At Google's Searchology event last month, VP of engineering Ubi Manber said that the current challenge in search is understanding people. Google doesn't understand yet, but he said the company is off to a good start.As a measure of how much Google doesn't understand, consider that less than 1% of Google users who search from Google.com click on "I'm Feeling Lucky," the search button that takes the user directly to the Web site of the top search result. Clearly, Google searchers recognize the limits of search engine smarts. Microsoft, with the launch of Bing, is also saying that search comes up short, citing ComScore data showing "that as many as 30 percent of searches are abandoned without a satisfactory result" and "that approximately two-thirds of the remaining searches required a refinement or requery on the search results page."

Yet Microsoft also acknowledges that people are generally happy with search, nothing in one video that something like 74% of people are generally satisfied with online search.

Microsoft argues that it can help searchers make better use of the information they find, but I suspect a more effective way to improve search would be user training.

Computers have failed to deliver real understanding, which is to say artificial intelligence, for decades. While further improvements in divining user intent will be developed over time, encouraging people to understand how to construct a precise query seems likely to lead to better results sooner.

People in other words should learn to understand search engines.