The Dirty Secret of Enterprise Search: It's All About Good Content
One thing never fails to surprise me whenever I work on implementing intranet or enterprise search in a large organization: a lot of people think search engines are some kind of magic, and believe they should be able to find their content “just like Google.”
My good friend and CMSWire contributor Martin White has been saying this for years — you need a team of people to provide good enterprise search capabilities. He has detailed the make-up of such a team and the individual roles in his book “Enterprise Search.” White is also a member of “The Search Network” – a community of experts who have collaborated to create a report on search satisfaction.
A 2-Pronged Approach to Resolve a Metadata Mismatch
Recent experience suggests a major step on the path to successful search in large organizations is having the search team educate its partners on the importance of good quality content and good quality metadata.
To illustrate this fact, let me paraphrase a recent discussion with a colleague which happened while doing some relevance testing for search results. My colleague noted the page he was expecting to find for a particular query was not appearing anywhere in the search results. He explained the query he was using was based on the business process specific terminology used by contact center agents. When I took a look at the page and its extended metadata attributes, I noticed the business terminology was not:
- In the page title.
- In the keywords metadata.
- In the search description metadata.
- Mentioned at all in the text on the page.
So it’s hardly surprising that the particular query was not surfacing this page in the results. It's worth noting the individual in question is not one who thinks search is magic and is very knowledgeable about content management. However, he isn't the content owner, and was unaware of the miss-match between the contact center business process and the metadata the content owner used to tag the page. Obviously, this conundrum has a number of facets: the content owner was writing and tagging for findability based on one or more particular use cases and was unaware or didn't anticipate a different target audience who'd need to use the information for a different purpose and therefore was entering a very different query to search for the same information.
What could we do to rectify this particular issue, which I would characterize as a broader information management issue? Two sets of actions or activities would help here, one involving the search engine, the other involving the content.
Search engine-specific elements:
- Use synonyms and vocabulary list features to map the new query terms to the existing metadata.
- Use entity extraction to map content on the page to the new query terms.
- Use “Best Bets” functionality to link the particular content item to one or more particular queries.
- Make sure the new search term(s) are added to keyword and search description metadata.
- Ensure the new business term is added to the appropriate content on the page.
- Use the social tagging facilities of the intranet to add additional user generated tags (another form of metadata) to the page.
Related Article: Intranet Search Is More Than a Technology Problem
Search Engines Alone Can't Produce Good Search Results
So what does working through this little real life story tell us? In short, any search engine — however good it is, whatever features and functionality it provides — needs good quality content to work with.
High content quality, including good metadata, is key to the usability of an intranet in many ways, including search. However “writing for the web” type training for content authors does not in my experience focus much if at all, on the elements of findability that can be embedded into content. These include making sure content authors use the right business terminology and language, and that they include certain words and phrases with both a frequency and proximity to each other that search engines algorithms can use them to provide relevant results.
However, as noted above, content authors might not initially understand all the potential audiences and use cases for their content, and so part of writing good content from a findability perspective is completing a thorough stakeholder analysis and ensuring you understand all the potential audiences.
Related Article: Your Intranet Is Only as Good as Your Metadata
Helping People Find the Information They Need
Whether your search team is a couple of members of the intranet team, or a team dedicated to managing and continuously improving your intranet and/or broader enterprise search capabilities, part of their duties should include educating content owners and authors on how to create content with an eye on findability — helping them help information consumers to easily find the content they need to complete their tasks and do their jobs. Good quality content equals highly relevant search results.
About the Author
Jed Cawthorne is Director, Security & Governance Solutions at NetDocuments. He is involved in product management and working with customers to make NetDocuments phenomenally successful products even more so.