Forum Moderators: open
1. Natural SERP clickthrough rates
Adwords is all about clickthrough, but what the regular search results? We all have this intuitive feel that a higher rank gets more clicks (undoubtedly true) but -- on average, how much more? Has anyone seen any research on this? (There's a really old thread in the site search, but nobody ever answered...)
2. Overall SERP accuracy
For what percent of searches do users not even click on a single link? We've all done it -- search for "animaniacs dvd" and get a bunch of spam, jump back and search for "wakko yakko dvd release date" and see if we get something better. (Aren't we all impatiently awaiting the release of Animaniacs on DVD?)
If anybody has any ideas, articles, or friends at Google who could answer either of these questions, we'd all be able to justify what we do much more clearly.
C'mon -- shouldn't we? If we were going to buy advertising in a magazine, and the only info the publisher could tell us was how many people read magazines in general, that wouldn't be very persuasive.
</rambling>
Wouldn't that be insanely valuable? No more fruitless wondering which engine was the "best" -- we could say, well, Google has 35% of people clicking on the number 1 result. AltaVista has 17%. Yahoo has 2%. Whatever.
That would be good for users, so it should be good for SE's. At least, that has been Google's philosophy, and it's served them well.
I guess we'd need some new vocabulary, for some behaviors that would be good to track across the engines.
Direct Clicks -- somebody types a search and clicks a result. Does not hit back, but moves on through the site. (Counts as a successful result.)
Reclick -- somebody types a search, clicks a result, hits back, and clicks on another result. (A string of reclicks could end in a direct click.)
Search Revision -- somebody types a search, clicks nothing, types another search.
I guess there are a million permutations, but with the Google Toolbar, for instance, Google could be (probably is?) tracking all of this.
A research company could make a toolbar and get a random sampling of people to install it, then monitor their search behaviour across a known set of search engines to determine the ultimate "accuracy" of the results.
I wonder if any search engine company would sponsor that research. Hmm...
If Google would say the number 1 position gets 35 % of the clicks then this would be interesting but it would always be an average of many factors:
- For one keyword the number 1 site with a very good wording might get 80 % of all clicks
- For one keyword the number 1 site with a very poor wording might get 0 %
- For one keyword the number 1 site might get 25 %
So the average is 35 % but what information do you get out of this average?
If you don't know from what industry / keyword the numbers are coming from it is difficult to say what information you could get out of these numbers.
I know of cases where a number 2 or 3 position got more visitors than the same text/url had as number 1. Human behaviour will always be difficult to predict...
Also a number 1 position might bring you more visitors but a number 5 position might get you more clients as some people look at the first few results anyway and then after having seen several sites they might stop and start shopping (so if you are happy enough to be the one when they stop searching and start buying you are the luky one).
So this seems to be correct.
