The (Invisible) eBay Economy

A while back eBay began mining and reselling information from its database of transactions on its site. At the time, while most people ignored the announcement from eBay, I said that it marked a massive change in the market for price information.

It is a kind of “informating” apotheosis. That is, eBay now throws off profligate amounts of information along multiple dimensions — prices, volumes, types, products, etc. — because of the way it has been automated. Capturing that information and reselling it, given eBay’s size and scope, is a big deal in the worlds of pricing, competitive intelligence, and even microeconomics.

Here are some nuggets from a new piece in USA Today talking about how eBay is beginning its mine its own systems:

  • At the beginning of 2003, BMWs, Gucci and Prada were among the 10 most-searched terms. Now the most-searched items have shifted to Fords, anything pink, and gold (the kind you store in a wall safe).
  • Word searches for all of 2002 reflect a society still spending freely. Among the top 10 searches for the year were BMW, Louis Vuitton, Prada and Coach.
  • Similar terms dominated the top 10 into early 2003, until August, when there was a sudden shift. The Iraq war was dragging on. Companies were still cutting jobs and keeping raises flat. The blackout hit. California was in political chaos with its recall vote. And just then the luxury names dropped off eBay’s top 10, replaced by more mundane words such as Ford, Chevy and diesel.
  • In September, “salvage” made it to the top 10.


  1. I propose that the shift in August was due to the Bush tax rebate. Rather than fantasizing about purchases, people could actually make one, only with much less money.

  2. Fair comment. It’s certainly possible that people suddenly had money and things to buy in August, so they stopped fantasizing about luxury goods. But that said, where would you have drawn the line? If they had been searching for boxed noodles rather than Fords, would that have meant more? I ask not to be politically mischievous, but because I’m curious how you might extract more meaning from trends in the eBay economic data.

  3. I said I would propose it, not justify it. It just seemed as good an explanation as a blackout.
    I suppose one would need to correlate searches and bids to obtain any really meaningful information. On the other hand, I’m a self-professed cheapskate. If I even remotely reflect other eBayers, extracting information from searches and bids would probably be meaningless in the macro economic sense.

  4. Yes, that’s true enough. eBay users are hardly representative of the entire population. They are more representative of people who will go through a few hoops to get a good price, i.e., cheapskates, to you use your word.