I’m interested in finding a place that sells the item a lot less often.

Another difficult search is often finding the answers to technical questions about computer systems and software. The most useful information is often buried in a blog post (“How I got around the infamous data loss bug with the Whizzbang 9000”) or in a mailing list archive. But the results there are much harder to pick out, and mailing list archives are often duplicated, so you have 10 hits (or more) for essentially the same information. In many cases, if you can’t find what you want in the first few hits, you have to dig much deeper in the search results to find something that suggests a useful answer.

This may seem like whining—after all, I’d be a lot worse off without Google making it possible to do these searches at all. I’m not actually complaining; Google is a terrific resource and I use it all the time. Rather, I think this is an interesting aspect of what gives Google its market leadership, and I’m wondering aloud what might disrupt that leadership.

I said earlier that my feeling about Google’s search results is entirely subjective. I don’t have a long-term actual data, it’s not based on any analysis, and there isn’t a randomized, controlled study with a sufficiently large number of people involved. On the other hand, that kind of feeling—that the search results just seemed better, was, I think, a large part of what drove a shift from AltaVista to Google some years ago.

From a business point of view, Google has
done an amazing job of keeping the interface
clean for its users, while providing ads that
make Google a lot of money.
Presumably, the ads are also
proving effective for the adver-
tisers as well, and one might

infer that Google is effective at connecting customers with products and services they want to purchase. In a competitive situation, a competing search service would need to master a revenue strategy as well as search technology. Competing successfully with Google is a serious challenge.

How do you provide better search results? That’s an open question, and many have tried without gaining much traction. Besides the algorithms for doing the search, you now need to have the infrastructure to index almost all of the publicly-accessible Web, and that’s a big problem all by itself. The network and capital investment needed for that is a much bigger challenge than it was when Google started.

A different kind of answer may also be with the “semantic Web,” an idea that Tim Berners-Lee, creator of the World Wide Web, has been championing for the past several years. The core idea behind the semantic Web is that websites can (and should) provide semantic information accessible to computers about the information content they provide. In theory, this gives search engines more information about whether “home depot” is talking about a store or a train station. It’s still early for the semantic Web, but so far text indexing is still the main tool we have to work with.

I don’t think we’re done with search on the Internet. As the Web continues to grow, finding things on it will be increasingly challenging, and there are likely openings for those who can really deliver consistently better search results. The payoff for that could also be big, with the possibility of a surprisingly rapi~d shift to an _{alternative search provider.}

 

DOI: 10.1145/1377016.1377020

Permission to make digital or hard
coPies of all or Part of this work for
Personal or classroom use is granted
without fee Provided that coPies are
not made or distributed for Profit or
commercial advantage and that coPies
bear this notice and the full citation
on the first Page. to coPy otherwise,
to rePublish, to Post on servers or to
redistribute to lists, requires Prior sPe-
cific Permission and/or a fee.
© acm 1091-556/08/0600 $5.00

Win treese is co-author of Designing systems For Internet Commerce. He can be reached at [e-mail].