Blogs

Would Lance Armstrong use a mountain bike to race in the Tour De France?  Of course not.  Why do we try to do the same with search in our enterprise?

When it comes to enterprise search most implementations are fixated on what metadata and what weighting should be applied to documents and other information - without ever stopping to ask “what is driving people to search”?

• Are people looking for documentation on a particular product or subject?
• Is the end user attempting to locate details about a support request that they filed?
• Or is the person looking for a particular discussion on a certain topic?

If you asked someone what they were doing on Google.com's search page, they will not generally reply "searching".  Instead users are "searching for cooking recipes", "looking for deals on stereos" or "finding news about Apple".  There is always a context to their search.  In our enterprises departments and roles all impact the specific ways in which people use search.

Without understanding the various “actors” in your enterprise and their intentions with search, we are essentially telling our enterprise to participate in the Tour De France with a mountain bike.  If we take time out to stop and think about what they are trying to achieve we can provide large time savings to similar users across the enterprise - giving them the right bike for the race they are trying to participate in.

Unfortunately a quick Google for "enterprise search scenarios" returns only technical details.  A search for "enterprise search strategy" does not fair much better.  Interestingly it seems that even though search exists to serve end users, minimal focus has been applied to their actual needs, beyond providing a generic search.

Thinking about my own experiences in search I have come across the following scenarios, with the following suggestions to increase the effectiveness of search for each scenario

1. Product Documentation for a Specific Product Version - if your organization makes heavy use of documentation in which versions are critical, make it possible to search just on a particular version using a dropdown menu that allows for specific versions to be targeted.  This saves a massive amount of time for users who previously needed to search and sort
2. Answer to a Technical Challenge - how many times have you searched to find a resolution to an issue on to end up locating a question on that topic - without any answer.  Imagine if in the search results the number of replies and an indication if a solution was reached was displayed.  This would save users at least 50% of their time in a search for the correct answer.
3. Marketing Materials for a Particular Product - this boils down to simple, metadata management.  How many times are users searching for particular content, only to find the subject, but from a different viewpoint (sales figures, opposed to product details)?  If the organization was to setup a taxonomy within their content repository it would be possible to infer metadata that would make this search much more direct in relation to the type of material being searched on for a particular subject.

There is no doubt that a vanilla enterprise search adds value, but with slightly more effort it is possible to unlock a substantial amount of value from the same search infrastructure by understanding who and what is involved in enterprise search.  We need to step back from the technical aspects of search and ask ourselves why are users searching and how can we better serve them.