homepage Welcome to WebmasterWorld Guest from 54.161.147.106
register, free tools, login, search, pro membership, help, library, announcements, recent posts, open posts,
Become a Pro Member
Home / Forums Index / Google / Google SEO News and Discussion
Forum Library, Charter, Moderators: Robert Charlton & aakk9999 & brotherhood of lan & goodroi

Google SEO News and Discussion Forum

    
Google Uses Bounce Rate!
TheMadScientist




msg:4294026
 5:34 am on Apr 8, 2011 (gmt 0)

NOT!
Made You Look Though, Didn't I?

Okay, maybe Google does use bounce rate ... I'm sure they have some extra hard drives, so maybe when they don't have anything better to do they put the bounce rate on those drives and 'organize it', just for fun.

All too often I hear about Google using Bounce Rate and it makes no sense, except for people looking at the web and statistics from such a narrow view point they fail to see the bigger picture, so let me highlight some of the huge issues with Bounce Rate as a ranking signal...

The bounce rates I'm using to highlight the issue are for every time it happens...

###

If a visitor lands on a product page and then adds an item to a shopping cart, views the cart, proceeds to checkout, finishes the transaction and lands on the 'thank you' page, they have 4 page views.

BUT

If a visitor lands on a product page and then adds an item to a shopping cart, views the shopping cart, which this time is remotely hosted, then proceeds to checkout from a different 3rd party processor, and arrives at the thank you page, the bounce rate from the item page is 100%, the bounce rate from the shopping cart is 100%, the bounce rate from the 3rd party processor is 100%, the bounce rate from the 'thank you' page is 100% and every single page in the chain was part of the correct answer, so a 100% bounce rate was ideal.

There are also many different variations of the preceding, which could change the bounce rate people see in their stats.

###

If a visitor lands on a content page with the content continued on 3 different pages, and spends 30 seconds scanning the 1st page for the answer they were looking for, then clicks to the 2nd page scans for another 30 seconds, then clicks to the 3rd page and scans for 30 seconds again, not finding the answer, the bounce rate of the initial landing page is 0%, the bounce rate of the second page is 0%, the bounce rate for the 3rd page is 0%, and the answer is not found.

###

If a visitor lands on a content page with all content contained on a single page, scans for 10 seconds, finds the information they were looking for and clicks back to 'search deeper' on the subject, the bounce rate for the page is 100%, even though it contained the correct answer.

###

These examples are the exact opposite of many of the conclusions drawn by those who think Google can or must use bounce rate, and they're only a few of the examples I can think of ... People bounce from pages for all different reasons, and sometimes the bounce could indicate the answer was exactly what the person was looking for and other times it could indicate it was no where close.

Google also does not have access to enough consistent and inclusive information wrt bounce rate to apply it to sites as a whole, imo ... To apply a factor to all sites and pages they have to have data for it, and thinking they have bounce rate data for all the pages online is, uh, well, politely, naive ... There are about a trillion web pages last I heard, and if we give Chrome installations a generous 100,000,000 (one hundred million) installations, then every single Chrome user would have to enter through 10,000 unique pages each, without any overlap before a single bounce rate could be determined for each page.

It's not like a template where it's the same site wide, or writing level of an author who publishes a site, which is consistent and either can be detected with a bot. Bounce rate is page specific and the best thing they have to go with to get the number is Chrome, and even getting the number from Chrome once borders on impossible, so how could they possibly have statistically applicable data for all pages in order to apply it as a factor? They couldn't, imo...

It's the reason I believe them when they said speed was only a factor used as a tie breaker ... The toolbar is what's used to gauge it, and although it's possible the toolbar has visited every page on the web, to derive usable statistical data web-wide, there needs to be more than one visit, so 'okay, we have speed numbers from these pages and not or slower numbers for the page it's tied with, so we'll give the nod to the one with the numbers or the best speed number', makes sense, but to try and gauge speed and apply it to every page on the Internet seems a bit of a stretch, so I can see the tie break factor, like they said, but more than that? Hmmm, idk if speed is a major factor at all, or if it will ever be that huge of a factor in the rankings...

People who think Bounce Rate can be applied by Google seem to think Google has access to the stats of every site on the Internet, and can somehow determine if someone bounced because they got the right answer and didn't want to do further investigation, got the wrong answer and needed to look somewhere else, had to look at 3 pages before realizing the answer wasn't there, which is worse than a page where it's obvious the answer is not apparent after a single view, because it's a colossal waste of time to view 3 pages only to not find the answer, or even if they bounced because they got a phone call and had to leave so they shut down the computer.

There are too many different reasons people bounce after a single page view or even don't bounce after a single page view, which could either be 'found the answer and kept reading the article', 'didn't find the answer on page one, so they continued to page two, and finally left after not finding the answer on page three', 'found the answer on page one, but returned to the results to get a confirmation', or some other reason ... IMO It's a very limited view point to think a low bounce rate is good all the time, because to me, in many situations it says you made the searcher do more work than if they had just found what they were looking for on the page they landed on.

Bounce Rate is a silly factor for about anything except internal use, imo.

I've stated previously, I think in some situations they could use 'click-back, re-click' rate, but then it would most likely apply to 'anomalies' where statistically the click-back re-click rate of a single page in the SERPs doesn't match that of the other surrounding pages, and the ONLY way to know which that is is to have the stats from the SERPs, meaning you the site owner don't have access to the click-back re-click rate of the full page of SERPs to know what's normal, good, or bad, so imo it's really only something you might be able to use internally knowing your site and what queries were used to reach a page.

Beyond internal use, imo, anyone who says a low bounce rate is good really hasn't thought the matter through and tried to apply it across the board in all situations and to all types of sites, because sometimes, the higher it is the better, but of course the converse is also true, and in a situation where either/or applies, it's usually better to use something else.

[edited by: TheMadScientist at 6:28 am (utc) on Apr 8, 2011]

 

tedster




msg:4294038
 6:19 am on Apr 8, 2011 (gmt 0)

I actively use bounce rate for any site I work with and I always have. That experience tells me that Google cannot possibly try to use bounce rate as a factor to compare pages or websites.

TheMadScientist




msg:4294063
 9:54 am on Apr 8, 2011 (gmt 0)

I think if people look at their surfing habits they might see where there's no way Google actually uses it too ... For example, my personal bounce rate is close to, if not, 100% for: w3, wikipedia, expert-exchange, ehow ... I bounce from the 1st two, because I usually find the answer on the 1st page I view, but I bounce from the second two because I can't find the answer on the page.

Sometimes I'll bounce from wikipedia or w3.org right back to the SERPs and click on a site I haven't seen before, just because I have a few minutes and like to see what's new, so even my click-back re-click isn't really 'pindownable'.

Where do I have a lower bounce rate? php.net ... I often have a lower bounce rate, because I usually research a bit more when I have a question, which that does not necessarily mean the 1st page I viewed was incorrect, it may have been, but it also may not have been, because I don't copy and paste the specific 'thing' I'm trying to do into the query, I generally type a query for a known function and see if there's anything related listed if I want something 'the same, only different' or with a slightly different application ... Sometimes the 1st page answers the question, sometimes I see another function that might fit the specific application better, sometimes I feel like reviewing a different function I use occasionally to see if there have been changes since I last used them, which may have been a year or two ago.

There is really no definitive way of knowing what my 'bounce rate' translates to based on my surfing habits, which is what someone has to try to figure out to use it externally, and to even be able to do that, they have access to all the data they need in the first place ... Sometimes I'll click on the references from a wikipedia page, rather than another page on the wikipedia site to get more information on something specific, or make sure the information is accurate, which is a bounce rate of 100% every time it happens ... It doesn't mean the wikipedia page was a 'wrong result', it just means I click somewhere besides another wikipedia page from the wikipedia page ... If someone doesn't click on a subsequent page of a site, it's a bounce, and there are so many reasons people could only view one page, or more than one page, to try and determine which is what and what should rank where because of a view on a single page, or more than one page is, imo, time that could be spent doing many other more productive things, like reformatting the hard drives the bounce rate is stored on, then re-saving it to them from another source, just for fun...

[edited by: TheMadScientist at 9:59 am (utc) on Apr 8, 2011]

aristotle




msg:4294064
 9:56 am on Apr 8, 2011 (gmt 0)

I agree that Google probably doesn't use bounce rate or click backs as ranking factors. But there are some other aspects of user behavior that they could use. For example, here are some possible positive signals

-- A user bookmarks a page as a favorite.

-- A user saves a copy of a page on their hard drive.

-- A user returns to the same page later.

In my opinion one of the main reasons Google created its Chrome browser was to be able to collect this type of information. I believe that they probably use it.

NOTE: When I say that Google probably uses some signals from user behavior, I'm not talking about bounce rate or click-backs. Instead, I'm talking about signals like the three that I listed above. I'm not talking about bounce rate or click backs.

indyank




msg:4294072
 10:35 am on Apr 8, 2011 (gmt 0)

Why do you cheat with that kind of title? I would have bounced off this thread if it were not WebmasterWorld and the mad mad TheMadScientist :) and if i had bounced off without making this remark, panda wouldn't love this page and bury it deep.

Shaddows




msg:4294178
 2:39 pm on Apr 8, 2011 (gmt 0)

TMS, I rarely disagree with you, but I must suggest the situation is slightly more nuanced than "Google doesn't use bounce rate"

Personally, I would imagine Bounce rate is a strong signal. Immensely strong. Just not as a RANKING factor.

Given that Google has already profiled the searcher (intent, habit, stage), bounce rate will help CATEGORISE the site they landed on. "Success" might be a bounce.

Once you start thinking about what Google already knows about the referal, especially what they statistically know about a set of referals, bounce rate becomes substantially less noisy. Its about predictions and variance, stuff Google goes well.

Unfortunately, from the site owner perspective, you just see "Bounce Rate". Which is meaningless.

netmeg




msg:4294185
 2:51 pm on Apr 8, 2011 (gmt 0)

Good post, TMS, and I agree wholeheartedly. My best performing sites have really really high bounce rates (as defined by, say, analytics), and they're still gaining.

randle




msg:4294201
 3:14 pm on Apr 8, 2011 (gmt 0)

bounce rate will help CATEGORISE the site they landed on.


Categorize as what?

StoutFiles




msg:4294202
 3:19 pm on Apr 8, 2011 (gmt 0)

I bet they DO use bounce rate for other categories, possibly to help flag MFA sites for manual inspection.

rlange




msg:4294206
 3:30 pm on Apr 8, 2011 (gmt 0)

Then you have people, like me, who open search results in a new tab, scan the page or check out other pages on the site, then close the tab. I would provide Google with no negative signal; they would see me click on the result and not return. Even if I clicked on a second result, that can't necessarily be taken as a negative for the first result.

--
Ryan

Shaddows




msg:4294213
 3:41 pm on Apr 8, 2011 (gmt 0)

Categorize as what?

To drop into site-type (or page-type) buckets.

So, say you had an info site. Normal index data implies it is an authoritive source of information. User data, including bounce-rate, shows it satisfies casual users ("bounce" but not reclick), with while causes query refinement or non-bounce in "researchers". This would look like a page with good, clear information that might provoke further enquries.

Same traffic set, opposite behaviour. Researchers non-bounce or new-search. Casual searchers bounce and reclick. Might be a data-heavy or technical (jargony, academic, etc) page unsuitable for casual reading.

Or there might be a larger than expected non-bounce and bounce-reclick, indicating index data might need rechecking. Might be well-linked MFA.

Again, bounce-rate only makes sense when you have profiled the referring traffic. Google has, the site hasn't.

Shaddows




msg:4294221
 3:43 pm on Apr 8, 2011 (gmt 0)

Then you have people, like me

Yep, they have a steady proportion of people like you, and statistically model you.

They have a much higher proportion of vanilla users, who don't obfuscate their data, deliberately or otherwise.

randle




msg:4294233
 4:08 pm on Apr 8, 2011 (gmt 0)

So, say you had an info site. Normal index data implies it is an authoritive source of information. User data, including bounce-rate, shows it satisfies casual users ("bounce" but not reclick), with while causes query refinement or non-bounce in "researchers". This would look like a page with good, clear information that might provoke further enquries.

Same traffic set, opposite behaviour. Researchers non-bounce or new-search. Casual searchers bounce and reclick. Might be a data-heavy or technical (jargony, academic, etc) page unsuitable for casual reading.

Or there might be a larger than expected non-bounce and bounce-reclick, indicating index data might need rechecking. Might be well-linked MFA.


And your saying this is a part of the present algorithm?

Shaddows




msg:4294238
 4:18 pm on Apr 8, 2011 (gmt 0)

I'm saying that multivariate testing happens. I'm saying that profiled traffic sets are given profiled SERPs (organised site sets), and the resulting user behaviour is fed back into the system. There was a thread about it a while back.

You would do this to test SERP satisfaction (NOT site or page specific) as well as split testing the specific results. Bounce rate would be one of many variables- but only to be interpretted in context of the traffic fed, against expected norms for the page/site profile.

TheMadScientist




msg:4294291
 5:56 pm on Apr 8, 2011 (gmt 0)

Personally, I would imagine Bounce rate is a strong signal. Immensely strong. Just not as a RANKING factor.

Here I was trying to make a post in English, generic enough so all the people who think bounce rate is a ranking factor could stop thinking that and stop posting about it every time there's an update and would stop doing silly things, such as splitting a perfectly fine page into 3 so their bounce rate would be lower, and along comes Shaddows who wants to confuse everyone and split hairs ... LOL

If there's so much as one post in the next update thread about how someone doesn't rank as well anymore and they include their 'really low bounce rate' as some positive about the site you're soooo getting a -1 it's not even funny.

And, yes I sensationalized the title, just like VanessaFox sensationalizes a bit now that she's a journalist, because I wanted people to read the thread and try to 'get' that I don't care what the size of your bounce rate is when you're telling me how great your rankings aren't but should be...

<NewTitle>
Google Uses Bounce Rate As a Ranking Factor!
</NewTitle>

tedster




msg:4294294
 6:08 pm on Apr 8, 2011 (gmt 0)

Hats off to Shaddows - he split some fine hairs.

Just so he doesn't split any of mine, I don't have as many as I used to

The point I see is that Google is always profiling data of any kind they can get their hands on - without rolling it into the ranking algorithm. Sometimes they find some new and strong correlation factor, and that might get rolled in.

We hear from Google engineers that bounce rate is too noisy to use as a ranking factor. How do they know that? Clearly they checked it out - something we all should be doing with our theories, too.

SanDiegoFreelance




msg:4294300
 6:17 pm on Apr 8, 2011 (gmt 0)

Bounce rate data is noisy (per google) it is also easy to game -- Spammers could easily reduce the bounce rate to near zero by doing javascript redirects.

It is also provides some really good data to determine the quality of the SERPs. I suspect the are or will at some point use it with other clues in a way to improve the selectivity of algos that remove pages deemed by the algo as not high content value.

In this case
If a visitor lands on a content page with all content contained on a single page, scans for 10 seconds, finds the information they were looking for and clicks back to 'search deeper' on the subject, the bounce rate for the page is 100%, even though it contained the correct answer.

the majority of visitors do not click on a different link for the same search.

TheMadScientist




msg:4294310
 6:42 pm on Apr 8, 2011 (gmt 0)

Yes, tedster they're fine hairs, and just in case people don't get how hard I was laughing from my reply to Shaddows, I'm still chuckling a bit ... I'm sure there are some ways Bounce Rate is and will be useful, but as far as a ranking factor or generically saying, thinking, believing N% is better than N2% it's silly...

Even if it was a ranking factor (it's not, but, even if it was) there is no way for the site owner to have any clue whether 20% is better than 80% in a set of SERPs, because they don't know what the bounce rate for all the other pages in the set of SERPs and closely related SERPs is, so imo, people should really stop worrying about it, except for internal use, and to say your page should rank higher because it has a lower bounce rate or even actively trying to decrease it, through some silliness like splitting a page into 3 when all that does is cause more work for the searcher is basically a waste of time, imo.

Use bounce rate internally, but don't try to manipulate it or nuance it so your page ranks better, because you're probably doing about as much (or less) good as re-working your keyword meta tag...

Roaming Gnome




msg:4294399
 9:54 pm on Apr 8, 2011 (gmt 0)

I hope they dont factor the bounce too highly. Sites that deal in affiliate sales dont need low bounce rates to be successful or relevant. The surfer clicks what they are interested in, buys and leaves. Why penalize my site for being streamlined and easy to navigate?

tedster




msg:4294458
 12:21 am on Apr 9, 2011 (gmt 0)

Roaming Gnome, Google does not factor bounce rate into the rankings at all.

SanDiegoFreelance




msg:4294471
 1:31 am on Apr 9, 2011 (gmt 0)

Are we comparing apples to apples and oranges to oranges? A page bounce being when somebody hits the back button vs leaving the page or turning off their computer.

For google they now show an option to "Block all example.com results" when they back bounce. This is very different than when somebody goes to an authority hub, then to a retailer for to buy the product. Although most stats programs combine the number into the "bounce" count (They are very different).

I am in agreement with tedster that these bounce back to google do not change the position in the search engines. As well as rlange: " ... they would see me click on the result and not return. Even if I clicked on a second result, that can't necessarily be taken as a negative for the first result. ...". I do research the same way, I never rely on a single source on the internet for information; Hence norms need to be considered. Bounce data is noisy.

When I was running a vertical directory and measured CTR and Bounce. People had all kinds of patterns - some would literally go through the whole list - and often sign up with a affiliate program or two. I did see some sites with low quality have high bounce rates - I removed them - some site closed and had high bounce rates - I removed them. I saw high quality sites with low CTR based on title and description (I adjusted the title and description to properly show their content and improve quality of the directory).

Bottom line is this the data is very noisy, and the hit miss ratio very low, but also very useful. It can not be used in any kind of raw form - but the spam would in the ad-norm group. I can see uses for it when it comes to improving algo filters; that is to say a site fails one test but other factors are normal treat it like a false position.

TheMadScientist




msg:4294478
 2:09 am on Apr 9, 2011 (gmt 0)

A page bounce being when somebody hits the back button vs leaving the page or turning off their computer.

They're both bounces, which is why some of us try so hard to use very specific terminology ... A click-back is a bounce, but it's not 'bounce-rate' (inclusively) it's a subset of bounces, or a type of bounce...

Global Options:
 top home search open messages active posts  
 

Home / Forums Index / Google / Google SEO News and Discussion
rss feed

All trademarks and copyrights held by respective owners. Member comments are owned by the poster.
Home ¦ Free Tools ¦ Terms of Service ¦ Privacy Policy ¦ Report Problem ¦ About ¦ Library ¦ Newsletter
WebmasterWorld is a Developer Shed Community owned by Jim Boykin.
© Webmaster World 1996-2014 all rights reserved