Auletta exposed very important facts that had been kept well hidden. His research and interviews made it clear that virtually everybody in consumer publishing is tottering on the edge of unprofitability, and everybody includes the booksellers as well as the publishers. His cogent examination of sales statistics provided by Barnes & Noble revealed that more than half the consumers dollars go for books that are more than a year old, despite an almost entirely frontlist-loaded marketing expenditure, which includes a frontlist-loaded returns burden, from publishers and retailers alike. He made it clear how little of the sales at Barnes & Noble -- less than three percent -- come from hardcover bestsellers.

Auletta describes a very ill book trade in his piece, but he fails to locate the causes of the ailment. They are really not obscure.

Big publishers are hurting because their book inventory, which they ultimately own until it is bought by a consumer, is out of control and because they have responded to that fact with some counterproductive strategies: the two most prevalent and damaging of these being to reduce the number of titles they publish and to push away more and more of the small-shipment reorder business to the wholesalers and to the distribution centers of the chains.

Big booksellers are hurting because they have adopted a stocking and merchandising strategy that grows market share at the expense of margin or efficient use of cash, which has huge built-in operating costs maintaining warehouses that require many of the books it sells to be shipped twice (and many of the books it DOESN'T sell to be shipped four times). And booksellers of all sizes are adversely affected by competition from mass merchants, who tend to sell only the best-selling books and drive the perverse reality of our business that our most desirable merchandise is sold with the least margin for everybody.

Wholesalers are hurting because the chains are trying to substitute for them with their own warehouses while the independents gasp for customers as the chains expand all around them.

The picture Auletta paints provides depth and clarity to the situation we sketched at Vistas summer conferences in NY and London. Our view of the solution hasn't changed either: it is vendor-managed inventory.

We will come at the subject of vendor-managed inventory for book publishing today from four perspectives.

First, well take a look at our consumer book supply chain.

Second, well look at the VMI solution from the publishers perspective.

Third, well try to demonstrate why retailers should *demand* publisher VMI if the publishers wont do it themselves.

And last, well examine the wholesalers role if the publishers and retailers cooperate to improve the supply chain in the ways suggested in this mornings examination.

The supply chain in the United States has changed drastically in the past quarter century. Up until the early 1970s, most books were shipped from a publisher's warehouse to a store's shelf. Wholesalers provided reorders on bestsellers, usually locally, and a network of wholesalers provided libraries. This was in a time when publishers were much more fragmented than they are now. Not only were the large ones not nearly as large, relatively speaking, but there were far fewer combinations of publishers for shipping or billing or selling efficiency than there are now.

The first great change in the supply chain, which we have discussed in these Vista forums before, was the rise of The Ingram Book Company as a trade wholesaler. By using the microfiche to let stores know exactly what titles were actually available, combined with admirable efficiency of operations, Ingram increasingly made it possible for stores to get most of their required stock quickly and reliably. And they offered stores the ability to combine many publishers books on one order, a more significant advantage 25 years ago than it is today.

Publishers were very happy with this development. Because Ingrams buying practices were sensible (they represented many titles, ordered frequently, and made relatively low returns) and their service to retailers was impeccable, it took the pressure off the publishers to ship quickly, which in the days before Ingram very few of them did. This is another thing that has changed since then; ironically, some publishers are really quite good at prompt shipping now.

And because fast and efficient reordering was even more important to a stores profitability than the stores or Ingram appreciated, Ingrams growth fueled a growth in independent book retailing, right alongside of the growth of the two big mall chains, B Dalton and Waldenbooks.

Publishers reps found themselves calling on an increasingly powerful independent network to introduce the new catalogs of books each season and, increasingly, the stores would go to Ingram or other wholesalers for the reorders. As computer assistance became available to more and more stores for inventory control, publishers became increasingly willing to trust to the stores systems and the wholesaler network to continually resupply books in the marketplace; their effort became more and more focused on the "launch".

Meanwhile, the chains continued to do most of their business direct from the publishers to the individual stores. Walden opened a warehouse in the mid-1980s, but a relatively small percentage of the total business was pushed through the warehouse. Although publishers did not SELL the branches of chain stores independently, they continued to fulfill to them directly. And they often knew the returns situation by store. When Waldens warehouse began consolidating returns in the 1980s, it began the march to where we are today, making it almost impossible for publishers to know from which particular locations returns came from.

Then, in the past ten years, B Dalton and Walden became Barnes & Noble and Borders, and both of them adopted a central distribution center strategy for restocking their stores, further reducing the knowledge publishers have of what is selling at the store level. At the same time, some publishers have cost-accounted their shipment activities and have seemed to discover that small shipments are unprofitable. Some have apparently made the leap to decide that small accounts are unprofitable.

What the experts in "activity-based" cost management failed to factor, however, as they accelerated the movement toward our current distribution malaise, was that lost control and the extended pipeline also have a cost. We have reached a situation where it is actually impossible for a publisher to know how many stores in Cincinnati or Denver have or have ever had a book, because they really don't know how the proliferation of distribution centers is distributing them. In a situation that could be considered comical if the financial consequences weren't so tragic, the voluminous data collection from store cash registers is resulting in less and less actual knowledge by the publishers about what books are selling where and when in the pipeline.

And the ratio of books in the pipeline, which we define as "having been shipped" from the publishers warehouses, to the books actually on sale, which we define as "on a retailers shelf" appears to keep rising. Does anyone doubt that a book on sale in a store is statistically far more likely to sell than one in a warehouse and not available to the public? What can possibly result from a growth of less-likely-to-sell books in our pipelines? Increased returns might be unwelcome, but they could certainly be anticipated.

A senior executive in a major American house recently told us that what sales was about today was to "fill up that pipeline and hope the book sells. If you keep that pipeline full, you're doing your job." The pipeline he refers to now begins with five Ingram warehouses, four Baker & Taylor warehouses, four Borders warehouses, four Barnes & Noble warehouses, and includes a not insubstantial number of additional warehouses for other wholesalers and book chains.

As we mentioned earlier, today's American book marketplace requires that almost every book we sell to a consumer must be shipped and received twice. The growing number of books the pipeline throws back must often be shipped four times. And those that don't sell that are shipped only twice are, ironically, the most wasteful component of the inventory because those are the books that went from the publishers warehouse to a distribution center and came back without ever having gone to a retail location!

So a series of rational business decisions: by publishers to avoid small shipments they saw as costly, by publishers to minimize their own warehouse stocks and carry less inventory, by retailers to respond by demanding better service from wholesalers, by wholesalers who responded to the market by providing it with growing inventories of the publishers books, by chain retailers to seize control of their own distribution from publishers apparently unable to provide the level of service that was required; all have combined to deliver a totally irrational supply chain from publisher to consumer.

And this irrational supply chain is a drag on the entire industry: publishers and booksellers, of course; but also everybody they depend on and who depends on them: authors, book packagers, printers, and the media in which books are promoted and publishers advertise.

What would be a rational supply chain? This is a question with thousands of better answers than the one we are working with now.

Of course, it is easier to imagine a rational supply chain than it is to describe how we'd get to one from where we are. It is possible that the most rational supply chain we could design would entail having publishers give up the distribution function entirely, in favor of consolidation of that function which would, in turn, permit some geographical dispersing of the books under publisher control. But because we want to describe something that can be got to from where we are, we will start with the assumption that publishers will have warehouses, most sensibly one warehouse for each shipping and invoicing entity, so that there is one primary inventory for each title and the publisher has it.

What should happen then is that the publishers warehouse should directly supply every retail location large enough to allow sensible-sized shipments frequently enough for effective inventory control. It would seem reasonable to assume that 25 assorted books is a minimum efficient shipment from a publishers warehouse and that biweekly shipping would provide enough control to get good stock turn and keep returns low. Since peak periods could require shipping some multiple of our minimum of 25 books or inserting an extra shipment between the regular ones, this would translate into accounts that sold approximately 1300 books per year, or probably billing about a minimum of $10,000 per year, receiving direct regular shipments from the publisher. Larger accounts could be shipped even more frequently and efficiently.

It is the size of the publisher as well as the size of the account that affects whether these "minimums" can be hit. What we're suggesting is that each publisher-point of sale combination that can be efficiently shipped direct this way should be. Doing so minimizes the number of books in the pipeline but not on sale, a practice that will reduce returns. And it also allows these books to be handled only once by the supply chain on their way to sale, reducing costs and adding margin for the publisher and bookseller to share.

The wholesalers role in the chain would be to supply the books to almost every store that come from smaller publishers and to provide the books from almost every publisher that go to smaller stores.

Of course, reorganizing the supply chain to this rational form theoretically would not require vendor-managed inventory to accomplish. And as we explored at some length in last summer's speech on VMI, the principal benefit of vendor management is to cut out the too-costly time and effort expended on all sides making small-consequence stocking decisions that can't pay for the time and effort, while, at the same time, it improves the overall quality of these decisions.

We have tried to show that the current inefficient supply chain is a consequence of rational behavior within the irrational distribution-by-negotiation system we have had for books throughout the 20th century. The converse is also true: movement toward VMI will not only increase inventory efficiency, it will be a catalyst for change to a more rational supply chain as well.

The fastest and most direct way to get to vendor-managed inventory, and to a healthier supply chain, would be for the publishers to take the bull by the horns and just do it. On this score, we are encouraged. In the few months since we last discussed this subject, there is some indication that some significant players are beginning to think about it. From conversation and from reading between the lines of public information, we can see movement toward VMI at Simon & Schuster, Random House, and HarperCollins, at a minimum.

In a related move that could also prod the book trade, the country's largest retailer of fabric and women's hobbies, Fabri-Centres of Cleveland, Ohio, is consolidating its book resourcing through The Butterick Company, a pattern company, which will administer a VMI program to supply the books. Since Butterick will be getting the books for this distribution from publishers, some of whom sold directly to Fabri-Centres and other accounts in this class of trade before, those publishers will have an opportunity to view firsthand how VMI affects sales.

Any publisher that attempts VMI for any account will want to start by creating benchmarks for the efficiency of the inventory going in and out of that account right now. Both the publisher and the store will want to see that sales levels are maintained or increased, that the books that are selling are in the store in quantities adequate to meet demand. And they will want to see inventory levels and returns reduced at the same time.

We have been engaged in such a benchmarking effort for one of our clients on inventory in one of Americas bigger chains. We found that current practices were yielding a pretty typical stock turn level for a chain operation of about 1.5. But delving deeper, we found that individual stores in the chain, being supplied by the chain warehouse, had annual turns as high as 40, 50, and 60. These levels are astronomical -- too high. They indicate frequent out-of-stock situations and significant lost sales for both the store and the publisher. So we have the worst of all worlds: a consolidated stock turn, including the books in the distribution centers, that is too low and, with the slow-moving stock in the pipeline that turn indicates, leads to high returns. At the same time, many of the points of purchase are starved for books.

Because this publisher has three divisions, each with separate sales staff working with this chain, and we see the same pattern across the three divisions, we believe it is typical of what most publishers would see if they did the same analysis.

When publishers do the analysis of their inventory productivity in their major accounts, they will see something else that they will want to correct: midlist titles, those below the top attention level reserved for the books with high author advances, big printings, and big initial shipments, sometimes sell out cleanly and are not reordered. Why? Because nobody really notices. Books that are distributed to only a portion of a chains stores often don't achieve exciting gross sales numbers, even if they are performing well in the locations they are in. We have detected instances like this in our analysis and even had the sales department say, "yes, but there's nothing we can do about it."

In other words, in a sales rep-buyer relationship built around what is shortly forthcoming and what will be done to promote it, there is little room for discussion of what was previously sold at a lower level of expectations and has exceeded them. Not only is the selling dialogue not structured to enable this conversation, the publishing processes aren't structured to handle this kind of opportunity.

But think about it. As often as publishers are proven wrong with high expectations that result in unwelcome returns, logic and the law of averages suggest that sometimes they must also sometimes be proven wrong about low expectations that are exceeded, and that should then result in a whole new push in the marketplace. Every publisher can think of a time or two when this has happened; analysis of real sales data reveals it doesn't happen nearly as often as circumstances would justify. And without VMI, it takes pushing against the inertia of a frontlist-oriented system to accomplish it.

Looking at the stocking decisions from the perspective of running VMI from the publishers end, particularly in an account that provides data weekly via EDI, will change the publishers whole perspective about the quantities that should be placed in advance of publication. Making shipments biweekly, as our ideal supply chain scenario suggested, means that four or six or eight weeks supply of any title in any location would avoid almost all lost sales to out-of-stock situations.

Of course, examining the data reveals very quickly that for the lions share of titles -- more than 90 percent of them -- four or six or eight weeks supply is still exceeded by one copy. That is why achieving good stock turns on the few titles that sell more quickly is so essential to good overall performance; the high sales and stock turns on the bestselling titles literally support a longer "tail" of titles that sell less well and cannot be turned quickly because they sell slowly. But they sell. Titles that sell a copy every eight weeks or sixteen weeks can still be selling many thousands per year, if VMI is used, in a way that is profitable for the stores and the publishers. And metered in the way VMI encourages, these sales will be achieved with minuscule returns.

Operating with VMI, every publisher would find it possible, within today's inventory levels, to stock more titles in more stores, have fewer sales lost to out-of-stocks, and boost sales substantially on midlist and bottom-of-the-list books that are now doomed by their low initial expectations.

Although it would require more frequent smaller shipments, theoretically more expensive, to implement such a program, it is not at all clear that overall fulfillment expense would go up. For one thing, the predictable shipments resulting from VMI are more economical than the unpredictable demands placed by customer-ordering. And the reduction of returns achieved by not having to speculate well into the future on sales will generate a substantial reduction of demand on these capabilities. And being able to plan and schedule returns processing, which VMI permits, also makes it much less costly.

And that is before the money the publisher saves not printing books not ever needed and it is also before the extra money the publisher earns on sales of titles now prematurely killed by the distribution system.

Let me give you some real-life evidence of today's distribution systems hidden costs. When we started getting the store data for our client, we set out to create a report that would be of immediate use to sales management. So we started by computing what we thought would be a weeks supply for each book, based on its sales pattern over the previous eight weeks, figuring then we'd isolate those books whose inventory fell below a certain number of weeks supply. THOSE would be the titles the sales rep would focus the buyers attention on for a reorder.

So we asked the Sales Director whose division was going to get this report: at how many weeks supply do you want the trigger? He said "Tell me all the books over a threshold sales level that have fewer than TEN weeks supply." He was figuring on the inevitable delays: getting to the buyer, getting the buyer to act, getting the action to show up as a purchase order, and then getting his company to ship it. As we said earlier, in a VMI situation, we'd be comfortable that four or six weeks supply would prevent lost sales. Here, even working with a special newfangled reporting tool to boost efficiency, we need twice that much inventory in the pipeline to avoid lost sales.

No publisher we know is equipped today to take on the responsibility for inventory management and to live with the consequences of the results with permissible inventory levels raised for good marks and lowered for poor ones. To develop new skill sets, which will be required, as well as to turn around the misguided thinking that it is more profitable to have somebody else handle the smaller shipments, will require an appreciation for the increased sales and reduced costs that can result. But with at least three major publishers beginning to contemplate the possibility, others may soon have the opportunity to learn from their competitors efforts.

At the top of this speech, I mentioned that list reduction was another counterproductive strategy for publishers groping for a way out of the current difficult industry conditions. What publishers practicing VMI will find is that the "bottleneck" they see, where good titles get choked off from their market, is not due to too many titles at all but to poor management of already available shelving opportunities. It may sometimes appear working with today's supply chain that more time and effort per title is what is needed to turn failure into success. Since much of that time and effort is spent selling within the supply chain itself, that equation is turned on its ear by VMI. It makes far more intuitive sense that you grow a company by increasing its product output, not by cutting it.

In fact, my own personal selection of the most frequently-repeated mistaken idea in publishing is "there are too many books published". That's not our problem. Our problem is that we have to put every book we publish into too many warehouses.

I fear sometimes that the retailers will look at my advocacy of VMI as an insult to their ability to make good stocking decisions. Actually, it is not intended that way at all. If it were possible to buy books effectively for multiple locations from myriad publishers, somebody would have done it by now. It is not for any lack of intelligent effort or dedication of resources that the problem has not been solved. Big chain bookstores have existed for nearly three decades; before that department stores had multiple branches stocking books, presenting a parallel challenge. Consistently high stock turns with minimal sales lost to out-of-stocks along with low returns have never been characteristic of these accounts.

But there is another point here about publisher VMI that retailers also have to appreciate. Even though our industry shies away from calling what we do selling on consignment, from an inventory responsibility perspective, that is what we do. Unsold books end up back with the publisher. Overstuffed accounts also frequently pay slowly, so the unneeded inventory can represent a cash drain on the publisher from start to finish.

Stores occasionally offer, or ask, to buy on an increased discount and no returns basis. But what do they offer to buy that way? Never the new fiction, that's for sure. They want to buy the backlist on a no-returns basis. That's great, except there should be low returns on the backlist anyway. Stores and publishers tacitly admit that new books are a risk until we gain some experience with them in the marketplace. Then they might establish themselves as being low risk, or in the case of Dickens or Shakespeare or Hemingway, perhaps eventually become no risk. Keeping those no risk titles moving through the system is important, but they constitute the minority of titles in any thriving bookstore at any time.

The retailer performs critical functions in the supply chain, maintaining the overhead and infrastructure necessary to present books to the public. In recent years, chain retailers have been highly skilled at drawing substantial investment capital into the industry to build more and more large and beautiful stores to sell publishers wares. That is their job, the retailers should focus on it, and publishers should support them in doing it.

What the big retailers have established pretty clearly over the past decade is that publishers are responsive to their desires. Sales organizations and catalog creation practices have been reconfigured to address the needs of larger accounts that buy centrally. Promotion and advertising expenditures by publishers have been redirected to suit the big accounts preferences. Discounts have been elevated and, in a particular irony, most publishers even have conceded additional discount points to support retailers distribution centers, even though our analysis would suggest their very presence actually costs publishers sales and adds to returns.

Meanwhile, the retailers are growing the support overheads they must maintain. They have large buying staffs to make the stocking and restocking decisions. Those staffs are overtaxed; they simply cant make enough decisions to keep stock moving efficiently. As we explained in last summers VMI speech, the huge number of decisions requires the buyers to aggregate data, to look at how a title is doing across a range of stores rather than store-by-store, to keep the number of analyses and decisions required barely manageable. The retailers cant afford any more overhead devoted to buying; the publishers cant afford to live with the amount of review and care the current expenditure on buying will permit.

This illustrates another point. We mentioned earlier that publishers were seduced by store computer systems, figuring, I guess, that because a computer recorded each book sold, what should be reordered would be reordered. This leap of faith forgot that the computers wouldn't make the reorder-or-not decisions, particularly for those books that had just been published with just-in-time-sized advances hoping for reorders. The data could assist the buyers, but only if they had the time to review it. Publishers cant trust to the systems; they have to invest the time to help store buyers whenever they can, even in accounts that wont today cooperatively share the data.

The retailers are now taking on the additional overhead and logistical challenges of running distribution centers. Increased discount from the publishers apparently subsidizes this effort, but that appearance is an illusion. Publishers must respond to shrinking margins somehow; they have done it by raising retail prices and cutting back title output, neither of which helps the retailer or, indeed, anybody in or depending upon this supply chain.

Of course, retailers didn't open distribution centers for their health. They did it because publishers didn't provide the service they needed for inventory management. In order to make operations manageable at the level of the individual store, it seems sensible to take on the overhead of a distribution center to meter in the stock, particularly if the publisher can be persuaded to part with the margin to support it.

In a recent Publishers Weekly examination of "just-in-time" inventory, David Cully, the president of Barnes &Noble Distribution, says the chain has devoted, in PW's paraphrase, "millions of dollars to its own replenishment infrastructure". Cully admits on Barnes & Nobles behalf, "We are bearing the cost of Just in Time" and he suggests that it will take further evaluation for B&N to decide whether the rewards to them are worth the costs. A model Cully suggests for the future is that the chain will handle the top 30-to-50 thousand titles from their distribution centers and have publishers directly supply the rest.

But there is another approach that hasn't been tried. Retailers could say to publishers: if you want more than the rock-bottom minimum of books in my stores, this is what you have to do: ship each location once every week or two (depending on store size); stay within the inventory ceilings we assign you; keep the proven backlist titles always in stock; and use your available dollars to give us the best possible sales with the rest.

Then the chains buying staff would no longer have to worry about hundreds of thousands of titles, with hundreds of new ones each week. It would have to manage twenty or fifty larger publishers, on the basis of total results, adjusting permissible inventory levels to suit actual results. It would worry specifically about only a few titles, those few that are current fast-movers or firmly established backlist, just making sure that the publishers were following through on the opportunity for everybody to keep those books moving. And the smaller publishers should work on the same basis through a wholesaler providing the VMI. Wholesalers might also fill in on the fastest-moving "cant wait" titles that run out of stock between publisher shipments.

In fact, we know of recent cases where pretty significant accounts, one a wholesaler and one a chain, have asked one or more big publishers to administer VMI for their books. At least one of the major publishers has a couple of experiments of this kind going.

When a chain retailer throws the challenge to a publisher in this way, we'd suggest starting with a 10- or 20-store experiment to let the publisher learn what's involved. The publisher will quickly find itself guided by sensible incentives to do the right thing. We've already seen that it would take fewer copies of big books to assure keeping them in stock, if the publisher knew more could be shipped in as soon as it was necessary. That would clear "space" to keep newish titles not yet established as backlist moving through the store. The inventory ceiling would make publishers very sensitive to the fact that each book in the store was keeping another one out, and that producing sales from the books that were in was the way to get the inventory ceiling raised.

Obviously, having the supply chain function in this way cuts out a lot of the existing wholesaler business. They could theoretically continue to resupply bestsellers between publisher shipments, but those opportunities would really be few and far between. One of the great misunderstandings bred by our current distribution chaos is that travel time is a serious obstacle to keeping inventory replenished. Only a sliver of the titles in any store sell more than a copy a week. Publishers who are stocking four or six weeks supply of their fast-moving ones and resupplying every week or two will very seldom be out of stock. There would be some exceptions, but not enough to build a business on.

The very smallest publishers will almost certainly need a distributing major publisher or distribution company to provide competitive supply. Smaller accounts would need wholesalers to distribute the smaller publishers or distributors. And the wholesalers would also have to provide VMI, or find their books hopelessly underbought in relation to the publishers supplying directly and using VMI.

Running such a system might require the wholesalers to increase the spread between what the publishers give them in discount and what they normally offer their customers. One way to do that, of course, would be to get the publishers to give them more. Providing a more significant service, which they would under this scenario, perhaps would make that make sense. Of course, the more the publisher has to give away, the more easily the publisher can justify direct-supplying the store as an alternative.

The other way the wholesaler could increase spread is cutting the discount to the accounts. This might not be as difficult, or as unpalatable to the accounts, as it might at first seem. The profit at the retail level is gross margin on sales times stock turn, minus the cost of doing the business. The wholesalers, by providing good stock turn, low returns, and minimal need for buying administration, could be delivering business profitably to retailers at lower discounts than they or the publishers provide today.

If the retailer looked for the same profit performance from direct-supplying publishers as from wholesaler-supplied publishers, and the discounts from wholesalers were lower, then the stock turn would have to be higher. If the shipping frequency were the same, that would suggest that the direct-supplying publisher could support a broader inventory assortment, which would show up at the margins on the slowest selling books. As a practical matter, this might mean that a wholesalers slowest-selling books might take 9 months to sell, whereas a publisher could afford to stock a title that would take a year. But that is the kind of difference we'd be talking about, and it is a climate of much greater opportunity than we have now.

Another opportunity that VMI generates is the ability to put books where there are no book buyers, something that just cant be done today. As a practical matter, that opportunity is more useful to wholesalers than it is to publishers, because it applies to non-trade situations that demand a range of titles across a specialized subject area. The Butterick initiative we discussed earlier is unique because they already have the accounts in their classes of trade, but wholesalers operating VMI could do wonders for book distribution in gift stores, stationery stores, record stores, and other places where books could be sold effectively like hotel newsstands.

Installation of VMI would also change the negotiating dialogue between publishers and their customers. The current negotiations between publishers and booksellers or wholesalers revolve around publishers offering incentives, in discount or marketing money, for their customers to take the initial quantities the publisher wants to place of new titles. There are occasionally, but very occasionally, incentives to promote backlist sales, usually across a list rather than targeted by title. It is evident very quickly in a VMI environment how simplistic and unproductive it is to contend over discount points.

We had suggested that the publishers efforts to manage VMI would be controlled by a store by giving the publisher an inventory ceiling. This is a bit of an oversimplification. Particularly in America at the moment, where superstores are growing and shelf space is plentiful, what stores are actually concerned with is not how much inventory they have, but how much money they have tied up in inventory. This is not the same thing. If publishers bill on 30 day terms and collect in 60 days, the books that sell within 60 days constitute positive cash flow to the store. Only those that sit there longer actually "tie up" money.

From a profit point of view, what stores will be measuring is the value of a fraction with the dollars of margin generated as the numerator and the number of dollars tied up as the denominator. Drive up the numerator or drive down the denominator and the result is improved.

Driving up the numerator requires selling more books, or giving away more discount. Publishers would always like to sell more books, and will always try to do so, but they'll never like giving away more discount.

So we don't think it would take publishers very long to see that driving down the denominator could be accomplished by delaying collection. In fact, publishers could deliver a superior profit to the store with LESS discount, by simply extending payment terms. The entire nature of the dialogue in a VMI relationship focuses both the publisher and the bookseller, and if they are involved, the wholesaler, on real profit. It enables seeing all of the factors that influence real profit: stock turn, discount, returns levels, and the speed at which money changes hands, in their true perspective.

A change as profound as a shift to Vendor-Managed Inventory could never be called "likely" to happen. we're clearly calling for it in this speech, but we're not predicting it. But we will predict that, without it, the slide that has taken trade publishing in Britain and America from a modestly profitable business to a modestly unprofitable one will continue, its root causes being fed by all of publishing's key players, with escalating risk for every one of them.