In the last few years, I've worked with a number of companies whose initial vision for their Enterprise Search is "like Google, but for our own organisation's data, including all that we've got on our laptops". Sounds wonderful, and certainly a great place to start as a vision (indeed, there's not really any further you could actually reach) - but when you start to think about some of the realities of achieving it within your company it becomes clear that a more practical initial vision might be needed, unless you are starting your search project by clicking the "I'm feeling lucky" button.
Specifically, here are a few things that you might want to consider when you set out your initial direction for Enterprise Search (and work out how much it might cost you):
- Some of your data is probably badly organised : You've probably got any number of holding areas for where the content that would form your Enterprise Search lives. Chances are that some of those areas aren't really managed very well, and will contain old content, items that don't have security properly applied, and things that would never really be of use to anyone if they came back as search results
- A lot of work goes into making Google as useful as it is : Google have a lot of people working on their search engine, but they only (largely) index web content, their business depends on the quality of the results provided, and people who believe they have the best results lobby their content through Adwords and other mechanisms. Any Enterprise Search project, to be really successful, needs to recreate similar mechanisms internally to make sure the results stay relevant to the users - it's never really just a case of putting the search box in and letting it run.
- You may be merging a couple of quite distinct use cases : by talking about indexing corporate stores and your own laptop, you may be starting to confuse the "I'm looking for something I know I have but can't find" use case with the "I'm looking for something in my corporate body of knowledge that I don't know whether we've got or not" and the "I'm looking for the definitive version of X" use cases. These things are different but related - a key point around Enterprise Search is how much of the first case you actually want happening compared to the other two (i.e. should the person in the first case actually be sharing that content with everyone else?)
- Expectations management : as with a number of projects in the ECM space (why do we seem to have a large proportion of projects where the stakeholders can get the most carried away?) good expectations management, very early on in the project, is absolutely essential. Getting this right takes time and effort and won't happen overnight. This shouldn't be difficult - very few projects offer quick returns for little input.
- Clarity on your use cases to understand the challenges : What real business problems will Enterprise Search solve, which content does that involve indexing, and what state is that content in? Being really clear on what problems search is going to solve for you will help you understand the size of the challenge in content terms, and prioritise the stores you should index for the most 'bang'. One of the ways I have used in the past for this is the concept of the search scenario, which I will cover in a future post.
"I am a great believer in luck, and I find the harder I work the more I have of it"Please let me know your experiences - how have your Enterprise Search projects gone? Do any of the challenges above resonate, or have you had the project where you just installed the software, ran it, and your business was revolutionised overnight (and, if so, can I come and work for you?)
No comments:
Post a Comment