Design & Development

Personalized Search: Coming For Search Engines?

Personalized search could be the next big thing in search engine marketing. In fact, we are beginning to see the shift from homogeneous search results to personalized results already, and all major search engines have tools in place that are collecting such data. A good example is Google’s Web History. If you run a search with it turned on vs. turned off, potentially you’ll see a slight difference in search results for the same keywords.

So what has led to the trend towards personalized search? The whole idea of a good search engine is based on the concept of “relevancy.” If you can quickly find what you’re looking for, then the “algorithm” is doing what it’s supposed to. Personalized search is really taking “relevancy” one-step further. Not only will Google, Yahoo!, MSN and Ask deliver results based on their proprietary ranking architecture, but they will also look at your Internet browsing patterns, and your search history, and then deliver the most relevant search results based on your past interests. Here’s a quick run-down of activities known to “personalize” your search results:

• Search history

• Community bookmarks (such as Google bookmarks in the Google toolbar or Yahoo! bookmarks)

• Personalized home page content (when you customize a page on Google.com/ig or MSN’s Live.com, both know exactly what kind of content you’re adding to those pages simply because you’re running it through their services)

• Engagement in other tools offered by search engines

When a search engine determines all of this (if it is able to collect enough data on your online-behavior and pinpoint certain patterns) it will act on those patterns and deliver customized search engine results that you have expressed interest in previously. Additionally, if your user-behavior profile fits into other “known” profiles, search engines might generate search engine results based on the activity of other users within that profile. Examples are bookmarks in the Google toolbar. If many people bookmark the same website, Google might assume the site is “valuable” to all those users, and therefore assume it should be ranking higher. Similarly, if you’ve expressed an interest in China for the past year, next time you run a search for “news,” Google might end up showing you additional results relating to China.

It’s important to understand that none of those changes will happen overnight; rather, search engine results will be “personalized” gradually with a built-in learning mechanism. What is the value for consumers, then?

While many have voiced their concerns about personalized search there are quite a few things that the Internet users will gain as a result of personalized search results, including the following:

Improved Search Results Relevancy

If other users’ activities are taken into account, you’re likely to end up with more useful information due to the fact they have indicated which information is helpful through their browsing patterns. This should produce search results that will contain less spam (pure algorithmic) search results.

Improved Content Selection

I want my Windows Vista Google News Ticker, for example, to know what websites I often visit. However, I don’t have time to sit there and preselect topics that I’m interested in. With personalized search, the data collection is happening in the background, and the end result is news headlines customized just for me.

Less Intrusive Interface

If you’re a fan of online videos but hate banner ads, search engines that use personalized search could ensure only the most appealing ad-types will be displayed. Consequently, you might end up with customized search results that display ads in video format with no banners. You can be sure what will display is the advertising format you are most likely to respond to.

At this point Google is the only search engine to actively implement personalized search methods. Presumably, Yahoo!, MSN and Ask all have programs in place that collect user-behavior information, however no active implementation of those records has been reported to date.

Greg Laptevsky
Bio   •   RSS Feed


x