Enterprise Search Engine Best Bets – Pros & Cons

Note: this article concerns enterprise search, typically for internal audiences and organizations. Best Bets have other purposes for public facing implementations, which I’ll likely cover in a future post.

In a previous role, I was the Enterprise Search Product Manager for a major financial institution, responsible for the vision and supporting strategy for search within the firm. I also had personal interest in improving the overall search results quality.

“Best Bets are valuable and should be considered when developing a search strategy. However, they should not be implemented to replace poor search results (that’s another, separate project). They should compliment the search results.”

After a brief analysis, I determined that there was an abuse of the Google Keymatch feature. Keymatches, or best bets are they are referred to in the search industry, are search results that are placed above all other organic search results. They are managed manually, and with that, have a high degree of overhead. However, they are quite valuable in certain situations. We had over 2000 best bets implemented, of those, many were obsolete. Additionally, at least 30-50% were not being searched on, providing no value to the organization.

After a clean-up of our best bets, I immediately changed our policy around their implementation. Best Bets would only be implemented if

• the online resource was not crawlable due to specific issues (behind security/single sign-on, etc.)
• the online resource could not rank well (listed on first page of SERP) due to issues beyond the control of the content owners (portal issues, etc.)
• overwhelming political pressure (yes, an unfortunate outcome of the corporate world).

There was an interesting response to our change:

• Content owners: “How am I going to get my content noticed?”
• Management: “Where’s the benefit in placing resources to regulate this?”
• Google (Vendor): “What’s wrong with Best Bets / Keymatches?”

My response: “We are not paying thousands of dollars for a search engine to then manually superceed what the search engine does naturally as long as the content is valuable, accessible and relevant.”

We immediately began an awareness campaign with our content owners and portal partners on search engine best practices, such as effectively using the <title> tag, headings, textual content replacing graphical text, JavaScript flyout menus, etc.

Best Bets are valuable and should be considered when developing a search strategy. However, they should not be implemented to replace poor search results (that’s another, separate project). They should compliment the search results.

Best Bet Best Practices (say that ten times fast…)

• Create a policy and procedure around the implementation of best bets.
• Communicate this policy and make sure all content owners are on board.
• Monitor search quaries and results to determine the need for best bets.
• Unless a self-service model is implemented, make sure to perform quality analysis on a regular basis.