Why Search Engines Suck, Part Two
Last week, we examined the flaws in the indexing cache methodology of search engines. Today, we look at the problems of search methodology. There’s much there that needs fixing, but it won’t happen. This post is only a wish list because nothing will be improved until searches are restricted to the main content on Web pages.
To begin the list, what is the point of returning matches if they don’t contain all the words in a search request? As if that’s not bad enough, why return partial hits before those including all the search terms? Better search tools could fix this. Why not an option for a strict search, i.e., one in which the only hits are pages containing all the search terms?
Another option would be a priority search, wherein search terms are arranged with highest priority first, then next lower, etc. If this were combined with a strict search, then only the hits with all search terms would appear first, followed by hits with all but the last search term, and so on.
This last suggestion is similar to the iterative search discussed in the post of 12/5/11. A strict priority search is like an iterative search except all the terms are present from the start. Searching one term at a time in priority order would be an iterative search.
Among the many problems of search engines, exact matches are often the most difficult to find. Although all search engines rightly ignore characters like dashes and commas in ordinary searches, why is this also the case when requesting exact matches? Since when does exact not mean exact? Search for “Smith, Frank” and the search engine ignores the comma! And searching for “Frank Smith” will return hits like “Robert Frank, Smith Bros.” Again, the search engine ignores the comma.
Related to the problem of exact matches, search engines have no method for accurately finding proper names, one of the most common searches. Generally, searching for a proper name means an exact match on the first name and the last name, but allowing any name or initial(s) in between.
After accuracy, the most important criteria for information is timeliness. What’s the point of allowing us to limit the time frame of a search, when search engines refuse to return hits with most recent first? If you want results for the past 24 hours, why would you not want them in order of most recent?
In the final analysis, there is one preeminent problem with all searching. In a word, terminology. How can you ask a question, if you don’t the wording of the answer? It’s like trying to find a word in the dictionary when you don’t know the spelling. When you don’t know the right words, you need all the help you can get.
Why don’t search engines do more to help us? As I’ve said in this blog many times, they don’t see that as their job. What they see are the dollar signs of advertising. A Senior VP of Google has said, “The more people that use our products, the more opportunity we have to advertise to them.” And the more time you and I spend using their products—searching and not finding—the more opportunity we have to see that advertising.