Search Lessons

I have been running Enterprise Search discovery workshops on behalf of Microsoft throughout this year. Now, this should not be a total surprise, but the technology focus of the workshop is Microsoft’s products and services. Specifically: SharePoint Server 2007 (I’ve gone off calling it MOSS at the moment), Windows Desktop Search and Live Search (that’s the web-based service).

But this post is not about a specific technology.

When discussing requirements and objectives for deploying an enterprise search solution, one request crops up far more than any other:

¨We want our intranet search to be like using Google¨

Yup, the ultimate wish is for searching the intranet to be as quick and easy as searching the Internet. (I ought to do my Blue Peter ‘sticky-backed plastic’ disclaimer here and add ‘…using any freely-available Internet search engine.’)

People are used to Internet searches and this is the base level of expectation for querying internal company web sites. It sounds reasonable enough. Google indexes billions of pages covering the broadest possible range of subjects. Indexing a measly few million* pages on your corporate intranet and returning decent results for search queries ought to easy, shouldn’t it?

Well… no. There’s a huuuuge difference between the Internet and your intranet. It goes like this:

The Internet = pages wanting to be found. The same cannot be said for most company intranets.

On the Internet, if your site doesn’t appear in search results, it doesn’t exist. Simple as that. Why do bloggers tag their posts? (See end of this one for a simple example.) To make them more discoverable – easier to find increases likelihood of being read and linked to increasing relevance rank on search engines making them even more discoverable. I doubt too many bloggers write just for finger and brain exercise. On a more commercial level, there’s a whole industry built around making sites and pages easy to find on the Internet, including making the wrong pages appear top of the list. It’s called search engine optimisation (SEO). Added to that, Google (and likely all other Internet search engines) invest huge resources to ensure SEO doesn’t create a negative impact on relevance (tweaking algorithms to push those wrong pages back down the list) Last time I checked, not too many organisations have SEO teams working on their internal information stores, let alone dedicated resource to continually tweak and improve relevance ranking.

That’s why achieving a Google-like search experience on your intranet is no mean feat. Step 1: Start thinking about how you optimise your information sources to return relevant results for search queries. How do you incentivize internal authors to classify information the way bloggers and photo-sharers happily tag their work when publishing on the Internet? The motivations are very different. Step 2: Consider how much effort you are prepared to assign to ongoing improvements, tweaking the search engine to improve relevance as the index evolves over time. This applies regardless of the technology you choose to use for enterprise search.

Technorati tags: search

* Insert any quantity, from a few thousand to several million. The results are the same.