Bounce Rate Questions

Marketing Pilgrim posted some questions about the bounce rate proposition from today’s earlier post:

Well, from their graphs, I see that their overall traffic, Google traffic and bounce rate all declined at the same time. But if I remember my Google Analytics correctly, bounce rate going down actually means that fewer people are leaving your site after viewing just one page, so your site is stickier. So Google rewards sites that people leave quickly with higher rankings and thus more traffic?

No. You are a little confused. Google leaves in tact the traffic that doesn’t bounce and takes away the traffic that does to test it across similar sites. It does not reward sites that make people leave, it does just the opposite (at least during these test periods).

I’m not convinced. Correlation still isn’t causation

True, but we’re not counting sun spots here. ;)

couldn’t it be that Google traffic dropped off (for whatever reason) and then a higher proportion of visitors are the loyal visitors who weren’t bouncing anyway?

No. These are not loyal visitors who weren’t bouncing anyway. These 2 graphs should clear it up for you a little more

bounce rate and google

Note the date ranges on each.

SEO bounce rate

Note the % New Visitors on both: A virtual tie (83% vs 80%). But if you look at the bounce rate from google traffic, you will see they lowered the overall bounce rate during the test period from 63% to 41% – which is statistically very significant.

Thanks for the feedback and be sure to post other questions or concerns if you got ‘em!

Both comments and pings are currently closed.

17 Responses to “Bounce Rate Questions”

  1. eugene-uk says:

    Hmm, but Google only knows the real bounce rate of a site if it has GAnalytics installed. That means unknown sites won’t get penalised in any case.
    Penalising the ones which are known is not fair.
    Only one option left: Google gives a ranking bonus to those who have GAnalytics installed and have a low bounce rate (people tend to stay).
    Right?

  2. _brad101 says:

    @eugene-uk: Google uses it’s toobar to measure bounce rates too: it measures the elapsed time between PR requests. It also uses this measurement to monitor redirection; that’s why delays are often used.

  3. SEOAlchemist says:

    This is fascinating, and the evidence is mounting. Great work! I’ll be following this one.

  4. Bounce rates are important just ask Tigger

  5. Gotan Raider says:

    Yow Brad just the toolbar? :)
    When you are logged in gmail or any other of the gazillion services G owns, can’t they track your behaviour?

    And more recently… Chrome.

    There’s an underlying concept behind all this and no one has mentioned it so far: is Bounce Rate a true indicator of positive user experience?

    Is quantitative data so straightforward and does it allow you to take conclusions that apply to all kinds of websites?

  6. _brad101 says:

    Arrrrggg…. Chrome! At the moment, as far as I can tell, G cannot automatically detect cloaking (content delivery not IP-based redirection)… so I wonder if Chrome is an attempt to combat it, plus monitor surfer behaviour even more! To protest, we must begin the Google Rebellion: every one should exclude Googlebot from their sites, especially in the financial niche :-)

  7. Dudibob says:

    @Quad – Are you saying that Google is effectively split testing it’s results for certain users (420,000 of them!) from your site to another one because they bounce quite highly on your site?

    If that’s true then using rankings as a metric is dead for sure!

  8. veke says:

    Hi, this is amazing.. I can not understand why google is going to do this…

    I think in this way:

    - A site get a low bounce rate for a particular keyword
    |_ Then probably that site is a good resource for that kw
    |_ If i would be google i will give that result an higher ranking.

    My ideas are the opposite of what google is doing…
    BUT

    Another way to interpret these data could be the following one:

    Suppose that you 30.000 Unique_visitor_per_day WebSite is taking a lot of time for loading and usually ( when you got 30 000 visitors every day ) it would be annoying to wait.. (personally if i’m not so desperate i will not wait more than 4 second on a web site for loading ) ..

    For every Day you get more than 30 000 visitors which a 20% are lazy enough for leaving the page if it is loading slowly.

    Now, Suppose that you got google-kicked about 3 position down for the most relevant kw .. You are going to get lower visitor from google…

    Now lets suppose that the lower traffic you are getting is leaving your website breathe a bit And consequently your webpage is loading much more faster…

    The 20% lazy-visitors are not leaving your website anymore… ( So you are getting a Lower Bounce rate)

    .. BTW another point is:
    Your graph is not proving anything.. You are proving that your website has a lower bounce rate when it gets less traffic … It would be a step by step process:
    1- You lower the bounce rate
    2- You get more/less traffic from google

    The process ‘1′ should be done before the ‘2′ one.

    If this comment was idiot tell me it..

  9. LebossTom says:

    once again, i disagree…

    visitors from google who always pick #1 result have an higher bounce rate then visitor who take time to read SERPs and pick number #3 or #4…

    What if in this case you get an temporary ranking penalty and you’re mixing causes & consequences…

  10. Steward says:

    Do you know, what exactly the bouncerate is? I’m not sure and give some examples:

    - visitor uses the back-button (imo that’s a clear bounce)
    - visitor clicks a link, target page opens in the same browser window (clear bounce)
    - link opens in a new tab/ window – did the visitor leave the first page or not?
    - visitor clicks on an ad and let’s say, it is from adsense. Is this a bounce?

  11. Contempt says:

    Very good Q&A as the follow up. I’ve suspected this myself for a while, but this puts the icing on the cake. Keep up the good work man.

  12. QuadsZilla says:

    I will add that it may not be bounce rate per se, but something that is a close corollary to bounce rates. So, for example, if someone clicks another SERP on the same search 30 secs to 2 mins after clicking the initial site, or any other trackable metric that indicates that the surfer did not spend much time on your site and is looking elsewhere for an “answer”.

  13. Steward says:

    QuadsZilla, thank you! This is for a search engine the most secure and easiest way to count bounce rates. It can be so simple indeed. :)

  14. LebossTom says:

    @QuadsZilla : ok, that’s sounds more realistic cause bounce rate isn’t a good metric at all to figure out user experience.

    sorry for my terrible english :)

  15. eagerterrier says:

    Surely what this means is that when google visitors drop off, bounce rate is reduced?

    Hence, people who have bookmarked your site, or who come from referring sites are more likely to stick around than google tarts.

    I don’t think there’s anything sinister here, as the bounce rate goes down, not up…

  16. [...] After my last 2 posts on bounce rate, Many more questions have been popping up about the bounce effect. [...]

  17. [...] Peter vom Blogpingr teilte unsere Meinung nicht. Er antwortete im Artikel , das einige seiner User wesentlich mehr Pings absetzen wie Andreas. Er verwies als möglichen Grund auf einen Artikel des SeoBlackHat Blogs , der sich mit der sogenannten Bounce Rate Frage beschäftigt. [...]