While Google is currently the dominant player in the Internet search engine business, that doesn’t mean that there is not a great deal of activity among companies vying for a piece of the action. As was recently reported, even 1% of the global search market represents quite a bit of money.

Not all search engine companies use the same strategies to capture market share, however.

Google’s main strategy has been to use a sophisticated algorithm to generate accurate search results. This approach has been so successful that Google commands a clear lead on the basis of profits, earning more in one quarter than Yahoo (the number 2 in search engines) does in a year.

In addition, Google has made use of the patterns of particular users to further target searches. Products like personalised search build upon a user’s prior search requests and selection of results to tailor subsequent searches. While this practice makes results more accurate, it has also generated some criticism regarding privacy implications.

Not long ago, Peter Fleischer, Google’s privacy counsel, published an interesting opinion piece in the Financial Times on the privacy issues involved in personalised search services, and Google’s own approach (to require users to explicitly opt in). This piece makes the general point that personalised search services are valuable because they help to make increasing amounts of data more manageable. By taking into account your search history when delivering results to your current searches, you are more likely to receive relevant results. However, the cost to users is measured in terms of their privacy–Google’s software needs to know the nature of your prior searches to make these judgments. A Financial Times writer has commented that personalised search is acceptable from a privacy perspective–but only so long as Google obtains informed consent from individual users.

For personalised search to work, it is necessary to retain users’ data for a period of time. This practice, according to Google, enables the provision of improved search results, as well as to defend Google systems and users from attacks, click fraud, and spam, as well as to comply with data retention obligations and the requirements of law enforcement officials investigating serious crimes. After this period expires, Google removes the association of the server logs with particular users (i.e., Google “anonymises” the server logs). In response to privacy concerns raised by the European Union’s national data protection authorities, Google has agreed to change its own policy to anonymise search logs after 18 months (rather than after 18-24 months as was previous practice).

Personalised search is not the only way to maximise useful search results. While some newer search engines are focusing on creating better algorithms, other search companies have eschewed this approach in favour of “social search” approaches, in which search algoritm results are combined with human intervention. These methods include manually edited search results based on submissions or user votes (Squidoo, Sproose, NosyJoe), editor evaluations (Mahalo), a combination of user and editor evaluations (Bessed), or the option of user interaction with a human (ChaCha). Many of these characterise themselves as “human powered” search engines. These approaches seem to have the advantage, from the privacy perspective, of not relying on a user’s search history to generate useful results.

It will be interesting to see how user-powered search fares. Both venture capitalists (Sequoia Capital is among those financing Mahalo) and Google itself is reportedly interested in these methodologies. It’s likely that the optimal solution is not just computer- or user-based, but a combination of the two–how that is to be achieved is still being worked out.