Forum Moderators: Robert Charlton & goodroi
Here's my view. You have heard many SEOs speculate about bounce rate - and speculate is the big word there. If bounce rate were an important part of the algorithm, then the scenario you describe could have a negative affect.
But the scenario you describe is one of MANY situations that introduce a lot of noise into bounce rate as a signal of quality. The metric is so noisy that it would be unlikely to ever be important. I also know that different sites - high quality sites - have very different bounce rates, and in many cases it is natural that it would be this way. That is, there is no benchmark that can be used to compare bounce rate from one site to another.
Even the more isolated signal of bouncing back to Google's search results to click on a different choice, while a bit more focused, will also be full of noise, and in my opinion, not a useful part of the ranking algo for the website urls.
If I were a search engineer, I might watch that second bounce signal on a high volume keyword to judge whether users liked an algorithm tweak - but not to rank the sites. In that kind of metric I would be comparing a SERP's present bounce rate to its historical value - in just the same way that any website can work with its bounce rate to better meet visitor expectations. But even then, the proof would be in the actual stats - and they also might be very noisy.
I run a portfolio of about 15 sites in a specific sector. One of those offers about 30 free games and very little text/content. There are a handful of natural inbound links but not very important ones. But the traffic way exceeds the other sites - about 2,000 organic uniques a day, 10x as much as the others in the portfolio and the main difference is the visit length - in fact the other sites are much more informative, text-rich and better structured.
It has about 25% bounce rate, 33% staying more than 20 minutes and 20% more than an hour. I can't see any other reason for it's high SERPs on some (fairly competitive) longtail terms other than user behaviour.
Six months ago I launched another with some niche games - about 20 in all. Same sort of thing only this time I started with zero text on the game pages....literally none - just the game and links to the others (had some homepage text). Had around 100 organic referrals a day inside 3 months, added a little text to a handful of game pages and now averaging around 200 uniques a day, no inbound links of note.
Last month (Nov) I launched another one - same principle, 14 different games, quick descriptions of each and two inbound links, and it's already up to about 100 organic uniques a day inside 30 days. Once again bounce rates are low and retention good.
Other sites I launched, similar with more text and minus the free games, grow at a much slower rate. That's just my experience but it was enough to convince me user behaviour, including bounce rates, are being used by Google. Although I'm always open to other suggestions for these results.
Perhaps, while a bad bounce rate may have too many causal links to consistently isolate the cause (insofar as Google would need to make a decision as to the value of the content), a good (very low) bounce rate will most likely be associated with a positive cause. Essentially, I wouldn't be surprised if Google ignored bad bounce rates as irrelevant/poor signal information but positively favoured low bounce rates.
