Bounce rates as an SEO factor

It’s been a long time prediction of many SEO’s that Google will start to watch out for bounce rates when indexing sites.. The underlying thought process being that a high bounce rate indicates a site which isn’t as worthy or interesting as a site with a low bounce. I’ve previously explained bounce rates and provided some advice on lowering your bounce rate.

Blackhat SEO showed us some interesting stats to show what they have found back in November in relation to bounce rates, but lets not get too carried away yet.

This article on bounce rates and google is definitely worth a read guys. I’m tending to agree with David, that behavioural metrics may affect SERP positions slightly but that they should be taken with a pinch of salt. They are as Matt Cutts has stated in the past “fuzzy or dirty” metrics that could be easily spammed in the future, and move away from the more algorithmic stuff that Google currently prides itself on.

Big thumbs up Dave to someone putting some perspective to this stuff within  the SEO industry. There are already too many myths floating around the web with regards this sort of thing.


Submit a Comment
  1. The only way Google would know your site’s bounce rate would be if you have Analytics code in your pages. If Google started penalising sites for a high bounce rate surely people would stop using Analytics and use some other analytic code. Equally the rubbish site that doesn’t have Analytics code would end up with an advantage because Google would have no idea what its bounce rate was.

    Saying that, I agree that if you have a high bounce rate it should be looked into regardless of what Google may think of it. Either the wrong people are coming to your site or something is wrong on your site – either way, it needs fixing.

  2. Hi Dean,

    Thanks for your comments. Google would be able to work out bounce rates via both Analytics and their site itself.

    For example lets assume I search for “shoes” the first result on Google takes me to a spam site. I click back instantaneously and Google is able to detect that I haven’t viewed much of the site in question. This could be used in combination with Analytics data.

    But as you say – a high bounce rate needs fixing regardless of its impact.

  3. Hi Paul… wow… Dave 1 Rest of the Loonies 4000

    Thanks for following along. I have spent the weekend talking to my IR engineer type friends and have read a TON of related research on all of this. I hope to get that up by the end of next week into the week after maybe…

    Ultimately, there is hope for the technology, but there are still hurdles for gleaning real actionable data (noisy) and very little work done on dealing with spam. In many cases there is more sense in a corporate enterprise search level where spamming is far less likely.

    So, from what I am finding in further research, it is as Matt said, noisy and spammable.

    (by the way, found my way here testing Joe Hall’s buzz monitoring toy – Who’s Talking – Google it, no need her me to be link dropping on my first visit :0)

    Thanks for the support tho… appreciated.

  4. Google can easily calculate a ‘type’ of bounce rate that is more commonly referred to as pogosticking. And it has nothing to do with Google Analytics.

    Remember that Google tracks every click on natural search results. There’s a subtle redirect that captures this information and records it. So, we know they keep track of our clicks, and they do so based on each query.

    One would have to assume that they’re writing a time stamp for each of these clicks. Hence, the ability to determine pogosticking behavior isn’t that outlandish.

    The mechanism to determine if a result is statistically outside the norm based on rank is the rub, but one that I feel Google and others are likely able to tackle.

    I’m not saying it’s a heavily weighted factor, but I’d be surprised if it wasn’t one of those 200+ signals that Google uses in the search algorithm.

Leave a Reply

Your email address will not be published. Required fields are marked *