How search engine searches can mislead and misinform

Ideas, thoughts, and resources from the permanently curious.

Living

World

Politics

How search engine searches can mislead and misinform (7/29/03)

Note: I typed this up in response to an email from someone who wanted to know about Search Engine Optimization. They'd made some bad assumptions so I was trying to expand them.

1. Google is the #1 search engine because of the way it ranks webpages. Why it's there is a bit counter intuitive though, in that Google could actually care less about what's ON the page it's "ranking". Google's trick is that it only (or mostly) pays attention to what the people LINKING to a site are saying about it. For the most part this works out fantastic. Only popular links are ranked high, because everyone is linking to them. And since popular links are probably good, the results are "good". This does have a downside though, in that if you can get together a couple hundred people with webpages you can have some serious fun. For instance, should all of those people put a link on their website with the text "purple monkey" and the link going to www.bentley.edu, anyone searching google for "purple monkey" would get www.bentley.edu as the top link. This is why the newly popular google search of "weapons of mass destruction" leads to a humorous website rather than a webpage about WMD {sadly, this no longer works -ed.}. There are just more people talking about that one site than any other site that has that keyword on it. This is called a "google bomb":

http://en.wikipedia.org/wiki/Google_bomb

This also brings up some interesting implications for searching. If a great deal of people believe in a myth, people searching for information are likely to encounter links that are going to *reinforce* that myth, rather than dispelling it. Medical information is notorious in this area, which is why it's best to use webmd.com or another service like that to get such information. Searching for one of the many email hoaxes circulating will also lead to many people discussing the hoax as fact, but a visit to snopes.com yields the truth. In the end it helps to be aware of the results google is returning and to always consider other viewpoints. While the viewpoint the person is searching for may turn up 10 results, the "correct" one may turn up 100.

2. Advertising (paid links). Thinking about this logically, if a person is searching for "best tv" their probably looking to buy a TV. So if they see an ad for a TV and in the blurb for the ad it says "independent testing confirms this is the best TV", people are more likely to click on that link. If, on the other hand, the ad says "cheap TVs", their very unlikely to click on it. This is another way Google changed our lives in that the ads that show up on search results are actually tailored to the search result. As such, their "click through" rates are much higher. Spending the extra 2 hours to come up with 5 ads rather than just one is worth it if the ads can be tailored to whoever is going to see them. Advertising has it's place it just needs to be "smart". Bombarding people with spam doesn't fall into that category, which is why people complain about it. I've yet to hear any calls for congressional legislation of google text ads.

3. re #2, the search for "best tv" will most likely never turn up the best TV, because whatever company makes the best TV has that reputation for a reason and in all likelihood, isn't doing much in the way of advertising at all. They might not even have a website! A practical example is a search for "best bed". I know from personal market research this is most likely the "Dux" bed made by the Duxiana company of Sweden, but a google search for "best bed" won't turn it up. In fact, the web is so "cluttered" these days that rather than find the best bed, I find instead links to travel destinations! So if I own a mattress company and want to advertise my bed... well, paid links of google probably shouldn't be first on my marketing budget. Instead, I'd probably try to get some journalists to write about it and link to my website in their stories, which would raise the rating in google once those stories go on-line. That'd hopefully get people talking about the beds on the numerous discussion boards on the web, which would then raise the rating even more. By then the beds might actually be the #1 search result anyway, even without the paid links.

4. Given all of the above... there's just no substitute for talking with an expert in a field. Those are the people who've been working on the search term for so long that their likely to have made it past all of the problems I've noted above. Experts can probably be divided between commercial and academic, where commercial experts would be a place like consumerreports.com, and the academics would be any professors doing research into the area being searched on.