Bounce rate is a metric that measures the interaction of a user with a page. If the user visits a page and then leaves the page without clicking on anything, then that user has left or “bounced” from the page.
If all visitors left the page without clicking on anything else, then the page would have a bounce rate of 100%, if every visitor of the page clicked on something, then the page would have a bounce rate of 0%.
An assumption on user engagement is that if a user is interested in your content, they would click further into your site. As a result, a low bounce rate is said to be desirable. The question now is, is a low bounce rate a positive ranking factor in and of itself? Will having a low bounce rate mean that you would also get a high rank?
For this test, 5 test pages were created and indexed, and the #3 ranking page was chosen as the experimental page.
Mechanical Turk was used for this test to create the traffic to these pages and to manipulate the bounce rate. Mechanical Turk is a platform where you pay real workers to do online tasks.
For this experiment page, users were instructed to go to the test page, wait 30 seconds, and then click the link on the page. By clicking the link, the goal is to reduce the bounce rate on the page to 0%, or as near as we could get.
For the other test pages, users were instructed to fo the the pages, wait 30 seconds, and then leave the page. With no engagement in any other actions, the goal is to have the test page as close to 100% bounce rate as possible.
We go into this test with the assumption that some users will make mistakes and that any tool used to monitor traffic will be an imperfect measuring tool and may miss some data. However, if we have an experiment page that is close to 0% bounce rate as possible and other test pages with as close as possible to 100% bounce rate, we anticipate the page with the 0% bounce rate to move up in rank if bounce rate is indeed a ranking factor.
Google Analytics were set up on the pages to monitor traffic and bounce rates.
The experiment page moved up to #1 showing that bounce rate is a clear ranking factor. As expected, we couldn’t get the experiment page at 0% and the test page at 100% but there is a clear margin between the bounce rate of the experiment page at 25% and the test pages at 73% to 87%.
A very interesting outcome is that the experiment page recorded fewer visits in Google Analytics than 3 of the 4 test pages. It’s extremely interesting that the quality of the visit would be more important than the quantity.
As the test environment uses a keyword that is not in Google’s database it stands to reason that an appropriate niche targeted bounce rate hasn’t been established for that term yet. As such, the lowest bounce rate will be seen as the best.
As Google gathers more data on a particular term, such as the search tendencies for users searching for that term, it would be reasonable to assume that what is the appropriate bounce rate, the rate that Google would use as a positive ranking factor, would shift.
If you decide to work on your bounce rates, make sure that you take a long term approach to decrease bounce rate over the space of weeks or even months, rather than a large, sudden drop. Determining a ‘good’ or target bounce rate is probably more tricky than determining if bounce rate is a ranking factor.
It seems unlikely that if you call to the top 10 sites for a particular search term, they will happily share their bounce rate data with you.
Start with identifying the type of site that you have and then do some searches for industry standards. Compare those standards to your current metrics and make a plan of action, if necessary.
In this SIA video, Clint discusses this test and his insights on bounce rate.
This is Test number 43. Is bounce rate a ranking factor?
I think we have to be a little bit clear about how bounce rate is measured. And to define the different versions.
If you’re doing a Google search for wooden pipes, and you click on the page, and you immediately go back, that is termed in the SEO community is ping pong. You ping pong and went right back.
I believe Google can measure that because when you do it, especially these days, you’ll click on link, you go to the page, you’ll go back, and then Google pops up related searches underneath the listing that you clicked. Obviously it can see that you’re doing that. So ping pong, as a theory, is actually pretty sound.
And then there’s bounce rate.
They say, the higher your bounce rate, the worse your page will do in SEO. But what if you have a landing page. If you have a landing page and you have a form, they fill out the form and that goes into whatever their systems are. You will have a 100% bounce rate on a really, really good landing page.
A really good converting website that’s generating phone calls can be 80% to 100%, bounce rate.
The way Google Analytics measures is you go website, you read it, and then you go to another page, then that affects your bounce rate percentage. If you go to a website, stay that page and don’t trigger any other events within Google Analytics, you very well could have 100% bounce rate.
So now it goes into let’s trigger other events. What do we do with that? You can actually lower your bounce rate into the 2% range if you had enough events firing off.
Does all of that matter? And is it working? Now, this is kind of a test, that would suggest that bounce rate is actually a thing. And if you manipulate your bounce rate, you’ll get higher rankings, and then it’ll be considered a ranking factor.
Honestly, even with this test, I’m not really convinced just because of the way that it was set up. There’s five identical test pages that were created. The third one was the experiment page. And then they sent Mechanical Turk projects to the web pages to manipulate the bounce rate.
What’s not clear in the write ups… And this is why we do a lot of testings, because we want to get better at the write ups and stuff is…
One, was there a set tasks that they had to do? Was there a time limit that they could be on the page? Do they just go to the page and sit? Were there other triggers set up to manipulate the bounce rate? Scrolling the page could be a trigger and that manipulates bounce rate.
How is that all done? And then was the exact same amount of people sent to each one of the experiment pages? Let’s say all five of the pages and then one of them was a test page, the other four would be the controls… all four of those got like half of the traffic of the test page.
Let’s say that happened. That could manipulate the test results because that test pages obviously getting more traffic and maybe it’s the traffic signal that resulted in the number one.
This is a great example of why a test write up needs to be clear. Why you should have the full specifications. Why you see what happens in the test, and how that was set up. And this is definitely the way we’re leaning towards in the 2021 test results and being a little bit more clear on the on the entire setup and the different factors that are moving it.
Honestly, I don’t think bounce rates is a ranking factor. I can keep mine at 45%. And that doesn’t mean I’m going to rank higher or lower than anything else.
I see that as a conversion tool. I’m more worried about the ping pong effect. Them coming to the site and then leaving and going right back to Google. And it’s a little bit harder to measure that other than they open the site and then they don’t do anything.
They don’t create either the triggers and they close the site. I could probably assume that’s a ping pong.
Then that might mess up my bounce rate. Don’t keep it up closer to 100%. So when I see that, I know that I have a conversion problem and not necessarily an SEO problem.
I’m probably ranking for keywords that are not necessarily what my page is about. And I somehow optimized for those. Or sent the wrong signals to Google and I’m getting the wrong kind of traffic. This happens way more often than we’d all like to admit.
Is the bounce rate a ranking factor? Test says yes. Experienced says no. It’s definitely something we need to look at again.