In the olden days (way back before the turn of the century), you could learn a lot by reviewing what search terms people used to find your Web site. When surfers go to Google or Yahoo! and type in a term, their query is caught up and delivered to your door in the slipstream of referrer data.

Nowadays, people don't find your site unless you have carefully engineered the terms people are likely to use to find you. If you sell tennis shoes and you haven't optimized, paid for inclusion or are high keyword bidder on “sneakers,” then you'll never know what you're missing.

What people search for on your site however, is another story. Granted, some portion of your visitors click to navigate and never, ever search. Some go straight to the site map without so much as a look at your links.

But a healthy portion of them enter a word or two in that search box in the upper right corner of your home page. If they don't like what they see, they'll modify their search, giving you a better idea about what they're after. Gold mine.

We've now entered into the realm of search analytics, which is made up of three components:

• What people search for

• What they find

• What they don't find

As in all measurement, a static number is not important; it's the changes that count. Are people looking for something in particular more often or less often these days? Is “deep-vein-thrombosis” inching out “leg spasms” on your Health-In-The-Air.com site? Is Creole Cream Cheese Ice Cream being creamed by Old Fashioned Butter Pecan? Are people searching your site for a phrase coined by one of your competitors?

Once visitors search, what do they click on? If 95% who search for “chicken stock” go for poultry investments rather than culinary delights, you might want your internal search engine to re-rank the results based on real, behavioral results rather than smoke-filled-room conjectures. (Quick show of hands: anybody remember smoked-filled rooms?)

Mondosoft's BehaviorTracking (www.mondosoft.com) automatically generates synonym search alternatives based on previous users' search terms. If more people who search on “application” are interested in employment possibilities than in software or spreading paint, the system allows the site manager (you) to review automatically generated synonyms. You set up the semantic links between “return” and “refund” or between “hard drive” and “disk drive.”

If your visitors don't find what they want, it's time for some serious soul-searching. Is your information hiding in a non-indexed database? That would explain why people can't find your retail store in Pacoima. Maybe the information is on your site but you call it something else.

I once called a warehouse electronics catalog company and asked about laptops. The young man on the other end (obviously fresh from a management meeting on nomenclature) said, “I'm sorry, we don't have laptops, we only have notebook computers.” I expressed my disappointment, hung up, and placed my $5,000 order with the next one on my list.

Perhaps the data is there, but users simply can't find it through your search engine. It may be time to standardize the use of meta-tags across your site to ensure visibility. Many search engines can lend a hand by auto-generating unique titles and descriptions based on page content.

Now it's time to link things together to prove that you are a Web analytics mastermind:

• Which promotions attracted people who searched for your best feature?
  
  
• Which promotions brought in the most people who searched for things you don't sell?
  
  
• Which search terms are indicative of buying behavior?

Over time, you'll get to know the linguistic tendencies of browsers versus buyers, and can use that knowledge to up-sell, cross-sell and service your customers better. Isn't that what we're all searching for?