CIL2007: Improving Content with Customer Conversations
Posted by comartslibrarian on April 18, 2007
Rich Wiggins of Michigan State University presented.
By analyzing your search logs, you can tell what your users are seeking and how to tune your site to give them what they seek the most. Blogger’s note: this analysis should include site ‘hit’ logs as well as search logs.
Search engines can confuse low-value content with vital content. Jakon Nielsen estimates that half of website users go immediately to a website’s search engine to find what they’re looking for.
Michigan State rolled out its “Best Bets” service. Did this by studying the most popular unique searches, mapped these to appropriate URL. Compaints went down! Developers continued to watch, learn, and respond.
Google is training users to expect high relevancy results from a search box! Check out Duke’s home page to see one university that has responded to this!
The long tail (Zipf Curve) for searches: most of the executed searches occur for the same thing and are one or two words long.
- 7218 campus map
- 5859 map
- 5184 im west
- 4320 library
- 3745 study abroad
Top 221 searches accounted for 40 percent of all searches at MSU.
Which common queries retrieve zero results? Which results are most frequently clicked-through per query?
What can you improve?
* Interface Design
* Retrieval algorithm modification
* Navigational design
* Metadata development
* Content development
* Spell checking for search engine
MSU builds an A-Z index automatically based on frequent queries.
Wiggins and Louis Rosenfeld are collaborating on a new book, “Search Analytics for your Site: Conversations with your Customers” due in 2007.