Warning – This is going to be a bit on the techy side, but as always I will try to simplify.  That being said, the method/solution I’m going to describe today is far reaching and doesn’t just apply to search engines. 

Search engines, like Google, use a system of values / algorithms to arrive at their site ranking.  This varies from search engine to search engine, but they all are similar.  They prioritize certain criteria, such as links to a site, or links from a site to another site, or any other readily available information about that site.  Where they differ is their prioritizing of one of those criteria over the other.  Now, that system is designed to go out to an internet that is just functioning as it should be, and gather up information so that the calculations Bing or Google are making are as accurate as possible.  Unfortunately, what happens is that certain folks out there are more concerned with their ranking on the search engines than they are with providing valuable content, and they study the system, and simply modify their site to maximize the value of the site on the search engine’s criteria.  The result is that the folks that are just worrying about the best content possible are beaten out in ranking by others who are basically just cheating. 

Now the solution is fairly simple.  I will use the analogy of a teacher (the search engine) and some students (various websites needing to be graded on their quality)  If you were a teacher in a classroom of 10,000 students that you couldn’t really regulate (like the web), you would want to reduce cheating on your tests, so that you could accurately grade a students progress.   Obviously cheaters would have an easy time of it.  Answer keys would pop up and be easily passed around.  What you COULD do, is mix up the questions – to a degree that answers couldn’t be easily shared.  You may not eliminate cheating completely, but you would reduce it significantly, and that’s the point of the excercise. 

So what should a search engine do?  Or perhaps what should any online entity that runs a standardized test or study of a large community of users?   They should mix up the questions.  Google and Bing have multiple data streams coming in.  Much of it is related.  I suggest that while the result may not be exactly the same by using various methods and measures, there is enough of a definite relationship between the data sources and measurement methods as a whole, that a search engine could continuously switch what they were measuring at some regular interval, and greatly reduce the effect of cheating as a whole.  They wouldn’t have to switch up every measurement for every site, just some of the sites, enough to confuse the observer as to what the heck was going on.  Then, within the program at the core of the search engine, the data could be compared and related back to a standard scale similar to what they were using in the first place, without switching things around.  For instance, for a football team to win against another team, there are more points scored by one team than the other.  There are a host of statistics included in that game.  If it’s a complete blowout, then all the stats will be in favor of one team.  If it’s close, ther may be some stats in favor of one or the other, but as a rule, the winning team will usually have the better stats overall.   You can’t say that the team with the most sacks wins, or that the team that had the least amount of interceptions won, but when you add those and other statistics together, and compare it with a history of the many games that came before, you could more than likely tell with great accuracy, who won the game, without needing to know what the score was. 

As I say, this applies to all online measurements of peoples response and opinion.  You design these things to be as simple as possible, and they usually would be very effective, if not for the folks that figure out how you are measuring, and cheat their way to the top.