CMSWire has a thought provoking article about the “Changing Nature of Search” The author Gerry McGovern notes that search changes over time and changes based on our behavior with applications and sites. Once we find a site to apply to a specific task like cheap airfare, then we stop hitting Google or Bing for example. But most intriguing was the relationship in how you modify your site based on insight gained from search analytics.
The BBC intranet used to analyze its top searches and then place the Top 10 on the homepage and other relevant pages. Once people could now easily find this Top 10, they stopped searching for them. So, over time, a new top 10 search queries emerged. And this caused a vicious circle because once they replaced the “old” Top 10 with the “new” Top 10, the old top 10 started getting searched for again.
Most site search is like a poisoned well: nobody wants to use it if they can at all avoid it because it’s just so bad. Thus, much site search statistics (particularly on intranets) are quite misleading. They give you a very skewed picture of what people need to search for, because people are only using them as an act of desperation.
What I see here is that it takes thought and effort to improve search. Much of that thought and effort revolves around actually improving the user experience of the site itself and not just a search tool. I can’t agree more with that concept. I also know that our Experience Design (XD) folks at Perficient agree with it wholeheartedly. They focus on defining that user experience, and it does make a difference.
That said, I also believe that search can be improved with type-ahead search, faceted search, etc. Our reality is that so much information exists that we cannot address everything through a better information architecture or better labeling. It’s a combination of it all that will bring success to your site, whether it be an intranet, partner portal, or customer facing site.