WHY WEB STORES LOSE MONEY.

It's Christmas, and Web merchants are having a merry old time, packing toys and books into gift-wrapped boxes faster than Santa's elves. Down on Wall Street, Christmas stockings are extra-full this year, stuffed with bonus checks from a year of wildly successful dot-com IPOs.

Of course, Santa Claus has always been a non-profit venture, and so no one seems to mind that precious few Web retailers are operating even close to breakeven. The cash is better spent on "customer acquisition," which--as far as we can tell--is supposed to be achieved with the help of amazingly weird TV commercials about gerbils shot from a cannon and naked pizza buyers (which presumably generate more bonus checks for the creative staffs, and perhaps an award or two).

The success of Web retailing is wonderful news, of course, except we can't help noticing that big online merchants keep losing money on operations as well as on their splashy marketing campaigns. And, even after a few glasses of eggnog, we remember that the Web wasn't supposed to work like this. We recall all the visionary talk about "frictionless" commerce and the "disintermediation" of inefficient channels. No more brick-and-mortar stores, no more warehouse inventory, no sales clerks, no local advertising: The new virtual storefronts would bring marketing and transaction costs almost to the vanishing point.

Instead, we somehow ended up with a channel so inefficient that the best-managed stores routinely lose two dollars for every dollar of sales. And when the dust settles on this blowout Christmas, we suspect the red ink will get even deeper, as Web merchants sort out backlogs of returns, customer complaints, and unsold merchandise. In the end, frictionless retailing looks a lot like perpetual motion devices-- always elegant in theory, but somewhat problematic in the real world.

Why should we care about the operating losses of the big Web merchants? In fact, these losses raise troubling questions for software publishers who hope to build Web stores that actually turn a profit. If Amazon, Egghead, and Beyond can't find the rainbow's end, it's fair to wonder if there's hope for sites that haven't yet hit a million-dollar run rate.

To be sure, many publishers do make money--sometimes a lot of money--on their Web store operations. But so far there's little consensus about why some stores succeed and others lose money. It's not always clear which metrics we should use to measure success, or even what role a Web store should play in a publisher's overall channel mix. Still, we can point to a few issues that have a major impact on profitability:

* Customer churn rate: Web stores typically spend about $40 to acquire a new buyer, roughly comparable to what traditional retail and direct response firms spend on customer acquisition. However, most Web merchants haven't figured out how to turn more than about 10% of these first-time buyers into repeat customers. The Internet universe may have a nearly-infinite supply of new customers, but the really profitable sites seem to be investing in services and resources (ancillary content, customer support, user communities, personalization) that give customers a reason to visit frequently over the whole life cycle of product ownership.

* Shopping-cart abandon rate: The Web has an embarrassingly high rate (50%-75%) of customers who select merchandise and then simply abandon the transaction. Pretty clearly, huge numbers of potential customers still find Web commerce to be unfriendly and baffling, especially when they reach the store's "shopping cart" order form. Small design tweaks can have an instant impact on the abandon rate, so it's not surprising that a growing number of Web merchants see usability testing as an essential tool for improving their bottom line.

* Customer service response time: Would-be buyers also walk away from sales because they can't get timely answers to questions about delivery, pricing, and product features. According to Forrester analyst Christopher Kelly, "37% of Web purchasers have used customer service while shopping on the Net"--and almost half of these buyers stopped shopping at a Web merchant because they were "unhappy with the service they received." Good customer service is expensive, but the alternative--a high rate of abandoned sales and constant spending on first-time customer acquisition--is probably even more costly.

* Fulfillment overhead: The Internet has dramatically raised customer expectations about how quickly a physical product should be shipped. Effectively, the new standard is 24-hour turnaround, which puts a good deal of cost pressure on order processing and shipping systems. So it's increasingly common for the fulfillment process to gobble up 20%-30% of revenues, particularly for Web stores that handle many small transactions. Squeezing a few dollars out of fulfillment costs can be tough work, but even a small cost saving is likely to be repeated on every transaction--so the long-term impact on profitability can be remarkably high.

* Average transaction size: According to outsourcers who process back- end transactions, software buyers spend an average of about $60 per shopping cart when they buy from a publisher's Web store. That may be a reasonable sale for a high-volume retailer, but it's peanuts for publishers who are used to dealing directly only with distributors and corporate accounts. Will online buyers spend more? In fact, companies like Dell and Cisco have shown that big-ticket customers are comfortable with buying technology products on the Web; the challenge for software publishers is to move their own customers up to a comparable price zone (in part through greater emphasis on online licensing and purchase order transactions). Ultimately, Web stores that move beyond single-copy credit-card sales are the ones that will see the biggest jump in average transaction size and long-term profitability.