A book in progress by Louis Rosenfeld & Richard Wiggins. Publisher: Rosenfeld Media. Anticipated publication date: 2008
Any organization that has a searchable web site or intranet is sitting on top of hugely valuable and usually under-exploited data: logs that capture what users are searching for, how often each query was searched, and how many results each query retrieved. Search queries are gold: they are real data that show us exactly what users are searching for in their own words. This book shows you how to use search analytics to carry on a conversation with your customers: listen to and understand their needs, and improve your content, navigation and search performance to meet those needs.
From yesterday's New York Times article on nude vacationing:
SpaFinder.com, an online spa search engine, recently created a separate category for “nudist spa vacations” after noticing an increase in searches for the term. Since November, searches on SpaFinder.com for such trips have averaged about 720 a month—beating out “pet-friendly spas” (284) and “waxing services” (298).
Thanks to Alexandra Fox for the tip!
Alex, a college student, writes:
...I'm building a web based library catalog for a special collection of books as part of a group software engineering project. My group has been really interested in UX methods related to conceptual models and task based evaluation (so on the qualitative end) but we are also interested in using some more quantitative methods. Since search is such a key feature of our site, and can provide such telling insight in general, I thought I would reach out to you.
Do you have any suggestions for what types of information to log beyond the query (and IP, host, browser info) or any other advice / cautions?
My quick answer for Alex:
Definitely try to log which pages queries originated from (and what those queries were). This will help you understand where navigation might be failing and what people search when it does fail. (Thereby enabling you to improve contextual navigation from those pages.) You should also log the other end of the process: what search results people click through. This will help determine what's popular; you can also compare which documents are clicked through with which ones you want to have clicked through (and thereby test your most frequent queries' performance). Also log the time and date so you can track sessions.
What might you add?
