07:00:43 am on July 22, 2008 |
Have we found what we’re looking for? It’s similar to the well-known question, “Are we there yet?”. You can’t answer the latter until you know where you’re going and you can’t answer the former until you know what it was you were searching for in the first place. Of course I’m talking about knowledge management, specifically as it relates to customer support, in this case and not the meaning of life or a lost TV remote (by the way, if anyone has a tip on either of those please let me know).
What I’ve seen in my 13+ years of experience working with knowledge management tools is that there has been a lot of attention given to search – the “search engine”, indexing, processing of queries and rendering search results – but very little if any thought to what is being searched for. At this point you may very well be thinking or even poised over the keyboard to write the comment, “It’s obvious what we’re searching for – the answer to a question, the solution to a problem.” Really? That is, I agree, usually the goal of the search but it is not, can not be, literally what we’re searching for, the query that’s entered into the search field. After all, if we knew the answer or solution well enough to search for it effectively in the first place, we wouldn’t be going to the support site or calling the support center in the first place except for the special case of just needing the details to an already identified solution.
Now I’m not attacking search or trying to downplay its importance in knowledge management and customer support. But the best search engine imaginable will be of little use unless it is guided and targeted by consistent structure in the content. This structure must go beyond “tagging” content with topics. It’s not enough to say an article is “about Product X” or “about Installation”. The particular questions that an article answers or the particular problems that a solution addresses must be explicitly listed. Those lists must be given priority when searching because a match in such a list to a customer’s question or problem description will have the highest value. Then the features and capabilities of a great search engine – for example, the ability to give a correct match to a query even when an alternate way of stating a concept is used – will have the most positive effect on the outcome of the search.
Think about how search is usually tested and evaluated. The tester typically starts with knowledge of the content and has in mind a particular item that they want to try to retrieve. Then they enter a variety of searches that more or less specificly match that item’s title or some text in its description. If the item is ranked with a high score when the search is more specific, lower as the queries grow less specific and the item disappears from the search results list when there are no matching terms in the query then search is considered to be working properly. Instead, think about testing search strictly from the perspective of the customer – What questions are they likely to ask? What problems are they likely to describe? Search should be considered successful only if items directly addressing the problem or specifically answering the question are ranked highest or are the only results returned. Items that merely mention the query terms, even if they do so exactly – but do not specifically address the customer’s issue should be ranked lower if an item that does specifically address the issue is available.
It’s a higher standard of success and, yes, it takes more effort to markup articles for this approach but it is the only way to get better customer support results from a good search engine.Advertisements