Google recently rolled out Google Instant which displays and updates search results in real time as the query is being typed. It is currently the default search strategy on www.google.com. The technology is proprietary and nobody outside Google knows how it really works. But the following can be ascertained with high confidence.

There are two components to this technology:

  1. Query completion. Given a partial snapshot of your query, Google builds a set of likely completions. This is based on generic query logs as well as personalized query history.
  2. Real Time Updates. Given potential query completions, Google delivers a set of results (web pages) to your browser and displays them immediately. The list is updated as you type your query letter by letter.

The first component is not very new. Google as well as other search engines (and also academic researchers) have developed query completion algorithms in the past. Given the massive set of queries available to Google it is likely that such completions can be computed with high accuracy. That means that the completions will be accurate for most users most of the time.

The second component is new. I don’t know exactly how they accomplish such fast updates but I suspect it involves smart Javascript and AJAX coding to obtain efficient communication between the browser and Google’s servers.

This is the background. Now to the story.

I was contacted yesterday by a reporter from the AJC (Jim Galloway) who wrote a story related to Google Instant. The story contains many interesting details, but the one I wanted to highlight in this post is as follows. A person claims to have succeeded in manipulating which completions (and therefore search results in Google Instant) are deemed likely by Google to the query “Nathan Deal”.

First the motivation: the person wanted to emphasize corruption allegation regarding Nathan Deal who is a politician running for Governer in Georgia. Apparently the person does not like Nathan Deal and wants to influence the election against him (see AJC story for more details). The idea is that when people search Google for “Nathan Deal” they should first see web pages indicating or discussing corruption allegations. Old style search engine optimization techniques would prove of little use here.

But Google Instant opens up a new avenue for search engine optimization: change the list of Google’s search results in response to “Nathan Deal” by making “Nathan Deal ethics” a likely query completion to “Nathan Deal”. The AJC story claims that this was done using the following two strategies.

  1. Creating multiple new websites in which the phrase “Nathan Deal” co-occurs with the term “ethics”.
  2. Having multiple people spend their spare time typing the query “Nathan Deal ethics” in Google.

It is quite likely that the above strategies would modify the set of likely query completions and would promote the completion “Nathan Deal” to “Nathan Deal ethics”. Consequentially Google Instant would promote webpages relevant to “Nathan Deal ethics” as users search Google for “Nathan Deal”. Indeed, as you type “Nathan Deal” into Google you get the most likely completion “Nathan Deal ethics”. Note however that the search results still shows Nathan Deal’s official website as first. Looks like Google did a good job on that one. But by now the story may have gained some publicity which would (1) cause more people to type “Nathan Deal ethics” to test whether the claims are true or not which would boost the “Nathan Deal ethics” completion even higher, and (2) potentially prompt Google to override their automatic algorithm with manually constructed alternatives for this query.

In summary, traditional search engine optimization have been limited to creating and modifying websites in order to boost up their ranking. Google Instant opens up a new optimization technique: promoting certain query completions which will bias the search results for shorter queries. It also raises an interesting legal question: could such activity be considered libel in some cases? Search engine optimization is a huge business industry with an estimated 2010 market share of $16 Billion). How powerful is this new optimization technique and whether Google will come up with counter measures is yet to be seen.