Actual Bounce Rate vs. Bounce Rate, and Why the Difference Matters for SEO

SMS Text

I’ve written extensively about bounce rate on my blog over the past several years, including how to lower your bounce rate, blog bounce rate, homepage bounce rate, etc., but I’ve never written about how it’s connected to SEO.  Bounce rate is an incredibly powerful metric that can tell you a lot about campaign performance and the quality of your content in a very short amount of time.  If you’re not that familiar with Bounce Rate, it’s the percentage of visits that arrive on your site, visit no other pages, and then leave.

Those visitors simply visit one page on your site.  From a conversion optimization standpoint, the more you can lower your bounce rate, the more opportunity you have to convert visitors.  But although bounce rate optimization is extremely important, it’s not the focus of my post today.  Instead, I want to focus on how bounce rate impacts SEO, and specifically, how ACTUAL bounce rate affects SEO.  More on the distinction between bounce rate and actual bounce rate soon.  First, let’s quickly explore an introduction to quality content from a search engine perspective.

Cutts, Forrester, and Panda

Over the past few years, there has been a lot of debate (and confusion) about whether bounce rate is a signal that the search engines use to determine quality content.  Matt Cutts has probably been asked this question ten thousand times, and has referenced bounce rate when answering some questions about ranking factors.  You can check out the part of the video about page speed to see what I’m referring to.  Then with the rollout of Panda, an important algorithm update that targeted low quality content, bounce rate and other quality signals entered the spotlight.

In addition, Duane Forrester recently wrote a post defining quality content (from a Bing perspective).  It’s a great post, and I highly recommend you go through it thoroughly.  There are some great points Duane underscores.  As part of the post, Duane explains that the engines can monitor “dwell time”, or the time a person remains on your page before clicking back to the search results.   If visitors are clicking through the search listings to your site, and then clicking back to Bing quickly (in just a few seconds), that can be a negative signal to Bing.  He clearly explained they have this on their radar from an SEO standpoint (and that it does matter for determining quality content).

So, when two of the top guys in organic search quality explain that you should keep an eye on bounce rate and quality content, you should probably listen.  But, is it as easy as checking your bounce rate in Google Analytics to determine quality problems?  Is the standard calculation for bounce rate the one that the engines use when determining rankings?  These are great questions, and the answer is most definitely “no” to both questions.  The engines have several better ways to determine actual bounce rate, engagement, etc., as well as how actual bounce rate applies to specific types of content.  This is something that I’ve come across many times while analyzing SEO performance for clients (across industries).  A high bounce rate in your analytics package does not mean that content has a high bounce rate to the search engines.  Sometimes the two metrics might be close, but other times there might be a large discrepancy between them.  Confused yet?

Bounce Rate vs. Actual Bounce Rate

As I explained earlier, I’ve known that actual bounce rate differs from bounce rate for some time.  I’ve had the opportunity to manage SEO for a wide range of organizations, from major brands to local businesses.  While helping these organizations, there were many instances of pages with high bounce rates (in Google Analytics and other analytics packages) that still ranked well in Search.  The content I’m referring to would clearly be identified as high quality, unique, and valuable, but had high bounce rates.  Many of the pages I am referring to ranked in the top three to five listings on page one of Google, Bing, and Yahoo.  That was the case across clients, industries, domains, etc.  So, it didn’t take the most advanced SEO analysis to understand that other factors were at play.  To me, standard bounce rate couldn’t be the metric that engines were using to impact rankings.  It just didn’t add up.

The problem with standard bounce rate is that it’s too simplistic.  Check out the following two situations and I think you’ll get my point.  Person A visits a page via organic search, takes a quick look at a page filled with ads and low quality content, and returns to the search results.  That person spent 6 seconds on the page before returning to the search results.  That’s clearly an “actual bounce” and would also show up in Google Analytics as a bounce.  Next, Person B clicks through a listing in organic search and ends up on a thoroughly written, unique blog post about a specific topic.  The content is supported by great visuals, addresses all the top concerns, provides definitions, etc.  The person ends up spending 21 minutes on the page consuming the content before clicking back to the search results to perform an unrelated search.  Guess what?  To Google Analytics, that’s a bounce.  But to the engines, that’s a much different situation than a bounce.  They know that you spent 21 minutes before coming back to Search (by analyzing dwell time).

The scenario I listed above is a great example of the difference between actual bounce rate and standard bounce rate. This is exactly why you might see pages with high bounce rates that rank well in Search.  They don’t have high ACTUAL bounce rates.  They have high standard bounce rates, which is too simplistic to take true engagement into account.  A quick example of content that might have a high standard bounce rate, but a low actual bounce rate, would be a detailed tutorial.  Someone has a problem, needs help, and the content provides a thorough solution.  A person visits the content, spends a good amount of time consuming the content, is happy with her experience, and leaves.  That’s not an ACTUAL bounce.  That’s a standard bounce.

Actual Bounce Rate = Real Signals to the Engines

The example I provided above covered someone clicking from the Search results to a page, and then clicking the back button in their browser to return to the search listings.  That’s obviously one strong way the engines understand that you bounced, but that’s not the only way.

There are several other ways that the engines can determine if a piece of content has a high bounce rate (actual bounce rate).  Using these methods, the engines can build a more rounded view of your content with regard to actual engagement.  And no, they don’t need to look far.  Let’s take a look at some of the mechanisms that are in place today.  Note, some of what I’m listing below cannot be proven.  I’ve noted when that’s the case.

1. A Bounce Back to Search

Bounce Back to Search

As I mentioned in my example, and what Duane listed in his post about quality content, someone clicking through a search result, only to quickly return to the search listings is a strong negative signal.  But, as I explained earlier, it truly depends on how long that person stays on your page before clicking back.  I’ve analyzed many pages where visitors are spending several minutes to over ten minutes on a piece of content, but that content has a very high bounce rate.  Those pages still rank well in Search.  This makes sense, since the content is ultra-high quality, provides solutions to problems, etc.  The engines know this and understand the actual bounce rate.  They can analyze the dwell time.

2. Glorious Toolbar Data

search engine toolbars

Millions of people have either the Google Toolbar or Bing Toolbar installed.  Although many people don’t know this, they might be passing information back to Google or Microsoft about the sites and pages they visit on the web.  Most SEO’s understand this, but I find most people outside of the industry have no idea.  This is an incredible way for the engines to understand actual bounce rate.  Imagine you visited a page from Search and spent 12 minutes on the page.  Then you either type in another URL directly in the browser or via the toolbar (Google or Bing).  The toolbar can pick this up and understand how much time you spent.  Again, that’s much different than seeing a “bounce” in your analytics package.

Also, if you read the privacy page for Google’s toolbar, it’s clear that the toolbar sends information back to Google about the searches you conduct, the pages you visit, etc.  Check out the section about “Enhanced Toolbar features”.  In addition, you can read the privacy policy for the Bing bar to learn more about the data it sends back to Microsoft.  Once you start realizing how much information those little toolbars are collecting, they start to look a little different to you.  You can almost make out a pair of eyes looking at you.  🙂

3. Chrome


In Chrome, the address bar essentially acts as the Google toolbar.  Auto-suggest works, Google Instant works, and now Instant Pages work in Chrome.  If Chrome is passing your data back to Google, then it has yet another signal about actual bounce rate versus traditional bounce rate.  There’s been a lot of debate about how much information is being passed to Google from Chrome, and Matt Cutts addressed this in a blog post back in 2008.  A lot has changed since then with Chrome (functionality-wise), and you can read the Chrome privacy policy to better understand the data that is sent to Google.  To me, the more functionality that’s available in Chrome, the more opportunity for data to get passed back to Google.  And that means more data that can be used to determine low quality content, high actual bounce rate, etc.  I’m ok with this, but that’s me.

4. Instant Previews

Instant previews

Earlier this year, I wrote a post about how Instant Previews could provide quality signals back to the mothership.  I like to call this Instant Bounce Rate, which is someone clicking an instant preview, viewing a snapshot of the page, and then not clicking through.

That’s definitely a signal to Google that the content isn’t meeting visitor expectations (at least based on the snapshot viewed in the search results).  Google might be incorporating this data into its actual bounce rate calculation.  You can think of Instant Bounce Rate as a micro bounce rate metric, since the person never actually visits the page at hand.

5. Google Analytics Data Sharing (when combined with other signals listed above)

google analytics

Disclaimer: I have no hard evidence that Google uses data from Google Analytics in its calculation for actual bounce rate, but it’s entirely possible.  When you agree to share your data with Google products, you agree to share your analytics data with Google to help it improve its services.

By providing this data, it’s just another possible way for Google to better understand your content and engagement. For example, a certain page might have a high standard bounce rate, but the average time on page can give them more information about how non-bounce and non-exit visits consume the content.  You can check out my post about time on page and time on site to learn more about the metrics.  Again, I’m not saying Google is doing this, but I’m also not saying it isn’t.  The data is there and companies are freely sharing it with Google.

Key Takeaways

I hope this post helped explain some of the mechanisms that are in place for understanding actual bounce rate versus standard bounce rate.  As I explained throughout this post, the search engines have the ability to collect in-depth data from a number of sources to determine your actual bounce rate, which can then be used for ranking calculations.  So, make sure you don’t simply analyze bounce rate in your analytics package and think that metric has an impact on SEO.  Instead, dig deeper to better understand the real metrics being used, because that’s exactly what the engines are doing.

Now, if you ended up reading this entire post, and then leave, you should know you didn’t really bounce, right?

Get the latest news from Search Engine Journal!
We value your privacy! See our policy here.
  • Anonymous

    Excellent article, Glen.  The search engines’ assessment of Bounce Rate is indeed too simplistic.

    Here’s another example: Suppose you’re eager to know the results of a soccer game, word definition, closing stock market price or celebrity trial.  If you Google for this, find the answer immediately and go back to what you were doing beforehand, that is considered a bounce — even tho’ the page gave you exactly what you were looking for (immediately)!

    To get this right, the search engine would need a mechanism that could distinguish between right/wrong answer.  Perhaps a ‘Was this helpful?’ checkbox or the more ubiquitous (but more ambiguous) Like or +1 button.  It’s certainly not fair to penalize the page just because it didn’t take you 20 minutes to find what you were looking for.

    • That’s a great point.  Actual Bounce Rate has to take the “type of content” into account.  I briefly mentioned that in the beginning of my post, but didn’t elaborate. You provided several examples of content that should rank, provided answers, etc., but probably has a 95%+ standard bounce rate. That can’t be held against it… Again, great point.

      • Hi Glenn,

        Great post, thanks for that.

        I would have to agree with the point made by SemanticAd also. This is an issue that has been of concern to me ever since first whispers began about bounce rate being used as a ranking signal. I think this is a definite flaw in the logic of using this data. I have a client for whom this is a constant concern. They are a plumbing company who offer 24 hour emergency service. The primary function of their website is to provide their phone number for clients as quickly and easily as possible.

        The idea that their rankings might suffer because the site does exactly what it is intended to do highlights another instance where it is not a good idea for Search Engines to try to shove every peg into the same shaped hole!

      • Yes, great point.  But, the actual bounce rate will differ from the standard bounce rate.  In addition, I truly believe the engines understand the “type of content”, and can adjust based on that. For example, quickly checking the weather or getting a quick answer to a common question. Regarding your client, the person would probably still be on the page for 30+ seconds, which can be a signal that they didn’t immediately bounce. 

    • That’s a great point.  Actual Bounce Rate has to take the “type of content” into account.  I briefly mentioned that in the beginning of my post, but didn’t elaborate. You provided several examples of content that should rank, provided answers, etc., but probably has a 95%+ standard bounce rate. That can’t be held against it… Again, great point.

  • Anonymous

    Another Outstanding Article Glen! Do you have screen shots of how to tell the difference between regular and real bounce rate in analytics?

    I think knowing what filters to look for would be really beneficial.

    • Thanks Ben. I’m glad you liked my post. Unfortunately, you can’t check any one metric in Google Analytics (although I wish you could!)  However, one metric that can help is avg time on page.  That will at least give you an understanding of how visitors are consuming your content (ones that stay on your site, since the metric excludes exits).  For example, you might see a standard bounce rate of 90%, but an avg time on page of 6:34.

      In addition, you can also check our real-time analytics packages that take a deeper look at engagement.  For example, Clicky’s bounce rate can incorporate time on page for one page visits by using “pinging”. You can learn more here:

      I hope that helps.

  • Anonymous

    Another Outstanding Article Glen! Do you have screen shots of how to tell the difference between regular and real bounce rate in analytics?

    I think knowing what filters to look for would be really beneficial.

  • I’m always looking for ways to decrease my bounce rate for my websites and these post definitely made me look at the numbers a lot more closer.

  • Anonymous

    Thanks for the helpful read!  Any business’ goal is to have less of a bounce back rate and you have brought up some good points.

  • Great info and a lot of help.  I had no idea that the Instant Preview affected bounce rate :/ 

  • Great info and a lot of help.  I had no idea that the Instant Preview affected bounce rate :/ 

  • frenchybee

    Great analysis!
    I now have a stupid question: what is a good bounce rate, what kind of number are we talking about here?

    • That’s not a stupid question! It depends on a number of factors, including the content, campaign, or traffic source you are analyzing. Aggregate bounce rate (at the site level) is almost worthless, but tracking specific campaigns, sources of traffic, and pieces of content can tell you a lot.  Generally, a bounce rate of under 35-40% is strong, where bounce rates of over 60-65% start to raise concern.  Again, bounce rate optimization wasn’t the focus of my post, but it’s extremely important…

  • news update Actual Bounce Rate vs. Bounce Rate, and Why the Difference Matters for SEO

  • Wonderful Post Glenn, It seems good to get in depth idea regarding bounce rate. It is vital for every website. One point i like the most that you explain “Bounce is differ as per the Google analytic and Search engine”  Which you called bounce and actual bounce.

    I just want to know it is vary by every search engine and how we can track those actual bounce. Thanks for your any input.

    • Thanks Hiren. Unfortunately, you can’t look at any one metric when determining actual bounce rate (although I wish you could…) As I explained in another comment, I would look at avg time on page and then check out some real-time analytics packages that can track time on page for single visits. Clicky’s bounce rate metric can do this, since it uses “pinging”. I hope that helps.

      • Thanks for the reply Glenn, I read your another comment as you mention and let me more focus on Clicky’s bounce rate metric. I think that will definitely help me out.

  • The Actual Bounce rate always matter cause the main thing should the bounce rate called the traffic should be converted to the sales or any motto that we are running for.So bounce rate matters most in the overall traffic generation to sites.

  • The Actual Bounce rate always matter cause the main thing should the bounce rate called the traffic should be converted to the sales or any motto that we are running for.So bounce rate matters most in the overall traffic generation to sites.

  • Great info and a lot of help.  I had no idea that the Instant preview affected bounce rate

    • Hey, thanks. I’m glad you found my post helpful. Yes, instant previews can be a signal that Google uses to better understand the quality of content. My post about instant previews as a signal explains more about this. You should check it out.

  • I love this post, it’s a useful idea as opposed to the theory and conspiracy that has plagued the search engine blogs recently.

    In the top image, did you customize Google Analytics to show Actual Bounce Rate or is that just for representative purposes? If you did customize can you show the steps you took to do it?

    • Thanks Nate. I appreciate it.  Regarding the image, it’s not real. I put that together for the post. 🙂 That said, wouldn’t it be great to have another column for actual bounce rate?

  • Yet more great content for SEJ.  I am redesigning my site right now and  appreciate the ideas posted here.


  • Preetam

    This post discover the answer which i am looking for it from many times but i can not get the answer but today Glenn give me in brief answer for it so thank you so much.

  • Anonymous

    Great points ! Absent the Instant Preview capability and, of course, ignoring Google Analytics, it seems to me that Search Engines probably don’t even have the capability to accurately detect what is and is not a ‘bounce’. From the Search Engine’s perspective, you clicked on a URL, and returned from a URL. If those URLs are not identical we know it wasn’t a bounce. If the URLs are identical with a very short dwell time we can say that it was probably a bounce. But as the dwell time gets higher, since the search engine doesn’t know if you visited other pages before returning to the original URL, it seems to me that it just cannot know one way or another…

    • Thanks for your comment. Regarding:

      “But as the dwell time gets higher, since the search engine doesn’t know
      if you visited other pages before returning to the original URL…”

      But they can know where you’ve been.  You can check out the various methods listed above. The engines won’t know this for everyone, but they don’t need all of the data to get a good idea for the quality of content.

      • Anonymous

        OK – reread the section on toolbars. I was thinking of the Search Engines in isolation and suspecting they highly dependent on the referrer field. But you are right – with local storage, like the toolbar, almost anything is possible.

  • The bounce rate is computed as a ratio. The numerator for the ratio or fraction is the total number of visits that resulted in a bounce for a specified period. The denominator is the total number of visits during the same period.

  • I was just going over this with a client the other day. This post will definitely be headed their way to clear up the relationship between bounce rate and SEO. Thanks!

    • Hey, no problem Jon-Mikel. I’m glad my post was helpful. Companies definitely need to understand what Actual Bounce Rate is. If not, the standard bounce rate metric could be misleading…

  • Kevin

    I’m not sold that bounce rates matter for SEO. Depending on your site/product/service, you may actually want your bounce rate to be as close to 100% as possible. 

    An example of a site that might want a 100% bounce rate could be a stock monitoring site with a personalized dashboard. In that instance, such a site might perceive clicking around as a sign of confusion, and a near 100% bounce rate for returning visitors might suggest that those visitors are finding everything they need on a single page.

    The idea in search should always be to give people what they’re looking for as quickly as possible. Google believes this. They want page views per visit on their site to be as low as possible. People should find their information quickly and easily. I have a hard time believing that they would believe that a low bounce rate, high pageviews per visit, or time on site are necessarily indicative of quality without taking into account additional factors. 

    • Thanks for your comment Kevin. I mentioned “type of content” in my post and in several of the comments here.  I believe the engines understand the type of content and can adjust based on that. For example, checking the weather, a stock quote, getting a quick answer to a question, etc. But -actual bounce rate- can be an important signal to the engines. That’s one of the reasons Duane Forrester listed dwell time in his post (referenced above). 

  • Great Post, this has always been something I wanted to find out, but of course searchenginejournal provides me with the answer once again

  • Curtis

    Like many before, thanks so much for this article! It cleared up quite a few things for me in trying to understand the difference between the two metrics.

  • Anonymous

    Thanks Glenn,

    Now I Understand a little bit more about bouncing rate. I was a little confuse before, but thanks again.

  • Great article and I like that you shared different instances and calculations of bounce rate. I still think one of the key takeaways, for me, from the article was simply

    Quality Content

    When you focus on content that is unique, adds value (i.e. as mentioned diagrams, good links, even videos on a page) you are letting the search engines and visitors know that your content is worthy of their time. While many people focus on “tricks” taking the time and energy to share quality content will certainly lower the bounce rate and keep the search engines moving you up in the rankings.

  • Hi Glenn,
    Interesting post. Content rich sites such as news, information and nfp sites that use Google Analytics can have poor BRates. This is because the standard GA tracking code cannot distinguish between a `real` bounce ie someone who came to a page and left quickly and someone who came, spent say 20/25 mins reading and then left. However by adding to the standard GA tracking code users can track the time on page for single page visits. This allows you to define a Bounce as someone who stays 10/20 seconds, whatever time you decide, on a page. IF Google is using your GA data as an SEO signal this data could help your SEO efforts. I have written a longer post about this on my blog.

  • Avishek

    We work with the analytics tool called Omniture Insight and due to a certain feature in it (or maybe how things are defined in it), it is not possible to measure the time spent on site for a bounce visit. We would rather look at whether or not a visitor interacted with any of the content on the site such as a lead form, a video or a download before leaving the page. I think that gives us a fair indication of what the “actual bounce” is for our site.