidence is non-existent.

After a book is published, of course, there is real sales data to guide stocking decisions. But except for the small percentage of the published books that are widespread bestsellers, stores depend primarily on their own sales data to decide what to reorder or keep as backlist. And the cold hard fact is that, for most stores, for most books, there is not sufficient data to tell anybody anything. A single store will have sales for a week or a month across its titles that show multiple copy sales for a small percentage of the titles they have in stock. The majority of the books they carry will sell zero in a week or a month. And, of those that do sell, the vast majority of those will sell one copy.

What’s a store trying to determine what to order next to make of that?

In the past decade or two, there have been a few tools developed to help the independent. The inventory management system Above the Treeline shares information among its stores so each one can get a picture of what is doing well in the aggregate. Stores can see what Ingram, in many cases the supplier from which they buy more than from any single publisher, is stocking and selling, which also provides clues for them.

But the best tool for more nuanced inventory management, for a long time, has been to have a bookstore chain. If you use it right.

This first became evident when the B. Dalton chain of mall stores hooked up computers to their cash registers in the 1970s. Although there were holes in the system, Dalton was then able — for the most part — to tally what they were selling across what was then about 300 stores (a number plucked from memory here, and this is in the neighborhood of 40 years ago, but I think this is right…) When I first started selling to Dalton in 1974, they had two lists of titles they considered worth tracking: a “hot” list and a “warm” list. On the “hot” list was every title that sold more than six copies in the chain in a week. The “warm” list was every title that sold more than six copies in the chain in a month. That’s one copy for every 50 stores to make these lists. And, of course, most titles didn’t make them. Learning those numbers did more to help me visualize how slowly books move than any other single piece of my education to that point.

In addition to those lists, Dalton introduced an automated backlist replenishment system based on “models”, or an ideal maximum inventory level and a reorder point. So the model for a title might be one copy in some stores, with a reorder when it sold, or it might be three copies with a reorder when sales took it down to one. Automating the reordering of “Romeo and Juliet” and “The Grapes of Wrath” enabled valuable buyer time to be freed to figure out what to reorder (or model) from the hot or warm list.

We had a very useful controlled experiment taking place in the 1970s because Dalton had a competitor chain called Waldenbooks, which had more stores and had been around longer but which, for several years, had no comparable computer-assistance for their inventory management. Reordering at Walden was primarily the responsibility of the store managers, and it will be no surprise to learn that some were a lot better at it than others. In the 1970s, it was clear to all publishers that Dalton bought more efficiently (fewer returns) and sold more books, particularly of the backlist.

In the 1990s, Dalton was absorbed by Barnes & Noble and Walden by Borders, the two superstore chains that effectively replaced the mall store presence of the previous decades. And, once again, the chains diverged in their inventory management capabilities. B&N invested heavily in what we were, by then, calling their “supply chain”. They built warehouses to serve as the resupply hubs for their stores and instituted systems by which stores got resupply from the distribution center(s) on a daily basis. They set up rigorous systems to manage models and to review each store’s performance annually. Whereas Borders had A, B, C, and D stores to denote a range of sizes and title counts, B&N had those gradings by store sections, enabling them to configure the stores much more granularly to their local demand pattern.

By the time the 21st century arrived, B&N had built a substantial advantage over Borders in its inventory management practices. Close observers of B&N’s financial reports would have seen that inventory efficiency — how much they could get in sales from their inventory investment — improved every year as they added more stores able to feed off the same central supply capability. Their inventory management cost — buyers to talk to all the publishers and the systems and physcial plant to administer all the stocking decisions and, if one were fair, the cost of returns — was consistently declining as a percentage of sales while Borders’s was not. Borders didn’t make comparable investments in plant and systems and their stores became less and less competitive.

This, even more than any failures in digital, is why Barnes & Noble thrived while Borders collapsed in the latter part of the last decade as sales shifted from stores to online.

The inventory management challenge is one that some publishers have tried to help bookstores with for years. My father, Leonard Shatzkin, instituted the Doubleday Merchandising Plan in 1957, by which the reps walked out of stores with physical inventory counts (there were no computerized inventory tracking systems back then) instead of purchase orders. Those were converted at headquarters into orders which the participating stores had agreed to accept. In the past decade, Random House has developed their own VMI (vendor-managed inventory) system (though my knowledge of details is sparse; this is proprietary information which has never been shared in any detail with me) which it has employed to help manage its books at B&N and is now, apparently, also using at Books-A-Million.

Even if a store knew, title by title, exactly what the right inventory level is (and they don’t, and it changes day to day, anyway), keeping the right books in stock is a challenge. The most sensible way for most stores is probably to order from Ingram (and/or other wholesalers) on a daily basis. Buying from publishers not only requires splitting up orders that will then arrive at different times in different packages, it also requires following ordering rules that are different for every supplier. A wholesaler can’t offer as much margin as a publisher, but the consolidation of the business both makes management cheaper and promotes faster stock turn, which almost always will more than compensate for slightly higher purchase prices.

(I have just made two books by my father available as ebooks. Both In Cold Type, his overall examination of the trade book business originally published in 1982, and The Mathematics of Bookselling, originally issued in the 1990s, contain extensive explanations of this fact with all the accompanying algebra.)

Independent bookstores generically resist vendor-managed inventory. Indeed, picking the books (also called curation) is both one of the great pleasures and great services that a bookstore offers its customers. It is understandably loath to delegate any aspect of that very important work to anybody else.

But the cost of buying is a real hurdle to running a successful bookstore or book department. Buying through what my father called “distribution by negotiation” is expensive for the store, expensive for the publisher, and, unfortunately, doesn’t result in the most productive possible decisions. Things can work much better if it is eliminated.

One example of this that I’ve been involved with is West Broadway Book Distribution. WBBD puts books into national chains that aren’t bookstores (JoAnn Stores and Hancock Fabrics have been their customers for years). WBBD makes all the stocking decisions, based on daily sales reporting it gets from the stores. The stocking decisions are highly automation-assisted, so that a couple thousand titles from over 100 publishers are managed in over 1000 stores with an extremely small staff at WBBD and virtually no buyer time required from the stores. And using a systematic approach means rules can be constantly improved. In fact, WBBD has improved the stock turn on the inventory it places virtually every year.

The cost of buying and maintaining the supply chain is going up steadily (as a percentage of sales) at B&N because they’re reducing their book inventory and closing stores. Soon what has been their competitive edge will turn into a competitive drain. Different supply strategies, such as having publishers ship more books directly to the stores, are already being employed and that will continue. But this a de-leveraging of the B&N core advantage. That’s one reason 2013 could be a difficult year for our last remaining truly national bookstore chain even if the sales of books in stores erodes more slowly than it has recently. A sales decline is very painful to an entity with high fixed costs and B&N’s supply chain is a big part of theirs. And if they were to suddenly close a substantial number of stores, climbing down from that infrastructure cost base might suddenly become much more urgent and very difficult.

As the shelf space for books being managed by retailers that accept the high cost of managing book inventory and commit to doing it effectively continues to decline, publishers need to understand that it will be really hard for non-book retailers to replace them.

Tweet about this on TwitterShare on LinkedInShare on FacebookShare on Google+Share on RedditShare on StumbleUponEmail this to someone

  Back to blog

  • The blending of the ‘science’ and the ‘art’ of book-buying and selling is something we talk about in our office a lot. We like data and the analysis of it, but it is really important that we don’t forget about the ‘art’ component.

    The decrease in book selling venues and the potential increase in ‘maintaining the supply chain’ is a huge challenge for bookstores of all kinds, and ultimately to the publishers that want to sell books through those booksellers. As we heard last week at DBW, and as our own data tells us, bookstores are still one of the best ways to create discovery that results in sales.

    PS. The Mathematics of Bookselling is required reading around here and should be for all retailers.

    • Noah, my own hunch is that the “art” is in the “architecture”: figuring out categorically what to emphasize and what to cut out and strategizing about how to carry extended backlist (which of Shakespeare’s plays do you stock all the time and which do you rotate?) Once you create the plan, you can manage it with automation assistance.

      I’m flattered on my late Dad’s behalf to hear about the fact that his monograph The Mathematics of Bookselling is core information for such an esteemed group as your team. Thanks for that.


  • Matthew

    Very interesting post. I’m in no way a fan of it, but this is the best argument I’ve seen for the Espresso book machine. If only it produced better quality books.

  • Bruce Batchelor

    Of course, the bookstores would benefit from an extra 5% to 10% of margin if they managed their inventory as retailers do in all other industries: by buying on a “firm sale” basis rather than “returnable” terms. Most publishers would happily grant retailers a larger margin (likely 50% instead of today’s 40%) if sales were firm. Booksellers could mark down items they couldn’t sell for full price to as much as 50% off, instead of returning (and paying the return shipping and packaging). If a local hardware store can figure out their buying of ten thousands items (all bought on a firm sales basis), surely a bookstore could do the same.

  • Pingback: One Way Ebook Subscription Services Could Work | Digital Book World()

  • Wow, Mike. This is one of your most informative posts ever. Terrific (though deeply sobering) intelligence!

    On Bruce Batchelor’s view that bookstores would benefit from greater margin if they bought their books on a non-returnable basis: is this sales model one publishers make available to bookstores? I ask because years ago when I worked at Barnes & my boss, the head of trade merchandising, claimed that Amazon was able to offer deeper discounts than because they bought their books on a non-returnable basis and so at a deeper discount. I’d never heard that claim before and haven’t since. Of course, I suppose Amazon has the clout to insist publishers sell them books this way even if the publishers themselves have no formal policy to allow it.

    In the end, though, I get the feeling — and I’ve had it for a while — that publishing, distributing, and stocking physical books is so expensive, and figuring out which (and how many) to buy and reorder is such a black art, that ebooks inevitably become the default regardless of what format the majority of book readers prefer. And with respect to the Espresso book machine, it would have to produce better quality books far more quickly to succeed in the mainstream. It’s hard to imagine someone shopping for the latest Dan Brown waiting 5 minutes for Espresso to print it out — especially if there were 2 or 3 other Dan Brown fans waiting in front of her.

    • Gee, glad you liked it!

      I’ve been out of the middle of things too long to generalize on current publisher policies. I think most publishers sell non-book retailers on a higher discount-no returns model. Bookstores are often reluctant to buy that way. And there’s a problem with it from the pub POV too. Remember that returns on *backlist* are already pretty low. So why should the publisher offer an incentive for backlist to be bought non-returnable. But *frontlist* is unattractive without return privilieges. So if you set it up across your list, you lose margin on the backlist and lose clout on placing advance quantities. Unattractive both ways.

      I think both and are often in a position to “know” a “safe” quantity to buy. But with BN you have the problem of separating (on terms) what the dot com buys from what the stores buy.

      My hunch would be that Amazon’s lower selling places are more about willingness to give up margin than because they can buy cheaper. But that’s just a hunch.


      • Good point: making frontlist returnable is like what we did at the McGraw-Hill Prof. Book Group in the ’90s by direct marketing books 30-day-free trial — providing an overwhelming added incentive to buy. Once a credit card was required DM sales were slashed, as bookstore orders would likely be if nothing could be returned.

        BTW: It’s possible BN has the problem of separating (on terms) what dot com buys from what the stores buy, but hasn’t had its own buyers for years — the store buyers also buy for dot com. (Dot com also used to have its own warehouse, but that was closed years ago as well.) I do recall that one major publisher, Pearson, offered an added “destination discount” (an extra 2 or 3% off) if the entire order was shipped to a single destination, which was typically BN’s main warehouse in New Jersey. But I suppose qualifying orders were mostly midlist or backlist, since major frontlist titles were being shipped directly to the stores.

  • Great article…and fun to learn that my books would be on the “hot” or “warm” list based on their current sales. (The numbers might be a lot bigger if that list was for today rather than the relatively tighter economic times of 1974, buit I’ll take pleasure in my fantasy.)

    • Sue, for all I know they’re still on a “hot” or “warm” list. Glad the piece generated some warm feelings for you.


  • Pingback: B&N Retail Decline Continues()

  • Pingback: Mike Shatzkin ยป Knjigarna Behemot()

  • Eric Riback

    As I remember it, back in the day Borders was such a leader in computerized inventory their BIS was used by a number of leading indies such as David-Kidd. That they let themselves fall behind was just one of the missteps leading to their demise.

    • That’s right, Eric, but that was really before the superstore era. The Borders brothers really developed a great inventory management system. I think the model was to lease the system to indie stores until enough financing came along for them to build out the superstore chain. Their inventory protocols broke down over time as B&N was building their capabilities up.


  • Pingback: Weekend Reading – The Real Advantage « the coffee philosopher()

  • Pingback: this went thru my mind |()

  • Michael

    Thanks for this excellent post Mike. I was privileged to hear your father speak in person many years ago and, as a (not-exactly) former bookseller and bookstore manager, I always had a copy of TMOB on hand – a short, yet seminal work for any bookseller worth their salt. I lived the daily reality of min/max computing and an 8-week sales forecast. Creating inventory turn was always my aim, at the store, category, shelf and title level. But I was a lonely independent, no scale, little efficiency and tiny margins add it all together with a single location and turns only get you so far. Thanks again!

    • Glad you liked it and were able to benefit from the insights of my Dad’s.

  • Aliya

    Good post! Are there ways for bookstores to pick order your novel without using ingram? Say, like through amazon?

    • Bookstores are reluctant to order from Amazon for two reasons. One is that they don’t want to support “the enemy”. But the other is that Amazon is a *retailer
      *and they don’t offer the wholesale prices stores need to make a profit when they sell a book. There are other wholesalers, but since indie books would most likely be printed on demand, probably only Ingram would be set up to deliver an indie book.