I’ve been working with several clients recently in an area that gets very little love, yet is critical for your Web site: Your own site’s search function. We love to talk about Google and search engine optimization, but most of us spend almost no time optimizing our own site search. So few companies work on this capability that we are in danger of teaching visitors not to even bother using our site search. That’s bad, because their alternative is to go back to Google and find someone else’s site. One of the things that kills us over Web site search is that if we ask our customers what is wrong, they are likely to give us answers that don’t help us. We must dig deeper than that, because if you really expect to optimize your marketing results, you can’t ignore something this important.

One of the things that searchers will tell you if you ask them about why they hate your Web site search is that there are “too many results.” You probably can guess that this isn’t the real problem, if for no other reason than Google provides millions of results for every search keyword and no one ever makes this complaint. I wrote a post a few years ago to explain what searchers mean when they say too many results.

But a real problem that searchers never bring to your attention is not enough results. If you’ve spent no time optimizing your site search, try this little test. Go through your top one hundred searches and see what your search engine returns. You might be surprised at the results. In many cases, you won’t see anything that seems like the right answer. That could mean that your search engine has a problem, but just as frequently, it is your content that is either missing from the site completely or so bollixed up that no search engine could ever find it for that keyword.

I wish that fixing these problems were as easy as it is to find them, but it’s unfortunately rather complex, especially for larger sites. (That’s why these clients have hired me to help.) But in Chapter 18 of the second edition of Search Engine Marketing, Inc., Bill Hunt and I walk you through how to diagnose what is wrong and what you can do to correct it.

There are three major steps in the process:

1. Determine the value of correcting the problem. You first must assess your situation and convince yourself that there is something wrong, that it’s important to fix it, and that you are willing to spend the time and money to fix it because of the return you’ll get on that investment.
2. Optimize your most popular search keywords. When I asked you to check the results of your most popular keywords above, admit it: you didn’t even know how to find them. Don’t be too upset with yourself—most people don’t. So figure that out and then set out to create and optimize the right content for each keyword. Work your way as far down the list as you think makes sense, based on the level of effort and the return you expect. It might be the top 100 or the top 1000, but at some point you reach diminishing returns.
3. Tweak your technology and process to improve the remaining keywords. When I managed site search at IBM, we found that the top 1000 keywords accounted for only 27 percent of all search volume. Clearly we needed to do something less manual for the rest of the keywords. The approach for your “long tail” keywords has to be about technology and process. You must focus on tweaking your search engine, your content management system, your e-Commerce system and other technology so that it does a better job. And you must address your content creation and update processes so that the content is search optimized from day one.

Again, this isn’t simple, but it is critically important for most businesses. Some companies have increased their conversion rates by 25% just by addressing the simplest measures in Web site search optimization. Or you can keep doing what you are doing, which is subtly requesting that customers go away.

Originally published at Biznology