Make Internal Search Search Friendly!

By Dave McAnally, Natural Search Supervisor, Content Solutions

In trolling the internet throughout the day (as I’m wont to do), I was doing some rep/buzz research for a certain pharmaceutical client of ours and hit up a very large and well known authoritative health website. I used the internal search function to search on the core product our client sells. When I clicked on the (one) result, I was given a HUGE amount of information on our client and their drug. It was written in the typical ‘just the facts ma’am’ fashion you expect from authoritative folks in the pharmaceutical biz, but it was great content no less.

My immediate thought was “I wonder why there’s no link back to the official client page information?” It seemed like a logical fit to have some knowledge share here (if you’re talking about a drug as an authority, shouldn’t a link to the manufacturer’s site bolster the case?). This prompted me to do some investigating as to the history and findability of the page itself. I noticed that the URL was the same URL from the prior page. That’s curious… A search through the code revealed that this was not a frameset or an AJAX concoction wherein the page lived inside another. I even did a site search on Google of the website (those surveyors of dynamic-ness that they are). As it turns out, Google knew nothing about our client’s product on the site, despite this massive tome I was just looking at! I did a few other searches on well known drugs just to make sure this particular one hadn’t slipped in from the fifth dimension or something, and sure enough, nothing was leaving any sort of unique footprint in the URL.

Now, let me make clear there may be a conscious reason for this. I have learned that the pharmaceutical world is rife with legalese and regulations that could very well require the company to bury the drug information (although highly unlikely, but hey I’m no lawyer). Be that as it may, I am more concerned with what technology is preventing any of this information from being visible to search engines, depriving visitors of having a URL to bookmark/send, and ultimately living as an island only accessible from an internal search.

The site was built using Java Server Faces Technology. Nothing wrong with the functionality, but think the ease of implementation with a component based UI. What appears to be happening in this case is that the parameters to request the information out of search are passed as an API of sorts. The net result is that as the elements of the page change (such as the search field magically turning into a trove of information about my client’s product) do not cause the URL to change. There are obvious benefits from a developer’s perspective to render pages in this fashion that have nothing to do with the URL. Still, my point is as thus: If you are using JSF internal search functionality (or anything that is feeding in results), MAKE SURE YOUR RESULTS STILL RETURN UNIQUE URLS!!! More specifically, give your users somewhere to go that doesn’t involve ‘#’at the end of the URL string. This is a usability issue as well-what happens if I want to email somebody with some horrible ailment the information I found on your page?

Internal search result pages are one thing if they cannot be indexed (seeing as how spiders can’t go arbitrarily searching for drugs to learn what you have), but your content around the subject should still be findable, link-able and bookmark-able. If information lives in a database that is fed via API to the front end, consider providing an additional repository for this information. Your users will thank you by accessing the content from your page instead of your competitors (and bookmarking your page instead of the other guy’s).