The following text is copyright 1997 by Network World, permission is hearby given for reproduction, as long as attribution is given and this notice is included.

The problem with findability

The November 10th issue of Time magazine has an article about increasing findability on the Internet ( Keeping Tabs Online, by Michael Krantz) but I think it misses an important impact of making things easier to find.

The article sees great promise in some new work of the World Wide Web Consortium (W3C - http://www.w3c.org) on an extensible markup language (XML). XML lets web users define their own document types.

If common document types were defined by network-based communities of interest, auto dealers for example, then it might be possible to have a search engine find all the auto dealers that had purple Dodge Grand Caravans for sale. This would be done by the auto dealers agreeing on what the basic format would be for a web page advertising autos, agreeing on what tags there would be on the page and what the format would be for the data in each tag (car type and color for example). The search engine could then list all the dealers, how many minivans each one had in stock that matched the description and what prices the dealers had put on each one. The customer could then very easily find the lowest cost dealer and place an order. Seems like the ideal way to shop for a car -- out of range of auto dealer salespeople. But why would the dealers agree to do this? More particularly, why would just about any merchant agree?

Actually, car dealers might. A buyer might tend to want a car dealer in the general neighborhood and one that has a good reputation in the service department. You don't want to have to get your car from a dealer 12 hours driving time away so price and availability will not be the only selection criteria.

But if you want to order Levis jeans that will be shipped to, you might as well go for the lowest cost merchant. If your browser sorts the returns by cost, a difference of even a penny out of $80 will change the order and change the likelihood that you will select that merchant. And if the browser can place orders for individual items from different merchants it will allow the buyer to avoid the cluster effect where a low price on one item is used to lure buyers for other products that may not be discounted quite so much. A loss-leader such as cheap milk in a supermarket can be used to sell expensive peanut butter.

I will say that I would not be shocked if merchants rush to not embrace the potential of XML. Getting into a price war does not seem to be a good way to ensure a profitable future, and decoupling the cluster effect ensures that little if any profit would be made on each item. I expect that we will continue to see an explosive growth in Internet-based commerce but I also expect that there will be little merchant enthusiasm for mechanisms designed to minimize profit.

disclsimer: The concept of minimizing profit is not taught at Harvard so the above understanding must be my own.