In Britain, the trade has been roiled by a different turmoil: the end of the net book agreement that effectively controlled retail prices. This newly-enabled price competition for books has encouraged the UK version of the US mass merchants, the "supermarkets", to expand into bookselling, like their US counterparts, bringing the top titles in the pyramid to a wider audience at lower prices. This is apparently also having a deleterious affect on the British bookselling independents, who have lost important chunks of their business.

For the publishers, particularly if they are not sentimental about whether bookselling is highly personal or highly corporate, this would, at first blush, appear to be only good news. Who wouldn't want to trade little editors for big ones? Who wouldn't prefer to get an order for 1000 stores in one place, rather than chasing around to 1000 different places? Who wouldn't want to put the books with the biggest markets beyond the limited venues where book lovers only may tread? And if the big retailers are willing to work on slender margins to offer books at cheaper prices to the public, why wouldn't the publishers love that?

So, why is it, if all this change is so good for consumer publishers, that so many of them--it seems like all of them--are suffering such dismal financial results? Why are reports of catastrophic returns of unsold books becoming so commonplace? And how can returns be going up at the same time that advance sales for new books, thanks to widespread use of computers and a "just in time" inventory philosophy, are going down?

These questions have a complex answer that can be very simply stated: the inventory in the pipeline is out of control. That is, more than any other player in the game, the publishers problem. And they can fix it: with vendor-managed inventory.

Vendor-managed inventory is not necessarily a panacea. If it is done wrong, it doesn't work. The history of the mass market paperback industry in the US demonstrates that very clearly. It exploded with growth after World War II because it employed vendor-managed inventory, on two levels. The publishers just told the local IDs: here are the titles, here's what you need to take. And the local IDs just put the books into the racks at locations they controlled. No muss, no fuss, no presentation, no negotiation. And, in the beginning, the money rolled in for everybody.

But over time, the system broke down. Publishers tried to jam more books into the pipeline than the pipeline could hold. The IDs, more used to a business of weekly and monthly magazines with predictable shelf lives and pretty predictable sales, stumbled trying to cope with thousands of different book titles that all had different marketing tracks. And, in a fairly short time, this led to a complete breakdown of the system. The larger locations that the IDs once controlled became targets for the publishers to sell direct; the owner of the location felt it could "afford" a buyer, particularly when the publisher offered much larger discounts than the wholesaler did.

So vendor-managed inventory first created the mass market paperback business and then, poorly administered, its disintegration destroyed it. It is no accident that what we now think of as the mass market channel sells much more expensive books; the higher unit cost is needed to support a much more expensive stocking system. And even the rack-sized books are, proportionately, much more expensive and close to hardcover prices than they were forty or fifty, or even twenty, years ago.

Making the vendors manage the inventory was an important component in the growth of Walmart, which Sam Walton made abundantly clear in his book. But the challenge for Proctor & Gamble, dealing with soap suds and shampoo and other items that sell repeatedly to the same customers, was not nearly so great as what publishers will face. And Proctor & Gamble had another advantage publishers don't have: the vendor asked them to manage the inventory, they didn't have to sell the idea.

Inventory management for books, whoever does it, is vastly more complex than it is for any other consumer product. Even superstore retailers in computers or home supply might carry 5,000 different items, or SKUs in retailer parlance. That number might cover the new releases a book retailer has to cope with over a couple of months. As for videos, Hollywood has made a total of something over 10,000 films in all of its history well under half, possibly a third, of the annual output of new consumer book titles. And there are a fraction of the number of recorded music releases as compared to books published.

And with the exception of a very small number of top sellers, the sales in each outlet are spread over a large number of titles. If a bookstore sells 10,000 individual books in a month, the sales will likely be spread over more than 9,000 titles. And the same thing could be true the following month, but 5,000 of the 9,000 titles could be different. So projecting sales from one month to the next, based on any one stores data, is a futile exercise.

And, of course, events influence sales on a title-by-title basis. Some of the events, like publisher-sponsored author tours or advertising, are themselves predictable, even if their effects are not. But many times the sales-catalyst event itself is not predictable, or perhaps not even detectable. A mention in a magazine, a tip in a newsletter, or take-up of an idea by a celebrity or opinion leader can all cause a spurt of sales in a book a store might have been selling modestly or not even stocking at all.

From the store end, the inventory management problem is further complicated by the sheer number of suppliers it generally employs. A superstore will be stocking books from thousands of different publishers, each with its own discount schedule, shipping location, speed of fulfillment, and credit policies.

So the books go out one by one and damn near come in the same way. No matter how much more efficient it would seem to move books around in 5s and 10s rather than in 1s and 2s, most of them won't actually sell that way. The massive returns in our industry are at least partly the result of a conspiracy of wishful thinking: both publishers and retailers would be much happier if books would move in 5s and 10s, but stocking them that way does not make it so. That approach might be made to work for a small portion of the title universe: the top of the pyramid stocked by the mass merchandisers or the relatively small number of titles available in the early days of mass market paperbacks. And it might work for direct response book clubs, who also work with the top titles in the pyramid.

But for the TRADE, the whole trade, selling and stocking everything as if it were all at the top of the pyramid wont work. If you want all the sales, you have to stock all the titles that will sell. That fact is what has driven the transition from mall stores with 25,000 titles to superstores with 150,000. And, let's remember, those 25,000 title stores were and are also characterized mostly by sales of ones and twos, not fives and tens.

Lets spend a few moments considering the economics of a bookstore operation. Bookstores convert money into books when they buy them and convert the books back into money when they sell them. At the end of a year, their profit is based on two things: how much they manage to make on each book they sell and how many books they sell. In semi-technical terms. we'd say that their total gross margin is the product of their average margin, determined by the discount they buy at, times their stock turn, or the number of times they re-use their inventory investment over the year.

Most bookstore costs are relatively fixed. Rent may have a variable sales component, but most of their other costs in staff, systems, and hardware, remain pretty much the same regardless of sales volume. That means that most of any increment in margin will drop straight to the bottom line.

If you look a little more carefully at the gross margin equation, you will see that the product comes from multiplying what will certainly be a fraction, the margin on each sale, by what should certainly be an integer, the annual stock turn. That tells us that the stock turn is by far the more powerful component of the result. Indeed, calculations available on our Idealog.com Web site demonstrate that at normal discount levels, it would take eight additional points of discount to equal the margin and profit impact of one additional turn of the stock. Check it out for yourself.

Bookstores and publishers have been doing battle over discount points for years. For the publishers, giving up discount points is very significant economically. But bookstores can increase their profits more quickly by improving stock turn than they can squeezing discount points out of the publishers.

The other component of the economics of inventory management that affects both the publishers and the booksellers is the cost of returns. Of course, all returned books signal a wasted investment in inventory plus wasted freight and processing charges in two directions, regardless of who pays the shippers. And they too often represent books the publisher could have avoided printing, if they aren't subsequently re-sold through another outlet.

In fact, they are sometimes re-sold through the same outlet. Publisher's customers frequently re-order titles they had once, sometimes very recently.

It happens that poor stock turn and high returns have the same root cause : a buying decision that is overly aggressive and speculative in relation to what turns out to be ultimate demand. The books sit there hurting stock turn until the bullet is bitten and they are sent back.

How speculative a buying decision *needs* to be, of course, is related to often you are willing to buy. If you buy once a year, you have to speculate quite a bit. If you buy every day, you can compensate for under-buying so quickly that the consequences are minimal.

The argument I want to make to you today is that our current buying-and-selling system, which my father Leonard Shatzkin has dubbed "distribution by negotiation", actually is responsible for the prevalence of unprofitable buying decisions and that a vendor-managed inventory system could make these decisions considerably more profitable for everybody.

The reasons for this are different in the independents than they are in the chains.

The independent bookstore is totally subject to the bookstore economics we described earlier. What they can stock is usually more a function of available cash and credit than it is available shelf space. While not every independent retailer may realize it, his prosperity depends almost entirely on stock turn. But the independent has two handicaps that make it much more difficult to select books at the individual store level than it would be for the publisher.

First of all, even if a single store captures data on every book it sells , it simply does not know enough to project sales meaningfully on most titles. If 8,000 titles each sold one copy last month, of which 4,000 didn't sell a copy the month before, which ones do you re-order? The sales evidence is so flimsy that some books that didn't sell any copies are more worthy of reorder than some that did.

Which leads to the second point: even if the decisions you would make would be good ones, how can a bookseller afford to spend the time to decide? Discerning subtle differences in sales appeal among thousands of titles that sell the same insignificant numbers of copies is a time-consuming pro position, even if it were possible to do it accurately.

The clues that would actually permit distinguishing among those titles, of course, are not revealed on the level of one single store. If your store was the only one in the country that sold one particular book but *everybody* had sold a copy of another book, that would make it pretty clear which one should be re-stocked. That knowledge constitutes the publishers' advantage, or the sales reps' advantage, or the wholesalers' advantage, or the chain stores advantage in making these decisions. It is the solo independent stores' albatross.

It is worth reflecting for a moment on how much money cash-starved independent stores spend meticulously tracking information that can ultimately be of little or no help in making good buying decisions.

Only some of the costs in the system are reflected by slow turn and high returns. There are also substantial costs in lost sales which are easy to visualize. When a store fails to re-order a book that is selling everywhere, both the store and publisher lose sales and sales momentum that may never be recovered.

But, having said that, we have some tall explaining to do. Because bookstore chains, which do know something about how a book is moving across many stores and which can spread their title reviewing costs over many stores, actually achieve much poorer buying results than well do well-run independents. How can that be?

We have been calculating stock turns for every store or sub-category of a store for which we could get numbers for years and years. Many independents achieve turns of 4 or 5 a year; almost all of them that stay in business for any length of time turn their stock 3 times a year. Although chain numbers are usually quite difficult to discern, we have consistently calculated chain stock turns between 1 and 2 a year. Well-run independents return 10% or 20% of the volume of their buys; chains generally return 30 % or 40% or more.

As a recent Wall Street Journal article was, I believe, the first to clearly articulate, it is precisely because the business is increasingly shifting from 10% returns customers to 40% returns customers that the economics has suddenly become so much more difficult for publishers. This is a shift that cant be made up for with increased volume. If the returns continue at their historical levels for this segment, and they appear to actually be rising, more business just makes matters worse.

A leading consumer publishing executive described this reality recently when he said, in effect: "It used to be that our bestseller business paid for everything else we do; now it is the other way around."

Until pretty recently, the difficulty of achieving efficient buying results at the chain level could reasonably be attributed to the sheer clerical and administrative burden of the task. Start multiplying tens, let alone hundreds, of thousands of titles by hundreds, let alone thousands, of stores, and you literally have tens, if not hundreds, of millions of individual stock levels to track. That problem is made apparently manageable by aggregating the data: the buyer first looks at how a title is doing in the aggregate in order to decide whether to even look at how it is doing store-by-store.

Unfortunately, aggregating leads to predictable buying errors that occur every day. A buyer sees he's bought 4000 copies of a book for his chain, and in the first week he has sold 240 copies. Not enough, perhaps, to be worthy of a more detailed look. But lets say a couple of cities, perhaps places that ran early reviews or where the author has stirred up some interest nobody really knew about, have sold out.

So the next week sales go down to 150 copies. The buyer sees this book as one which is "slowing down". It might be that it is actually "picking up"; the two sold-out cities have been joined by a couple more. If the book had been replaced in the cities in which it is selling, sales might appear to be accelerating rather than declining.

This is a lesson in the dangers of over-interpreting from aggregated data that will be familiar to any rep who has sold a major account.

But pleading too much information to digest is a poor response in the digital information age. And while there are some obvious computerized fixes to the problem we just considered that I don't think have been implemented in any bookstore chains procedures, bad systems are not the only barrier to efficient inventory management in the chains. Lets reconsider the bookstores economic realities from the chains perspective.

The chains, particularly the biggest ones, see themselves in a market share game. Because he is painfully aware that books that sell big numbers are the slender minority of the titles they stock. the biggest nightmare of any chain buyer is that a book takes off and they don't have enough of it and can't get more quickly. This is even worse if the competition does while they don't.

And while the individual retailers constraint on how many books it can stock is usually available cash and credit, chains play that game by different rules. Their cash is coming from Wall Street, which appears to respond to the top line, sales and market share, more than to the bottom line. And their sheer size makes it possible to stretch payment to the publishers.

The equation for a stores profitability has "dollars of inventory investment" as a key component; we must remember that the investment only REALLY passes from the publisher to the store when the store pays for the goods. If a store pays its bills 150 days after it receives shipments, all the books that sold in the first five months had a turn of infinity, and the only investment the store need worry about is for the books that take more than five months to sell.

So both the market share consciousness and the ability to pay more slowly could diminish the chain stores sense of urgency about stock turn. There is another differentiating factor working against the chains concern about returns.

Returns cost stores money, as we discussed before. But if a store pays slowly enough, that cost is limited to the actual cost of processing the books and shipping them back. Inventory cost--which could also be viewed as the opportunity cost of lost sales or lost interest--may be cancelled out as a factor. And offsetting any returns cost for the chains is their stake in the remainder business. The silver lining for retailers that sell remainders is that the books they ship back to the publishers today w ill return to them at remainder prices tomorrow. And because of their size and merchandising ability, chains have the best opportunities to profit from remainders.

But trying to figure out WHY stock turn is so bad and returns are so high is speculation. What is not speculation is the fact that these are the facts, they seem to be getting worse, and publishers must do something about it.

That "something" is moving to vendor-managed inventory.

This is not the forum to discuss exactly how to build an inventory control system from the publishers end, or, as we will discuss briefly in a moment, from the wholesalers. There are a variety of ways to operate such a system, but first a publisher needs to re-orient his thinking to use any vendor-management opportunity correctly.

That begins with a clear understanding of what the objective of each titles stock level is. For titles that are moving, it is to have sufficient inventory to have the book in place, perhaps in various places around the store, without running out of stock. How many copies that is clearly depends on how fast the book is selling and what its display opportunities are, of course.

But it also depends on how fast you'll spot movement and how quickly you can replenish supply.

It is those last two points that change dramatically in a vendor-managed situation.

Being able to spot movement quickly and inexpensively is the new opportunity afforded by technology. A publisher receiving sales and inventory information via EDI, could get it and analyze it as often as the store would agree to transmit it, perhaps weekly or even daily. Before EDI, it was only as often as a publishers rep could get around to count the books, seldom as frequently as monthly.

Of course, publishers COULD get EDI data daily or weekly, or even hourly, without converting it to use with a vendor-managed inventory situation. They could just get it and analyze it and digest it. But this begs the next question: how fast can you replenish supply? In a vendor-managed situation, that is purely a logistics question. In today's world, it is a sales and bureaucracy question. Taking data to a buyer, at BEST, causes delay, even if the buyer acts quickly in the way the publisher would want. Too often, it results in no action at all or even some other wrong action.

The publishers *only* defense in today's world is universally employed: try to sell in enough copies to compensate for the time it will take to know you need more and to convince the buyer to order more. The problem with that defense is that, when it succeeds, it puts more copies in the retailer than even the publisher would think was prudent, if the publisher were operating with the right to restock the title whenever it was necessary.

And that, in a nutshell, is why vendor-managed inventory would increase the frequency of ordering, which would in turn, almost inevitably, increase stock turn and reduce returns.

Let me enumerate what I think are the key steps for any publisher seeking the vendor-managed inventory solution to the inventory problems that are plaguing consumer publishers.

The first step is to become more conversant than most publishers are with bookstore economics, so the relative impact of turnover improvements and discount changes are well understood. This is necessary to even speak the language of vendor-managed inventory properly. When you are negotiating for an overall inventory level based on overall performance, rather than for 10 copies of a forthcoming book instead of 5, your sales pitch will be rooted in historical numbers rather than prospective words.

The next step is to get a real handle on turnover and gross margin calculations, which, in my experience, no publisher really has. Calculating stock turn and return on inventory investment, for a store, a chain, a store section, or for a particular promotion, have to be routine exercises that every rep can do. I would hasten to add that reps will not actually be DOING these calculations; for the most part, computers will. But in the world of vendor-managed inventory, everybody in sales needs to understand them.

The third step is to analyze results in those present accounts that will provide sell-through data via EDI. Any store which is not achieving a stock turn of 3, an annual returns level of 15% of purchases, or an annual return of $2 for each dollar invested, is a likely candidate to have its economic performance dramatically improved by vendor-managed inventory. That will be almost every account. Comparing stock turn levels by store will reveal a great deal about where your sales opportunities really are.

What your vendor-managed inventory system will do is to react quickly to actual sales for a title, based on its history in the account, its history in the universe, and what you know about its future based on seasonality, promotions, and so forth. For titles that have not appeared in a store, you will be looking at the stores relative success with related titles. For example, a store that is achieving a higher turn for cookbooks than travel books should probably have its cookbook selection broadened before its travel book selection, which might even be shrunk to make room for more cookbooks to achieve a better overall result.

Of course, stores no longer holding the order pen will need a different way to control the vendors managing their inventory, or they've let the fox into the chicken coop.

We'd envision that the stores control over the publisher would come from limiting the total amount of inventory, while keeping an eye with the publisher on the economic results of the program. Obviously, a publisher that can't achieve better margin results from an equivalent or lower inventory investment will find its program eliminated, or its inventory levels cut back. But from what we've seen of publisher inventory levels in relation to sales in many stores, publishers could find it very easy to reduce inventory without losing sales, once they absorb how easy it is to replace stock when they don't have to do battle with buyers.

Of course, everything we have said about the publishers opportunity to control inventory for the retailer could be said for the wholesaler as well. They share the ability to view a title across a wide range of retail locations and they offer the added advantage of being able to provide all of the titles a store needs, which no publisher can do.

In fact, there have been short-lived stabs at wholesaler inventory management. The Paperback Booksmith and Little Professor chains of the 1970s, and then the beginnings of Borders Bookstores with Book Inventory Systems, were built on it. These systems were very successful, for a while. For a long time, stores that had started with Booksmith or Little Professor systems were among the most useful and successful independent bookstores.

The problems were partly created by the wholesalers need to cut discounts to create margin to operate what were expensive systems. Even though we now know that turnover and inventory efficiency can compensate for a lot of lost discount, stores were understandably tempted to go around the system when publishers were offering as much as 10 more points of discount to buy direct. And once that process began, one of the chief attractions of having this service provided by a wholesaler, the simplicity of consolidated supply, went by the wayside.

As far as we know, there have been no serious attempts to install vendor- managed inventory programs, by publishers or wholesalers, since the information revolution of the 1990s has made it easy to follow sales and inventory from a distance. Technology today also enables much more complex and timely algorithms to calculate appropriate stock levels.

Of course, any publisher or wholesaler offering vendor-managed inventory will have to be prepared to account for information that might come from the retail level. It may be that the STORE will know about a sales-affecting event or a local school adoption. It is certain that "special ordering" capabilities for books customers are willing to wait for that aren't at an instant moment in the store, will continue to be required.

Building a truly automated inventory management system, calculating optimum inventory levels from store cash register data, overall sales data, and pending promotions, as well as taking into account the publishers supply and reprint situation, is a challenging task. But achieving it will payoff handsomely. The most obvious benefits are that sales will go up and returns will go down, but these might not be the most important benefits.

A publisher with such a system in his selling arsenal is suddenly able to put books where there are no book buyers, a critical tool in an era where publishers are trying to grow their sales outside the book trade. And the ability to "meter" books into the supply chain can be a substantial ad vantage if there's a temporary stock shortage waiting for a reprint or responding to a publicity break.

And if a vendor-managed approach were to become standard practice, an array of costs related to catalogs and sales conferences could be reduced or eliminated. Publishers would look at spending on sales representation as an investment to be recovered by improved sell-THROUGH, not sell-IN.

For most of us, the fear of serious consequences usually provides more motivation than the attraction of potential rewards. We are publishing books in a time when the consequences of the status quo could not be more dire; the storms are washing so much water onto the boats that a sea-change is desired by almost every consumer publisher. The change must come from the publishers, because they are the ones who really pay the price for the inefficiencies of the current system. No matter how promptly or slowly the store pays, no matter how much publishers try to punish accounts with high returns, no matter who pays the freight to ship product around; the fact will always remain that the books that don't sell are, most of all, the publishers problem.

Vendor-managed inventory is complicated to tackle and in the beginning it will be hard to sell. Many accounts will be nervous about it, and so will editors who are asked to believe that smaller advance placements--which would be the case with almost all big books--will not doom the ultimate sale.

But the alternative is to continue as we have been, selling each order one by one in a process that has become increasingly expensive and decreasingly effective, to the point that squeezing out a profit has become nearly impossible for most publishers OR booksellers. Change has to come, and a distribution chain that features millions of decisions too small to be worthy of the time it takes to discuss or consider them, but yet which must be both considered and discussed, seems like the best place to start.