Enterprise search - what's your haystack?

Imagine you'd just decided to redecorate your house, and had bought in a crack team of guerilla interior designers (perhaps some form of Laurence Llwellyn-Bowen and Sarah Beeny dream ticket for just the right blend of style and re-sale potential). You're having the initial planning chat, and the only thing you can tell them about what you want is that you want to be able to 'find everything' in your house when it's done. What's the more likely outcome at the end of the show - a complete success or an unmitigated disaster?

Strangely, enterprise search projects have started on slimmer requirements sets. The trick is being able to identify the specific areas where business value would really be unlocked through better search, and what work is required to get you there, before you start. After all, searching for a needle in a haystack is a lot easier if you can work out which of the 10 haystacks you're supposed to be looking in, isn't it?

Enterprise search is one of those technologies that sounds fantastic when the salesperson describes it. Why wouldn't you want a tool that you can use to get Google-style answers out of all of your organisation's data? Surely what's more surprising is that everyone hasn't already got this up and running, and is now existing in a search-driven nirvana where all relevant information is only one phrase and a click away?

Obviously, the reality is some distance from this, and successful search projects are those that haven't taken the "put it in, index everything, see what comes out" maxim (for which the answer is generally "an unlocked file containing everyone's salary data that someone put on a shared drive by mistake"). To start off in the right direction, consider making the business value of enterprise search more clear by defining and prioritising some search scenarios that cover:

• Who is looking for the information
• What business purpose they want the information for
• Where that information currently comes from (which systems, which external sources, which people)
• What the quality/recency of that information currently is and whether it is suitable enough to deliver business value
• How that information would need to be presented to make it useful

Doing this analysis up front will help you focus on where the real value of enterprise search might be for your organisation, and often shows that indexing just one system and presenting the data in the right way will address the majority of scenarios that your stakeholders can think of. In some cases, I've also seen this help organisations understand that the things that are really affecting productivity have nothing to do with getting an advanced search tool.