Hissingsid - 8:44 am on Oct 4, 2011 (gmt 0)
@Tedster That's the question we've all been asking since Febriuary. It's essentially the same question as "what does Panda measure", and we've got no solid answers.
Just thinking aloud.
Perhaps we should ask the question: If you were designing Panda what would you have it measure?
But first the question: Why?
For as long as I've been watching Google updates there has been a wall of silence and/or misinformation from Google on each change. They clearly were trying to stay one step ahead of us so that their changes were not in vain. "They" must have got increasingly frustrated as we, by sharing information, were ever more able to react and circumvent the changes that were essentially designed to try and stop webmasters from being able to change their web pages, back links etc in order to improve their rankings.
I think that "Why?" is to have a system that cannot be manipulated by webmasters and SEOs.
Bearing in mind that Panda is only part of the algorithm and all of the other things that we have always done are still important, if I were designing Panda here are the things I would measure.
1. Bounce rate vs other sites that rank for a search term and broad topic area.
2. % of new visits vs other sites that rank for a search term and broad topic area.
3. Number of pages visited vs other sites that rank for a search term and broad topic area.
4. Average time spent on the sites vs other sites that rank for a search term and broad topic area.
5. The time spent and number of pages visited by a user before they disappear into a transactional area vs other sites that rank for a search term and broad topic area.
6. Where users go to when following a link from your site and the 1 to 5 measure of that site.
The learning and iterations that Panda is going through could be changing the mix of these six things and the balance between specific search terms and broad topic. Or perhaps they started measuring 1. to 5. and then added 6.
As I say just thinking aloud.