Jump to content

Cre8asiteforums Internet Marketing
and Conversion Web Design


Photo

Is A High Bounce Rate Inherently Bad?


  • Please log in to reply
49 replies to this topic

#1 Robert_Paulson

Robert_Paulson

    Gravity Master Member

  • Members
  • 220 posts

Posted 27 March 2016 - 04:38 PM

Over the history on the sites I've run, it seems that most of them have had a pretty high bounce rate.  80% or more.  And I've always read that that's bad, but I'm questioning the wisdom of the global assessment. 

 

Would it not depend on the type of site? 

 

For example, my main site over the last 13 years has had a bounce rate of 80-90%.  However, many of the pages visitors land on have an average stay of 3-8 minutes. 

 

My hypothesis is that some sites (content-driven, I'd think) have pages that adequately answer a visitor's question in that single page.  They (visitors) were never destined to register or visit frequently; they were always going to be a One-and-Done visitor.  And my site happened to give them the info they needed.  So they absorbed it and left. 

 

So I guess I'm wondering if the bounce rate is really a valuable metric for all site types.  Instead, maybe some sites can expect high bounce rates and the more important challenge is how to monetize those sites.  Because so much of the behavior on my sites follows the 80/20 rule, I'm thinking maybe I should serve 80% of my ads to the drive-by visitors.  ;)



#2 Black_Knight

Black_Knight

    Honored One Who Served Moderator Alumni

  • Hall Of Fame
  • 9416 posts

Posted 27 March 2016 - 06:12 PM

No.

 

If your site's function is as customer service - to answer user questions so that they don't take up valuable time talking to limited staff on limited telephone lines then having a site built to answer and settle those questions such that customers leave, sated, is the point.

 

Alternately, if your site were one of the classic old 'arbitrage' sites designed solely to take searchers from one search to land on your page and immediately exit through a sponsored link then that too would be successful by having people not browsing around.

 

But, if you are not one of those two very specific case types (and the vast majority of websites are not) then a bounce is not a good thing.  Not particularly because you can say you disappointed them, or failed to capture any value at all from their visit, but because you are completely in the dark.  You have no idea who they were, or what impression they took away.

 

So the first thing that you absolutely must do is attempt to capture more information.  You need to answer what value you gained from those 80% that bounce, because for all you know right now, those 80% were all disappointed and all they will remember in future is "Oh, I vaguely remember going to that site once, but it was rubbish".  They'll have forgotten that the site was good, but just not relevant to their needs of the time.  They'll only vaguely remember your brand and their disappointment.

 

Why is the rest of your site so unappealing that 80% of all visitors won't look at a second page? Are you forgetting to invite them to?  Are you not providing content that is relevant to what you do, so the article that brings them in doesn't naturally lead to any other part of your site?

 

You need answers, and that means you need to start devising and running experiments to give you some.



#3 Robert_Paulson

Robert_Paulson

    Gravity Master Member

  • Members
  • 220 posts

Posted 27 March 2016 - 07:37 PM

Why is the rest of your site so unappealing that 80% of all visitors won't look at a second page? Are you forgetting to invite them to?  Are you not providing content that is relevant to what you do, so the article that brings them in doesn't naturally lead to any other part of your site?

 

 

That's an excellent question (or series of related questions). I should state first that the site for my brick and mortar business has less of a bounce rate, typically is focused in its content, and visitors often move from portfolio page to portfolio page, telling me they are qualifying us as a prospective provider of services for their needs.  My question pertains more to my informational sites and a forum site. 

 

Now, on to your questions - is the content of the heavy landing pages relevant to the rest of the site?  I feel confident saying yes, very relevant.  Is the site so unappealing they don't peruse a second page?  I suppose, but the length of stay on the 100 or so top pages seems long enough that that isn't the issue.  I can't say with certainty, but I've always thought that exits in under 30 seconds mean a visitor left disappointed.  Conversely, visitors that stay on a page for up to 8 minutes on average must've found something they liked.  Either that or they're all deciding to go make a sandwich after they click on my links. 

 

Do I do an adequate job inviting them to dig deeper?  This I can't answer.  Which then tells me that the answer is probably no. 

 

If I'm doing an awful job inviting further reading, how do I test that? 

 

I've noticed Google is offering a similar content widget to allow you to place on a site to help visitors find more of the same kind of content (from your site) that page contains.  Maybe Google is trying to tell me something? :blink:



#4 iamlost

iamlost

    The Wind Master

  • Site Administrators
  • 5377 posts

Posted 27 March 2016 - 08:34 PM

No.

 

We've had a few discussions on bounce rate over the years.
A few that might be of interest:
* Bounce Rates For Social Media, November-2007.
* How Many Bounces Could A Bounce Rate Have?, January-2012.
* Reducing The Bounce Rate On A Mobile Site?, August-2015.



#5 Robert_Paulson

Robert_Paulson

    Gravity Master Member

  • Members
  • 220 posts

Posted 27 March 2016 - 09:58 PM

Thank you for assembling those links.  So far in my reading, I don't see discussion of time on page relative to bounce rate. 

 

Let me give a specific example - I have a page that was originally authored in about 2007 or 2008.  Since then it's been ranked #1 in Google for many direct and similar/related search terms.   Avg time spent on that page over those 7-8 years has always been 6-8 minutes.  The presidents of the largest mfgs of the product being discussed on that page know me by name because of that article.  So I feel confident in saying the article falls into the "authoritative" category.  If it's maintained that level of traffic, visitors reading for that amount of time on the page for the life of the article (and I won't even go into the hundreds of emails and calls I get about it, despite my asking in the article NOT to do those things), I think any argument that claims the article isn't relevant to the searches where it turns up or didn't serve the needs of the visitor because of a high bounce rate fails logic. 

 

I can spin a yarn, but not a yarn so good that on average I can dominate the serps and con people to stick around for 6-8 minutes, every single day for 7-8 years. 

 

So I'm wondering if in fact the article is TOO complete.  For that particular visitor's Google query that led them to that page, their needs were met and they didn't have to return.  Ammon's question is on my mind, though - have I done a crap job of showing visitors why they need to see the rest of my content? 

 

From you, in 2007: 

 

"The more a page is designed to provide an immediate answer, including immediate conversion, i.e. ad or affiliate link, the higher the bounce rate. The more a page is designed to 'pass on' traffic to (an)other page(s) for additional information or conversion opportunities the lower the bounce rate. Finally, something within our control! And knowing the difference between good bounce (conversion), so-so bounce (external link click, no conversion) and bad bounce (no external link click, no conversion) are important statistics."

 

I think that article, all nearly 3,000 words of it, provides most of the answers visitors may have on the topic.  It was never intended as an immediate converter or affiliate link.  Just solid content.  But maybe that was the mistake. 


Edited by Robert_Paulson, 27 March 2016 - 10:01 PM.


#6 EGOL

EGOL

    Professor

  • Hall Of Fame
  • 6317 posts

Posted 27 March 2016 - 10:00 PM

Some people have a really high bounce rate because they don't market their own content to the people who visit their site. 

 

What are you doing to market your own content?  Do you have related boxes?  Suggested additional readings?  Do you have hypertext in your articles to related content?   What are you doing here?

 

Keep in mind that bounce rates are measured in different ways by different analytics programs.  A frightening bounce rate from GA might be a much lower number in clicky.  They each have arbitrary definitions.

 

In my mind a bounce rate is crappy if you are not doing anything to help the visitor find more.  I don't care how big or small the number is.



#7 Robert_Paulson

Robert_Paulson

    Gravity Master Member

  • Members
  • 220 posts

Posted 27 March 2016 - 11:11 PM

Hi EGOL - Good to see you still kicking around these sites. 

 

What I try to do with these pages, to varying degrees of depth and success depending on initial success and my assessment of their future potential, is to anchor text link portions of the content to relevant threads in the forums of the site.  On the other side of that coin, I find relevant threads in the forums on that same topic and go into the OP and edit it to include a link back to that static page.  So in total there are probably 15 outgoing (to pages on my own site) highly relevant links on the page, and about 20 links from highly relevant threads linking back.  Some of those forum threads rank highly for related searches, so it seems like that page is in a pretty good place, search-wise.  I have quite a few other pages in the same situation, but this one brings a disproportionate amount of search traffic to the site, so I'd really like to make more of it and it's little brothers and sisters, either in terms of more pageviews or more on-page monetization. 



#8 EGOL

EGOL

    Professor

  • Hall Of Fame
  • 6317 posts

Posted 28 March 2016 - 05:46 AM

How about the design of the site.  Do you have a sidebar where you can promote related content, popular content, trending content.

 

You have probably seen the "more from the web" and "sponsored content" links that are distributed by outbrain, taboola and related sites.  They have image that are usually highly provocative and attention grabbing.   I tried some of those programs on my site and learned that they are clicked like mad - even when they are posted at the bottom of a page.  The problem that I found is that they - at the time that I tried them - paid very little.  So I was losing an awful lot of traffic for very little income.   So, I decided to have some widgets like them on my site that promote my own content - using the images that illustrate my articles.

 

If you have a forum without a sidebar, consider creating one.  Make it wide enough to display the 300x250, 336x280 and 300x600 ad units.  Then create a method to display your own content in there using attractive images for your most popular, newest, most active, or trending content.  If there are topics that repeatedly crop in in your discussions, writing basic articles about them to display in the sidebar might be a profitable use of your time.



#9 Black_Knight

Black_Knight

    Honored One Who Served Moderator Alumni

  • Hall Of Fame
  • 9416 posts

Posted 28 March 2016 - 07:08 AM

As a general rule, I tend to say that questions are often more valuable than answers.  That's because some questions may be rhetorical, others may have no single right answer and need revisiting often, and that even in considering simpler questions with more straight-forward answers, that consideration may often inspire new insights or ideas.  Questions open up our minds, where answers may sometimes blinker them.

 

 

 

is the content of the heavy landing pages relevant to the rest of the site?  I feel confident saying yes, very relevant.

 

Remember that relevancy has to be more than just being on topic.  Context and mind-set are just as important.  You may need to further test that the people who were asking the questions that your content seeks to answer match the mindset you'd imagined and predicted.  

 

For example, there was an informational site all about a particular subject matter that hoped to sell books on that subject.  The logic of its creator was that people who were interested in his articles on the topic would be ideal customers for the books on the topic.  The reality however was that people already knew where to buy books, and a large percentage of those going online for fast answers were not those really interested in the subject itself, just wanting a fast answer.  A lot of students doing homework hit the site, copy pasted sections, and left.  They had never had the interest in the topic, only in getting the answer for their homework as easily as possible without reading the textbooks they'd been given for free.  The content was superbly relevant to the topic, but not to the mindset of deliberately not turning to books.

 

There are as many ways to test this as you can dream up, and then more that only others would dream up.  Just bear in mind that the question itself, and how it opens your thinking up in asking it and seeking answers has value in itself.  Think, invent, posit theories.  Put in links that would appeal to other mindsets or motives and track clicks on those (using redirected links allows you to track 'hits' on the redirection script).  Ask yourself "If these didn't have the mindset and intent I'd expected, what might they be intending?  What might help those people and intents?"

 

 

 

 

Is the site so unappealing they don't peruse a second page?  I suppose, but the length of stay on the 100 or so top pages seems long enough that that isn't the issue.

 

So, first I want to know what method is used to calculate the time spent on the pages, because in most analytics software (before being heavily modified) a bounce is a one-hit deal, with no second activity to time the period between.  Thus time on page/site metrics automatically exclude bounces, and are based only on those giving a second page-view, and the time between requesting the first page, and requesting the next.  There are methods to use javascript to start a timer when a page is loaded, and to then actively 'ping' the analytics software at set intervals of say, 30 seconds, to report that the visitor is still there, but these tend to be custom installs (and most don't differentiate between the page being open in a background tab, over being the focal tab).

 

Obviously, if those times are based only on the 20% who don't bounce (which would be the norm of analytics) then those are highly questionable figures based on a small minority of your audience.  Further, such a small sample is far more easily skewed by an even smaller proportion within that small sample, such that it could be that just a few people are leaving the tab open for very long periods while reading many articles, and the average of even the 20% minority without those is just a minute or two.

 

 

 

If I'm doing an awful job inviting further reading, how do I test that?

 

In part, by simply asking yourself the question and seeking your own answers.  This is the value of questions that I speak of in my first lines of this reply.

 

Obviously, by being aware of the possibility, you'll be more conscious of what "Calls to action" to read more pages, or just learn more about the author, or see a clearer picture of figure 1, or whatever.  And don't be afraid, for the purposes of both testing, and of providing the best experience, to even suggest reading elsewhere and other articles (using links you can track the clicks on).


Edited by Black_Knight, 28 March 2016 - 12:55 PM.


#10 Robert_Paulson

Robert_Paulson

    Gravity Master Member

  • Members
  • 220 posts

Posted 28 March 2016 - 08:48 AM

EGOL, I'm wondering if that Google widget I mentioned earlier might work.  I haven't tried it, but as I understand it, it's function is similar to those links at the bottoms of pages you mentioned, and yeah, they're SO DAMN CLICKABLE! Now I don't have a bevvy of bikini-clad women anywhere on my sites (well, maybe on one), so it may not be as fruitful, but I'm feeling like this might be something to at least try.



#11 EGOL

EGOL

    Professor

  • Hall Of Fame
  • 6317 posts

Posted 28 March 2016 - 09:01 AM

I tried Google's widget when it was first released.  I didn't like it because the images were cropped automatically and poorly.  The captions were grabbed from my content and cropped as badly as the images.   It didn't get clicked much.   It didn't look nice.  Maybe they have improved it since them.

 

So, I am using my own images, my own captions and am much happier with it.   Donna made a system for me that ranks and reranks the images on a schedule using simple algos that I control.   I like the results.



#12 Robert_Paulson

Robert_Paulson

    Gravity Master Member

  • Members
  • 220 posts

Posted 28 March 2016 - 09:03 AM

Black_Knight, thank you for the thoughtful response.  Plain ole Google Analytics is where I'm drawing my data from. 

 

The book site analogy is relevant, because there are certainly some parallels.  Based on the phone calls and emails and forum threads that get started on the topic that all are a result of this particular page, the reason this page is visited is not because of an interest in the topic - it's because there's been a problem with the installation of the product this page discusses, and people are searching for fixes to the problem. 

 

I like the thought train you put in motion here.  I'll have to spend some time on it to hone things.

 

I think part of my frustration is in talking with these other guys that run similar sites, even now, a decade later, they seem to know so little about their sites and how they work and what SEO is yet they keep having money fall in their lap. 



#13 Robert_Paulson

Robert_Paulson

    Gravity Master Member

  • Members
  • 220 posts

Posted 28 March 2016 - 09:06 AM

So, I am using my own images, my own captions and am much happier with it.   Donna made a system for me that ranks and reranks the images on a schedule using simple algos that I control.   I like the results.

Did she make it just for you, or is it something she's offering to a broader market?



#14 EGOL

EGOL

    Professor

  • Hall Of Fame
  • 6317 posts

Posted 28 March 2016 - 09:37 AM

She made it for me.  But she might be able to make something similar for other websites. 



#15 tam

tam

    Honored One Who Served Moderator Alumni

  • Hall Of Fame
  • 2334 posts

Posted 28 March 2016 - 10:09 AM

It depends what you want from your visitors. If you just want to educate them, e.g. you are the local govt. teaching them which recycle bin different rubbish goes in, then having them spend 10 minutes reading a page and then leave is great. If you want to convert those readers into shoppers, e.g. have them buy your amazing recycling bin system, and they never move on to a buying screen it's terrible.



#16 JEHochman

JEHochman

    Unlurked Energy

  • Members
  • 9 posts

Posted 28 March 2016 - 12:55 PM

I always tell people concerned about a high bounce rate to paginate their content and make the navigation more confusing. This will increase average pages per visit and lower bounce rate.

On the other hand, my own site has 90% bounce rate and does exactly what it is meant to do: answer the questions of those who can't afford to hire me, while intriguing those who can.

So, do you want to have an effective website, or beat an arbitrary metric? What matters is why bounce rate changes and whether it makes sense got your objectives.

#17 bobbb

bobbb

    Sonic Boom Member

  • Hall Of Fame
  • 3305 posts

Posted 28 March 2016 - 02:19 PM

I always tell people concerned about a high bounce rate to paginate their content and make the navigation more confusing. This will increase average pages per visit and lower bounce rate.

I see sites like this. I have the tendency of not going back realising what it is all about: ads and plenty of them before eyeballs.... and they are probably very effective at RPM and make a lot more $ than me. A 1200 word article in 3 clicks.

 

You get to be able to spot them from the snippets. Like: The 10 best ........ in the world. You know it is 10 clicks to the last item.

 

Just curious. At how many words would you tell them to paginate? Do they?



#18 EGOL

EGOL

    Professor

  • Hall Of Fame
  • 6317 posts

Posted 28 March 2016 - 02:40 PM

From my experience, you get better google rankings by keeping all of an article's content on a single page.  In addition, you get more long-tail traffic because of the keyword diversity.

 

So, folks gunning for ad impressions, might be cutting their own rankings and keyword reach.



#19 glyn

glyn

    Sonic Boom Member

  • Hall Of Fame
  • 3209 posts

Posted 28 March 2016 - 02:45 PM

While it might not be inheritly bad, Google may use that against you, but frankly I would not wonder about it. I would however think about offering an app so you can deliver that content without going through a search engine, and make it as useful as your content.

#20 bobbb

bobbb

    Sonic Boom Member

  • Hall Of Fame
  • 3305 posts

Posted 28 March 2016 - 03:01 PM

While it might not be inheritly bad, Google may use that against you,

I'm wondering how they would figure that out as opposed to a "related article" link in the content. We can tell by reading it and I'm sure Google does  not read the article.



#21 JEHochman

JEHochman

    Unlurked Energy

  • Members
  • 9 posts

Posted 28 March 2016 - 03:31 PM

 

 

 At how many words would you tell them to paginate? 

 

I personally dislike pagination.  My comment was a bit sardonic, but if the content was very, very long, I might paginate after something like 500 words, just to make it easier to deal with.  That's the issue: paginating to artificially drive up page views and ad revenue, versus paginating at a length that is convenient for the reader.  A very good option is to offer pagination or "view as one page".  Business Insider does that with their mobile app, and I appreciate it.



#22 bobbb

bobbb

    Sonic Boom Member

  • Hall Of Fame
  • 3305 posts

Posted 28 March 2016 - 07:03 PM

 

I personally dislike pagination.  My comment was a bit sardonic, but if the content was very, very long, I might paginate after something like 500 words, just to make it easier to deal with.  That's the issue: paginating to artificially drive up page views and ad revenue, versus paginating at a length that is convenient for the reader.  A very good option is to offer pagination or "view as one page".  Business Insider does that with their mobile app, and I appreciate it.

I had figured you were not pushing pagination, just curious about length. 500 words. That sounds like ask.com and the minefield they lay to get to the end of an article.



#23 Black_Knight

Black_Knight

    Honored One Who Served Moderator Alumni

  • Hall Of Fame
  • 9416 posts

Posted 29 March 2016 - 09:33 AM



Google may use that against you, but frankly I would not wonder about it.

 

I know you'll already know the various wrinkles of this, but wherever stuff comes up that is commonly and widely misquoted or misunderstood you know I like to be thorough with the caveats and conditional stuff. :)

 

Google use bounces back to SERPs as a quality control measure on their algorithms.  They have repeatedly said they do not use 'bounces' as they are reported in analytics (where they leave your site but not back to search results), nor even track the data for anything but your own analytics reporting.  Nothing in individual analytics gets fed back into the algos.  I believe them on this, if for no other reason than the sheer waste of processing it would take to try to figure every single online click into an algo, and for very questionable, vague, metrics that could have so many possible explanations as to be worse than useless (they add confusion more than add clarity).

 

Even where Google use the bounce-back-to-SERP (aka Pogosticking) as a metric of quality control, this is not run on a query by query basis.  That again would be just too intensive to get anything meaningful on.  This is a metric to judge the improvement or otherwise across a significant proportion of Google's quality, looking at literally millions of diverse searches.  The closest would be when they A/B test some feature that works in a particular scenario (such as the "Your money or your life" searches, where queries about finance and health should involve additional attention to factual accuracy and proven authority).

 

Bounce back to SERP is used solely to measure how many results a user needs before being satisfied, how far down they will have to go before selecting a result, and of course, the positional bias effect is taken into account.  It helps Google check that, on average across millions of searches made, the algorithm that put the best sites in the right places were tested.



#24 Robert_Paulson

Robert_Paulson

    Gravity Master Member

  • Members
  • 220 posts

Posted 29 March 2016 - 04:26 PM

Lots to read here (thank you for the insights!) but a couple things caught my eye I wanted to respond to -

 

Because the site has dominated the terms it seeks to rank for for about 8 consecutive years, and it just survived a 301 to a fresher look and an update to the content, I feel confident saying Google isn't penalizing me.  I also know from conversations with companies (of sizes ranging from about $10M - $100M) whose whole livelihood is made on selling the product this page discusses, it's killing them that none of them seem able to outrank this page.  I don't say this to brag, because it's one page covering one product.  Big whup.  I simply say it to express that it appears Google is pretty satisfied that the page does what Google hopes it would, and despite companies with a lot of skin in the game trying to outrank it, they can't.  No page's rank is guaranteed I know, but this page for whatever reason seems pretty fortified. 

 

I'm always riding the fence on how to present content.  For my warm, fuzzy 20% that create 80% of the forum content, I want to treat them well.  Minimal ads, fast page loads, etc.  For the one-and-dones, I guess I'm often feeling like I'd just as soon paginate by paragraph if I could, to force pageviews and ad impressions, make the nav convoluted page to page, so I can squeeze every drop of value out of them while they spend their few minutes on the site.  I'm kidding of course, but only a little. 

 

I always wonder why some sites provide nifty calculators (volume, area, step rise/run, that kind of thing), because in the history of my life, all I've ever done is Googled the calculator I need, gone there, used it, and never returned.  I almost expect to be punished for my behavior by having to suffer through a truckload of ads, but often that doesn't happen. 

 

There's lots more to consider in your responses, and thanks again for taking the time.  I'm going to revisit all this when I have a few free minutes later.  Thank you!



#25 bobbb

bobbb

    Sonic Boom Member

  • Hall Of Fame
  • 3305 posts

Posted 29 March 2016 - 05:01 PM

 

I always wonder why some sites provide nifty calculators (volume, area, step rise/run, that kind of thing), because in the history of my life, all I've ever done is Googled the calculator I need, gone there, used it, and never returned. 

Because at one time Google and Bing did not have that online and sent you to sites that did so people made those calculators to take advantage of the empty niche. Like you said you did a Google search and went there. Same for time sites.

 

Today all that conversion stuff is native in windows 10

 

I have an online network calculator site for IP stuff and I don't bombard people with ads. Just one and it's obvious it's an ad. That's all it does. 2 English calcs and 2 French plus stats on usage. Yes I know monitize, put up more content, etc, etc.

 

#2 on Y, B, and the Duck but 3-4 on G

Tried the others. Mine is simpler. But I'm biased


Edited by bobbb, 29 March 2016 - 05:07 PM.


#26 Robert_Paulson

Robert_Paulson

    Gravity Master Member

  • Members
  • 220 posts

Posted 29 March 2016 - 06:47 PM

I'm curious then - what's the bounce rate on that?  If the average Bob behaves like I do for those kinds of calculators, bounce rate has to be really high and not much monetization.  I think everyone should pay if they receive a benefit, and for those that make calculators, especially for mobile when it can be so easy to accidentally click on an ad if you don't have dainty fingers, you should require shmucks like me to have to navigate a maze of ads to get to the thing I want, because I'm just gonna use it like (insert something colorful here) and throw it away when I'm done, anyway. 

 

In the end, noble poverty is still poverty, right?  I'd prefer to don PJs made of money like EGOL does.  B:)



#27 bobbb

bobbb

    Sonic Boom Member

  • Hall Of Fame
  • 3305 posts

Posted 29 March 2016 - 07:27 PM

Have no analytics on this. Would not know the bounce rate but it would have to be high since that is all it does. There is nowhere else to go except out.

 

What I do know is the usage by country reads like the United Nations and known domain names as in names you can identify ( governments, banks, universities, etc ) read like a who's who.

 

I could, I guess, very easily increase the impression rate and not  go through a maze of ads.



#28 Robert_Paulson

Robert_Paulson

    Gravity Master Member

  • Members
  • 220 posts

Posted 29 March 2016 - 07:50 PM

I'd love for you to make a killing on that.

 

(And then tell me how you did it, so I can replicate your success.  ;) )



#29 bobbb

bobbb

    Sonic Boom Member

  • Hall Of Fame
  • 3305 posts

Posted 29 March 2016 - 09:07 PM

Don't get the wrong idea. I have no tricks in my bag.

 

When I built the calculator I wanted it to be as fast as I could make it. When you compute a CIDR address by submitting your data I use a POST to the same page but only spit out the response in an iframe usually ~ 200 bytes. Soooo if you hit the submit button 50 times I never refresh the page thus only 1 ad impression for as long as you stay on that page. I guess that could be 50.

 

Hmmm... now you have me wondering.



#30 Robert_Paulson

Robert_Paulson

    Gravity Master Member

  • Members
  • 220 posts

Posted 30 March 2016 - 09:45 AM

It could be 50, and in the spirit of making your tenants pay rent, the results could be presented somewhere else on the page.  Lower, probably.  ;)



#31 glyn

glyn

    Sonic Boom Member

  • Hall Of Fame
  • 3209 posts

Posted 31 March 2016 - 08:34 AM

 

I know you'll already know the various wrinkles of this, but wherever stuff comes up that is commonly and widely misquoted or misunderstood you know I like to be thorough with the caveats and conditional stuff. :)

 

Google use bounces back to SERPs as a quality control measure on their algorithms.  They have repeatedly said they do not use 'bounces' as they are reported in analytics (where they leave your site but not back to search results), nor even track the data for anything but your own analytics reporting.  Nothing in individual analytics gets fed back into the algos.  I believe them on this, if for no other reason than the sheer waste of processing it would take to try to figure every single online click into an algo, and for very questionable, vague, metrics that could have so many possible explanations as to be worse than useless (they add confusion more than add clarity).

 

Even where Google use the bounce-back-to-SERP (aka Pogosticking) as a metric of quality control, this is not run on a query by query basis.  That again would be just too intensive to get anything meaningful on.  This is a metric to judge the improvement or otherwise across a significant proportion of Google's quality, looking at literally millions of diverse searches.  The closest would be when they A/B test some feature that works in a particular scenario (such as the "Your money or your life" searches, where queries about finance and health should involve additional attention to factual accuracy and proven authority).

 

Bounce back to SERP is used solely to measure how many results a user needs before being satisfied, how far down they will have to go before selecting a result, and of course, the positional bias effect is taken into account.  It helps Google check that, on average across millions of searches made, the algorithm that put the best sites in the right places were tested.

 

Yes Mr BK, I am glad you clarified this point, also because I don't believe what comes out of PR-HQ too much and prefer to run tests myself, like many.

 

In them I found that query based improvements happened, rather than the side-wide lift on the results which might be a knock-on if a website had increased stickiness for queries. Some kind of query based ranking happens, also because it can be gamed and has been gamed fro quite some time already. There are so lots of factors at play but I've been able to isolate cases so am pretty happy that it happens.

 

Is it something of longevity no...but neither is stuffing your brand name alongside keyword terms through proxies, but I reckon lots of people are doing it for short term gain.

 

G.



#32 Black_Knight

Black_Knight

    Honored One Who Served Moderator Alumni

  • Hall Of Fame
  • 9416 posts

Posted 31 March 2016 - 10:18 AM

We talked briefly on this during the Q&A session, because Rand too has run a series of well publicised tests.  Some of those tests worked and some didn't.  That immediately tells us that there isn't some system-wide click metric in play, but something that is a little more complex.  Additionally, the effects were usually rapid, and temporary.  My best guess at this time is that a combination of things are in play.  For a start, I believe the rapidity of reaction and decline are showing us this is possibly a measure for adapting to 'Burstiness', and secondly, I think only certain types of query are easily gamed in Burstiness, just like certain types of query are categorized to need QDF (Query Demands Freshness).

 

Now, Burstiness is more often associated with actual data bursts, and is a bit like the old electricity burstiness effect that might happen during the ad-break of very popular TV shows as a million viewers all took the same moment to turn on an electric kettle.  But search queries also have a lot of burstiness.  Just like the way (and reasons) that a particular hashtag can start trending on Twitter.  When something goes viral, lots of people can suddenly be searching for pretty much exactly the same thing, and that thing may well be one specific article, video, tweet, or whatever.  An actor relatively unknown gives an amazing performance in a popular TV show, suddenly millions of viewers may search for that actor's IMDB page to find other things they have done.  Plus, of course, people searching for actual hashtags as they trend.

 

This creates a similar effect, and indeed uses a CTR metric too, if my guesses are correct, but explains the reason for the very short, temporary duration of the lift (which is usually just days at best, and sometimes more like mere hours).  If it was a CTR quality control part of the algorithms, there would be a more notable delay in effect, as you'd have to outweigh the entire past click-history of the term and results within, and there would be no reason for it to 'unlearn' or forget what it had just had happen.  It would take far longer for usual traffic patterns to reassert the old positions.

 

This is, however, merely my own best guess and speculation based on what I have seen happen in so many tests, the background reading in papers and patents, and of course my 'feel' for Google's perspectives on signals and motives.



#33 Robert_Paulson

Robert_Paulson

    Gravity Master Member

  • Members
  • 220 posts

Posted 31 March 2016 - 01:30 PM

BK, you have a genuine skill at explaining the complex with simple analogies to make it all make sense. 



#34 Black_Knight

Black_Knight

    Honored One Who Served Moderator Alumni

  • Hall Of Fame
  • 9416 posts

Posted 31 March 2016 - 04:43 PM

I always believed Einstein's saying that if you can't explain something in simple terms, you don't really understand it yourself.



#35 Robert_Paulson

Robert_Paulson

    Gravity Master Member

  • Members
  • 220 posts

Posted 03 April 2016 - 01:29 PM

This one page I keep discussing brings in up to 8% of my site traffic, on a site that has somewhere in the neighborhood of 10,000 pages.  It has more importance than I'd like, but I've gotta make hay while the Search sun shines, right?

 

So, testing has commenced.  I've implemented Google's (dubious?) Matched Content widget, and also done a better job of highlighting linked forum content to invite more pageviews.  Will report results in a week or so.    Last week, bounce rate of this page was 90%, time on page was 6:00+.



#36 Robert_Paulson

Robert_Paulson

    Gravity Master Member

  • Members
  • 220 posts

Posted 05 April 2016 - 04:58 PM

Bounce rate has dropped about 10% and time on page has dropped about 90 seconds.  I'd like to think it's because they're clicking internal links. 



#37 tam

tam

    Honored One Who Served Moderator Alumni

  • Hall Of Fame
  • 2334 posts

Posted 05 April 2016 - 07:30 PM

Bounce rate has dropped about 10% and time on page has dropped about 90 seconds.  I'd like to think it's because they're clicking internal links. 

 

Should be able to segment your audience (option at top of the report pages) and found out - look at time on page for people who bounce and those who don't before and after the change.



#38 Robert_Paulson

Robert_Paulson

    Gravity Master Member

  • Members
  • 220 posts

Posted 05 April 2016 - 09:35 PM

Thanks Tam - that sounds like a lovely idea but my brain hurts a little right now trying to figure out exactly how to sort that out.  So much information but feels like a house of mirrors.  Ack!



#39 tam

tam

    Honored One Who Served Moderator Alumni

  • Hall Of Fame
  • 2334 posts

Posted 06 April 2016 - 07:44 AM

Thanks Tam - that sounds like a lovely idea but my brain hurts a little right now trying to figure out exactly how to sort that out.  So much information but feels like a house of mirrors.  Ack!

 

Aww, google actually makes this one pretty easy. If you load analytics, there is a blue circle below the heading with a label that says 'all users'. Click that and a drop down will appear with various options - tick 'bounced sessions' and then apply. Now you'll be able to compare analytics results for people that bounce and everyone.

 

It basically splits up results by different types of user so you can see specifically what people that bounce do or people that use mobiles etc.



#40 Robert_Paulson

Robert_Paulson

    Gravity Master Member

  • Members
  • 220 posts

Posted 06 April 2016 - 12:28 PM

 

Aww, google actually makes this one pretty easy. If you load analytics, there is a blue circle below the heading with a label that says 'all users'. Click that and a drop down will appear with various options - tick 'bounced sessions' and then apply. Now you'll be able to compare analytics results for people that bounce and everyone.

 

It basically splits up results by different types of user so you can see specifically what people that bounce do or people that use mobiles etc.

Tam!  I think I'm in love.  :wub:

 

Worked beautifully, and yes, it confirmed a lowered bounce rate on that page.  :beatingheart:


Edited by Robert_Paulson, 06 April 2016 - 12:29 PM.




RSS Feed

0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users