What Google doesn’t tell us about Bounce Rate and…

Some days ago, I was reading the transcription of You & A with Matt Cutts on SearchEngineLand and I found a very interesting question by Danny Sullivan that Matt Cutts brilliantly evaded. This is the passage I’m referring to:

DS: What about tweets earlier today about using bounce rate? You don’t look at how quickly someone bounces from a search result and back to Google?
MC: Webspam doesn’t use Google Analytics. I asked again before this conference and was told, No, Google does not use analytics in its rankings.

I don’t know whether you’ve noticed it or not, but Cutts has given to “bounce rate” the typical meaning of a web analytics metric and has asserted Google doesn’t use Google Analytics data for ranking purposes. I think this is true, also because not every website uses Google Analytics. Moreover Cutts also added:

Bounce rate doesn’t measure quick answers you get. You get the answer and leave, so it isn’t a good metric for Google to use.

I think we can believe him on this too, but I also think Cutts hasn’t told us everything.
Before you misunderstand what I’m going to say, however, let’s give some coordinates to what we’re talking about.

Bounce Rate

By definition, bounce rate represents the percentage of visitors who visit a single page of a website (according to web analytics tools, such as Google Analytics). It doesn’t matter how much time a visitor actually remains on the page, since there’s no way to get this information without the timestamp of a second visited page. All your tool can tell you is your visitor hasn’t moved to other pages of your site.
So, how could this faint information be used as a ranking factor? Maybe that visitor has read your page for five entire minutes before moving away, and your tool simply doesn’t know it.
Of course, Matt Cutts doesn’t tell us anything new asserting they don’t use flimsy elements such as Google Analytics datas, but let’s think about some of the possible events that might cause the end of a visit:

• Closing the page
• Opening a new window or a new tab in the browser
• Clicking on a link
• Typing a new URL
• Using the Back Button

If one of these actions happens without a second page is visited, your web analytics tool registers a bounce.
Might Google know whether one of these events has happened?
Well, I think if the browser is Chrome they have these informations. But not everyone uses Chrome. And even if everybody did, they would be unuseful per se: did the visitor close the page because they weren’t satisfied of its content or because they found the exact information they were looking for? No way to undestand this without any further informations.
These events could have a slightly bigger meaning combined with the time spent on the page, maybe, but also put together they aren’t a certain signal. If I searched for Obama’s birthday, once I’ve got this information, I would probably close the page. And what if I searched for a list of chinese restaurants and I got a page with a list, I chose one of them and clicked on a link?
Bounce rate means nothing without context. And Google hasn’t context in most cases, so I think Cutts was telling the truth about Google not using bounce rate for rankings.
In a specific case, however, they have a strong signal. This is when a user visit a page and goes back to the SERP. If they open other results this might indicate that the previous one wasn’t relevant for the query: in this case, Google could use bounce rate as ranking factor. And so we arrive to a more specific concept that’s search pogosticking.

Search Pogosticking

In 2008, Bill Slawski reported of a “Search Pogosticking Benchmarks” patent, assigned to Yahoo!, writing:

Search pogosticking is when a searcher bounces back and forth between a search results page at a search engine for a particular query and the pages listed in those search results.

Maybe nowadays, with search engine users that open several results in different tabs, pogosticking is a bit dated concept, but that patent demonstrates that search engine have been interested in how users interact with their results since years ago.
The topic was then investigated again after the Panda Update in 2011, and Jim Boykin wrote a post about the concepts short clicks and long clicks, reporting a brief paragraph from In the plex, Steven Levy’s well known book about Google:

On the most basic level, Google could see how satisfied users were. To paraphrase Tolstoy, happy users were all the same. The best sign of their happiness was the “long click”. This occurred when someone went to a search result, ideally the top one, and did not return. That meant Google has successfully fulfilled the query. But unhappy users were unhappy in their own ways, most telling were the “short clicks” where a user followed a link and immediately returned to try again. “If people type something and then go and change their query, you could tell they aren’t happy,” says (Amit) Patel. “If they go to the next page of results, it’s a sign they’re not happy. You can use those signs that someone’s not happy with what we gave them to go back and study those cases and find places to improve search.”

We know Google track user behavior on SERPs and during the years they have improved their datas storing capacity so I think they can compare SERPs tracking logs with statistical models for an enormous number of queries and define whether the interaction with top results for a particular one differs or not, and how much, from what was expected. If the difference is significant, this could be a negative ranking factor for those results that show a scarce user retention and a bigger pogosticking activity.
On the opposite, a positive flag could be assigned to those results with a minor bounce back.

What about Post Click Optimization?

If you were thinking in a simplicistic manner (high bounce rate equals bad and low bounce rate equals good), you’ve always been wrong. Forcing a user to visit two or three or ten pages doesn’t bring you any ranking benefits: what you’d want to do is avoiding visitors going back to the SERP and “long clicking” another result. The only way you can achieve this? It’s by giving them what they were looking for. If they bounce you should do some post click optimization.
Are you interested in it? Come back in a few days, subscribe my feed or follow me on Twitter, I’ll show you a possible process to reduce search pogosticking (maybe this could help for rankings and maybe not, but giving your visitors the right answers is a good idea in any case).