Hettinga's Best of the Month Internet Auctions

Robert Hettinga
rah@shipwright.com

44 Farquhar Street, Boston, MA 02131 USA (617) 958-3971

Visit for more related articles at Journal of Internet Banking and Commerce

The invention of flat pricing (as a store-wide policy, by Woolworth's in the mid-nineteenth century) was a great marketing innovation, up there with the invention of franchising (by Singer) and auto loans.

However, compared to its alternatives, haggling and auctions, flat pricing has the disadvantage of throwing out a great deal of market information. A vendor setting a flat price for his product has no very clear way of learning how many people might have been willing to buy at double or triple the price or how many who refused to buy might have been willing to do so for a quite small price adjustment, like a 10% cut.

In short, flat pricing keeps the price "data grain" suboptimal. This is unfortunate. Price signals illuminate and represent a society to itself; in the terms of that metaphor the pixels of pricing should be as small as possible. Flat pricing, for all its convenience, is irremediably low-bandwidth.

The spread of auctions on the internet inspires one to wonder if it might not be possible to use the net to replace flat pricing in some sectors without paying with added time and inconvenience. At the moment net auctions are confined to the obvious: basically collectibles and resales of the sort that appear in Want Advertisers. Would it be possible to sell production items, new cars, hotel rooms, clothing, via auction? Would both buyers and sellers have incentives to switch over from fixed pricing?

For instance, suppose some manufacturing company were to auction off each day's production. People who wanted a car (say) would log onto the auction site and enter either a date of delivery or an amount. Once they entered one value, software would estimate the other within a range of uncertainty. (If this range was unacceptably broad you could narrow it by buying insurance.) The entire transaction need take no time at all.

The advantage of this system to consumers is that they would be able to play with two 'resources': time and money. Purchases could be pushed a lot closer to their true utility function because consumers could maneuver in two axes instead of one. (Granted they can do something a bit like this now by waiting for a specific model to appear on the used car lots, but the system I'm thinking of is more formal, general, and certain). Producers would probably make more money, since it seems likely that the extra extracted from the impatient rich buying early in the cycle would exceed the money lost from selling below cost at the end. Even if they didn't they would get marketing information to die for. The only losers I can think of would be rich people who now get away with paying less they are prepared to pay (and therefore less than the value of the item to them) and therefore are being 'subsidized' by flat prices. Perhaps there are others.

The auction model seems to fall apart with goods so cheap that no one in the population has any interest in delayed delivery. It is hard to see how auctions might be used to distribute newspapers and soda and fast food. Medical care and search and rescue services are perhaps not in all cases appropriate for auctions either.

> The difficulty with introducing auctions centers around who benefits.

> Since you aren't changing the goods sold, you aren't creating much

> value there. The new value comes from obtaining better information

> about people's utility functions. But then, who gets the value from

> that information?

I dunno. I don't think there's a right or wrong answer here, beyond the obvious point that both buyers and sellers have to want to participate. If the seller tilts the playing field too much in his direction he'll be playing by himself, is all.

It's not as if he doesn't have competition. I would strongly urge all the members of this list to set aside an hour and tour the auctions now running out in net country. According to Yahoo there are 65+ sites out there, some of them quite large indeed. Some disclose bid history, some do not. Most sell the expected collectibles, but there are a number of sites dealing in overstocks and refurbished goods of a very wide variety of types. At least one promises to start handling industrial auctions, as in sea/land containers.

The assumption of the original post is that auction pricing is socially better than flat-rate pricing because the information you get from auctions is of higher quality. (I also assume that everyone benefits from a smarter economy in the long run.) The question was whether the net could in some fashion allow AP to nibble around the edges of the sectors colonized by FRP, particularly 'many of a kind' products (as opposed to 'one of a kind') like cars or refrigerators.

The question is still open, but it's interesting that the auction sites dealing in overstocks often sell several units of the same item in parallel auctions (i.e., not as a single lot). One system is to rank the bids and distribute items down the list until the stock is exhausted. If there are N items the top N bids win. There are others.

One benefit to buyers (so far as overstocks go) is the satisfaction of getting the goods at about 25% of list price, which means that the system is to some degree parasitic on FRP. Another benefit is feeling that a price set by bids represents a truer estimate of value than one set by vendor fiat. You have more confidence that you are not being ripped off...

Of course sellers also have to feel they're making out -- they have to have some reason to think their aggregate return from AP will be higher than it would have been by setting a flat price. This is easier to imagine with small lots of a dozen units but I can sort of dimly imagine distributing an entire production run this way. Again, the marketing edge is that buyers are more likely to feel that a high price is legitimate.

Fred