Originally posted at MYDD but I figured I should go for my first crossposting.
I've spent the last six years working as a search engine optimization (SEO) consultant. Prior to that I was running an online marketing group for an advertising agency. While running the group I realized that search engines had a very bright future and I focused on SEO. It's kept me quite busy for those six years. Many companies still are not utilizing search engines to their full potential. I also strongly believe that no one is using them properly on political campaigns. And thus I started doing some research to try to make a case for some of the democratic candidates on what they could/should be doing. While doing that I uncovered some interesting information on how search engines are being used to get information on candidates. I also believe there may be potential to use this type of information as a new form of polling.
The most important part of using search engines to reach your target audience is figuring out what words your target is using to search. In certain fields the "consumers" are using very different language than the "providers". For example, patients speak differently than doctors. So, if you want to reach patients using search engines, don't talk like a doctor (or at least don't talk like a doctor on your website). This isn't as much of an issue with political campaigns because people are mostly just typing in the candidates names. However, what I like about this method is that these people are not being prompted or polled. I'm not a polling expert but we all know some people are going to say one thing and then vote another on certain issues. And I used to work in consumer product marketing and I've seen people lie in focus groups. We don't have that issue here since searchers are at home or work and they want to find out more information about a candidate. Are these types of people probably "likely voters"? Probably more so than the average registered voter if they are looking for information at this point in the game. Perhaps they'd be considered the "early adopters" in politics. Regardless, I think they deserve watching.
Now a bit about the numbers before I get into them. These numbers are provided by a service called Wordtracker. It is a subscription service that is used to estimate how many searches there will be a day for a given search phrase. They actually look at the past 90 days to generate their estimates so that is something that must be considered. This is unlikely to show any short term bumps like normal polling.
I initially took a look at the "top 3" candidates for each side in early February and found the following information for estimated daily searches for Democrats.
Hillary Clinton 9,674
Barack Obama 6,367
John Edwards 2,187
And the 2/5 the Rasmussen Polls:
Hillary Clinton 34%
Barack Obama 18%
John Edwards 10%
The numbers for my next data point (after I read Bowers' Inflated Clinton Poll Theory post which got me thinking about checking the search frequencies again, were on April 16th) the Rasmussen numbers showed a big jump for Obama with no real movement for Hillary, and a relatively big move for Edwards.
Clinton: 32
Obama: 30
Edwards: 16
The Wordtracker data for April 16th showed a similar situation.
Clinton: 9,262
Obama: 8,917
Edwards: 3,121
I found the correlation to be pretty interesting and I will be watching this. Could the massive amounts of people using search engines every day be used as some type of polling? Seems worth considering.
Now, there are two other very interesting (to me) things that I found. Democratic candidates are being searched at a much higher level than Republicans. And Fred Thompson is rocking on the search engines (since I didn't know about him back in February I didn't track his numbers then, but apparently neither did Rasmussen so I don't feel so bad).
Here are the February search numbers for the top three Republicans.
Mccain 1,171
Romney 997
Giuliani 305
Here are the February Rasmussen numbers
Mccain 19%
Romney 9%
Giuliani 27%
Now here are today's search numbers
Mccain 1,788
Romney 2,123
Giuliani 1,061
Thompson 4,086
And the most recent Rasmussen
Mccain 19%
Romney 12%
Giuliani 33%
Thompson 13%
Now they don't correlate as well as the democratic numbers and Rudy doesn't seem to generate search interest anywhere near his polling numbers. But if I was a Fred Thompson supporter, I'd be pleased by these numbers. And looking at the Democratic numbers vs. the Republican it would see that there is a lot more interest in the Dems, which is encouraging.
So, fire away and let me know what you think.