Does click-through rate affect your search engine positions?

[tamecrow]

Hi all,

Trying to get a bit of a discussion going over at Sphinn on the subject of click-through rates affecting organic search positions. It would be much appreciated if you could offer your own opinions and Sphinn the article if you find it interesting

Article at Sphinn: Does click-through rate affect your search engine positions?

[mjtaylor]

I absolutely believe that the clickthrough rate is taken into account on Google. I think that may account for, in part, the minor variations in SERPs from day to day.

I also believe that Google tracks a visitors behavior once on the site and takes into account whether you come back the the search page or follow a link from the site.

Cheers, MJ

[Webnauts]

One thing I am pretty sure about is, that click-throughs are a part of the "SiteLinks" algorithm.

[mjtaylor]

Originally Posted by Webnauts

One thing I am pretty sure about is, that click-throughs are a part of the "SiteLinks" algorithm.

And the SiteLinks algo is?

[Webnauts]

Originally Posted by mjtaylor

And the SiteLinks algo is?

Google Adds Sitelinks Control To Webmaster Tools & Much More

[wige]

There was an informal experiment done a while back see here that looked at the effect of clickthroughs on a listing in Google, and the effect of bounces recorded in Google Analytics that seems to indicate both these factors may play a small role in how a page shows up in Google's SERPs.

NOTE: Bear in mind that unless the user has the Google toolbar installed or the webmaster uses Google's Analytics, Google has no way to track user behavior beyond the landing page.

[ctabuk]

Wige said
'NOTE: Bear in mind that unless the user has the Google toolbar installed or the webmaster uses Google's Analytics, Google has no way to track user behavior beyond the landing page.

That should be made a sticky somewhere - rep points awarded - but it will probably say -spread around more before giving more etc etc

That should be impressed on the foreheads of so many posters

[crankydave]

This is a slippery slope for a lot of reasons.

Click throughs are easily emulated.

Time spent on a site does not indicate searcher found what they want.

Top 3 sites are going to get the majority of clicks.

Dave

[mjtaylor]

Originally Posted by crankydave

Click throughs are easily emulated.

Time spent on a site does not indicate searcher found what they want.

Top 3 sites are going to get the majority of clicks.

Dave

All true *and* I am still betting that search behavior is noted and calculated by Google ...

[crankydave]

Originally Posted by mjtaylor

All true *and* I am still betting that search behavior is noted and calculated by Google ...

I don't doubt that.

First thing they'd need to do is determine who (or what) is doing the searching.

Next they'd need to *guess* at the purpose for the search.

Without those 2 things happening first, I can't see where their *behavior* tells them much of anything especially as related to the ordering of the results.

"Armies" of searchers, alive or automated, clicking on results, following some and bouncing on others. Hmmmm... Click fraud round 2.

Dave

[mjtaylor]

Originally Posted by crankydave

I don't doubt that.

First thing they'd need to do is determine who (or what) is doing the searching.

Next they'd need to *guess* at the purpose for the search.

Without those 2 things happening first, I can't see where their *behavior* tells them much of anything especially as related to the ordering of the results.

"Armies" of searchers, alive or automated, clicking on results, following some and bouncing on others. Hmmmm... Click fraud round 2.

Dave

Well, they know something about the searchers who use gmail or the toolbar, as they have tracked past behavior and they do have the search terms to determine the purpose of the search.

Armies clicking would probably be detectable, and would not mimi normal search behavior, plus could create a tremendous bounce rate factor that would not help the site.

MJ

[espmartin]

While click-through rates do lead to daily swings in your SERPs, and it is a fact that
manipulation can and does occur (black hat style) in click ratios for many sites, I would like
to note the importance of website accessibility in regards to click-through - and how it
comes into play with SERPs.

The new feature Google has implemented in it's tools area, sitelinks, according to their own
documentation:

...If your site's structure doesn't allow our algorithms to find good sitelinks, or we don't think that the sitelinks are relevant to the user's query, we won't show them...

Note the first part of their statement. "If your site's structure doesn't allow..." My point is
that if you do not use semantically correct markup, which is the first part of accessifying a
website, not only will you have an UN-optimized UI (and cost you money. Have you been
following the Target.com lawsuit?), but your site will also suffer - even if it's the slightest
bit - with the ease of a SE (Google used here as an example) to crawl and gain all the
information it typically uses to "rank" you.

Hope I didn't deviate too much, but I feel that accessibility is a valuable issue in regards
to SEO. I think the worlds of semantic HTML (which is the foundation for accessibility in
my eyes) and SEO are closer related than most people think. Just think of all the click-through
you'll get with an accessible web design...

[crankydave]

Originally Posted by espmartin

While click-through rates do lead to daily swings in your SERPs,...

Speculation on your part or fact you can verify?

I rarely see "daily swings" in top 5 listings. These get the most clicks.

Dave

[espmartin]

I have seen my secondary (and third and forth) keyword phrases that I optimize for swing
from 127th to 133rd, to 138th, back to 127th - on a near day to day basis. (All organic SEO
on these terms).

I can see by this - in my case (YMMV) - is there is a difference in the SERP results for
those terms I target versus my top 10 results. Those keywords in the top 10 fluctuate
nowhere near as much as my secondary (and third and forth) keywords - but they have,
and again - in my case - moved on an almost weekly basis at some points.

Now, you can say that algos change frequently, that SE's results vary due to the change in
the weather, or that your top #5 sites never move (or move only monthly, and that very
minor.) because of a strong PPC marketing plan you have - whatever, but this is what I've
seen.

Now, let me theorize (again, YMMV) that all my efforts are VERY organic. I have not used
any PPC, or other paid marketing for my sites. So maybe that has had an effect to what I
have seen. So I guess you can call that speculation on my part?

I feel like I have to put my Politically Correct hat on with "in my case/YMMV"

[crankydave]

You're certain it's a case of your pages moving and not a case of the ones above you moving?

Pages above you moving into the SI, getting filtered, dropping out of the index, losing links, algo tweaks, filter tweaks, threshold changes, different DC's accessed, etc.

PPC plans having an effect on organic rankings?

It's not a matter of being "politically correct". You started your post out stating, very definitively, that click through rates lead to daily swings in the SERP's.

I questioned that statement of fact.

Dave

[espmartin]

You're certain it's a case of your pages moving and not a case of the ones above you moving?

Now that is something I clearly have under-estimated. Although, serious concentration of
accessifying my sites has - in my opinion - could be a major factor in given me a higher
ranking (as my statement about click-through; accessible navigation leads to more pages...accessed).

However, even that is open to speculation, as much as anything else SEO related. I mean,
was a "movement" in SERPs in fact due to making a site more accessible - or was it due to:

• The age of the domain (Google trusting one of my sites due to "getting older" by the day)?
• A certain article has been organically "growing links" - thereby giving quality backlinks?
• You name your favorite SEO myth or "factual" technique here...

As far as PPC marketing having effects on organic rankings, what I mean is if your PPC
marketing has been giving you steady incoming "targeted, relevant" visitors, this will
hopefully lead to:

• Return visitors
• Quality backlinks (your user linking back to your site with what they've found to be a quality site to them)

And so included in PPC goals is to not only sell your product to that user, but in
return readership. This generates (hopefully) a buzz, which leads to quality, relevant
linking = better off-page optimization = higher SERP.

So my statement about PPC and organic ranking is that I haven't used PPC marketing,
and so I base my results as such.

[mjtaylor]

Originally Posted by wige

There was an informal experiment done a while back see here that looked at the effect of clickthroughs on a listing in Google, and the effect of bounces recorded in Google Analytics that seems to indicate both these factors may play a small role in how a page shows up in Google's SERPs.

NOTE: Bear in mind that unless the user has the Google toolbar installed or the webmaster uses Google's Analytics, Google has no way to track user behavior beyond the landing page.

Or gmail.

[Webnauts]

Very interesting thread. Lets keep it going.

[Webnauts]

Originally Posted by mjtaylor

Or gmail.

Or Web History?

[espmartin]

I think I found that GoogleBot crawling my local hard drive!

Well, in a way that is true. You use the Google "internet" (referencing toolbar/analytics/web history/webmaster tools
- all that Google stuff) - and they have claim to all that "traffic" and it's analytic qualities.

Ok, back to reality...

[Webnauts]

Originally Posted by espmartin

I think I found that GoogleBot crawling my local hard drive!

The best SEO joke of the year! .lol:

[cass-hacks]

Originally Posted by Webnauts

The best SEO joke of the year! .lol:

It needn't be that big of a joke. Many years ago, back when servers cost a small fortune and the promise of giving over your first child, I used to serve web pages from a box at my home using dynIP, or whatever it was called.

I can honestly say that I had Google making visits to my home almost every day.

Back on topic though, I don't know for sure one way or another but considering all the data Google has access to, including user stats, it wouldn't surprise me if they use it for some part, even if very small.

On the other hand, they definitely use user stats for personalized search but how much that bleeds into other search types, I haven't a clue.

[Webnauts]

Originally Posted by cass-hacks

On the other hand, they definitely use user stats for personalized search but how much that bleeds into other search types, I haven't a clue.

That is what I was waiting to hear. Web History!!!

[espmartin]

Originally Posted by Webnauts

That is what I was waiting to hear. Web History!!!

And let's not forget "Google Desktop":

Multiple File Types

Google Desktop automatically indexes, and allows you to search the full text over, the following types of items on your computer:
Outlook Email Netscape Mail / Thunderbird Outlook Express Firefox / Mozilla / Netscape Word PDF Excel Music PowerPoint Images Internet Explorer Video MSN Instant Messenger Zip AOL Instant Messenger Text and others Google Talk Even more...

This is under user agreement of course, when you install it onto your system

[crankydave]

Desktop is interesting.

Recently had a conversation with an SEO/Webmaster that uninstalled it because it was constantly grinding away at his hard drive.

Dave

[kgun]

Originally Posted by espmartin

The new feature Google has implemented in it's tools area, sitelinks, according to their own
documentation: Note the first part of their statement. "If your site's structure doesn't allow..." My point is
that if you do not use semantically correct markup, which is the first part of accessifying a
website, not only will you have an UN-optimized UI (and cost you money. Have you been
following the Target.com lawsuit?), but your site will also suffer - even if it's the slightest
bit - with the ease of a SE (Google used here as an example) to crawl and gain all the
information it typically uses to "rank" you.

Hope I didn't deviate too much, but I feel that accessibility is a valuable issue in regards
to SEO. I think the worlds of semantic HTML (which is the foundation for accessibility in
my eyes) and SEO are closer related than most people think. Just think of all the click-through
you'll get with an accessible web design...

It is mainly about content (how it is presented is in most, but not all cases not so important).

<cite from my own forum>
XML is a large family of directly and indirectly related members like CSS and DOM. The XML Information set is an abstract description of the information contained in an XML document. It is the information contained in the document that is important. Applications are often not interested in the syntatctic details of a document, but only in the information contained in the document. XML informations set was designed to serve this purpose.
This information is used by other applications and parsers like DOM.
Related links:
XML Information Set
</cite from my own forum>

XML is a in some instances (not validation) a superset of HTML. Google should be able to extract the information from a web site. It should be more and more able to figure out whether that information is manipulated in one form or another.


I think Google is getting better and better. Yesterday I tried the following searches:

tex

latex

"scientific word"

All hits were very relevant. I agree, the examples are too specific and too few to generalize. The hits I have seen, for my KW searches, during the last months, have been very relevant IMO.

[kgun]

Originally Posted by espmartin

I can see by this - in my case (YMMV) - is there is a difference in the SERP results for
those terms I target versus my top 10 results. Those keywords in the top 10 fluctuate
nowhere near as much as my secondary (and third and forth) keywords - but they have,
and again - in my case - moved on an almost weekly basis at some points.

KW's

financial stability (little fluctuation in the top positions there).

object oriented programming

How Object-Oriented Programming Started

still on page 2. It should be on page 1 in my view.

[kgun]

Originally Posted by Webnauts

The best SEO joke of the year! .lol:

No, GoogleBOT has started crawling my brain.

And they watch me daily through Google earth with a resoultion of 1 cm (for internal purposes) when I take my coffee outside my house

[mjtaylor]

Originally Posted by Webnauts

Or Web History?

Web history *iff* someone has some sort of gmail, toolbar, Google desktop or other Google account and is signed in ...

But I have often wondered whether Google can see searcher behavior even without those.

MJ

[cass-hacks]

Originally Posted by mjtaylor

But I have often wondered whether Google can see searcher behavior even without those.

Every request comes from an IP address. Even if some are proxied or whatever, there is a lot of information that can be obtained from large random samples so even an IP address is not necessarily required.

When you consider the number of searches Google services in a day and all the data that generates, patterns within data sets of that size have a tendency to jump out and slap you in the face. For some thing, being able to tie some patterns to specific samples is useful/important but at the same time, there is quite a bit of information that can be extracted from anonymously sourced data.

[wige]

Google can only track what goes through their servers. The Google toolbar, and Analytics, as well as anything you click on while visiting any Google site. And of course, Adwords ads on any web site. Beyond that, Google can not track you.

Google Desktop has not been shown to send any data to Google's servers, and Google Sidebar and other apps have not either.

[mjtaylor]

Originally Posted by cass-hacks

Every request comes from an IP address. Even if some are proxied or whatever, there is a lot of information that can be obtained from large random samples so even an IP address is not necessarily required.

When you consider the number of searches Google services in a day and all the data that generates, patterns within data sets of that size have a tendency to jump out and slap you in the face. For some thing, being able to tie some patterns to specific samples is useful/important but at the same time, there is quite a bit of information that can be extracted from anonymously sourced data.

That's what I thought, though I couldn't have expressed it so clearly.

So, let me propose a hypothetical: a surfer (who doesn't have gmail, G's Desktop, toolbar or any other Google product) types a query into Google for the first time on a new machine. The surfer clicks through to a site, and bounces back to the query page and then clicks through to another site, but leaves that one via a link to a site in their resources page.

Do you believe Google can see that behavior?

And if so, might Google not register that the second site helped the user more than the first?

And if I understand you correctly, you are saying that even without following that IP or single user, Google can see bounce rates on a site, and can see when, for example, outbound links are clicked ...

This is the argument I use to encourage clients to link from their site to trusted relevant sites. I believe Google can see that the user doesn't return to search the same term again in Google or to click through to another site in the SERP, but instead finds a link that will perhaps take them closer to what they want.

On the other hand, if they don't find what they want and return to the SERPs, Google gets the message (sees the pattern) that the site didn't satisfy the user's query.

Thanks, MJ

[crankydave]

Originally Posted by wige

Google Desktop has not been shown to send any data to Google's servers, and Google Sidebar and other apps have not either.

I've not used desktop so please correct me if I'm mistaken.

Isn't there an advanced feature which allows access (indexing?) of files on multiple computers?

If so, where is the information stored? How is it transmitted?

Dave

[crankydave]

Just found this wige...

What is the Advanced Features option?

Enabling Advanced Features will allow Google Desktop to collect a limited amount of non-personal information from your computer, which is then sent to Google. This includes summary information, such as the number of searches you do, the time it takes for you to see your results, and application reports we'll use to improve the program.

Dave

[kgun]

Originally Posted by wige

Google Desktop has not been shown to send any data to Google's servers, and Google Sidebar and other apps have not either.

I think that Google toolbar and desktop are reverse engineered long time ago, so if there were any problems they would have been known months ago.

Originally Posted by crankydave

I've not used desktop so please correct me if I'm mistaken.

Isn't there an advanced feature which allows access (indexing?) of files on multiple computers?

If so, where is the information stored? How is it transmitted?

Dave

Never heard of that, but may be a network solution.

Non personal information. Is that so very different from your online surf behaviour / pattern?

[incrediblehelp]

Do you think they count clicks more:

1. When logged into Personal Search
2. Logged out of Personal Search:
3. Both equal?

[Webnauts]

Originally Posted by incrediblehelp

Do you think they count clicks more:

1. When logged into Personal Search
2. Logged out of Personal Search:
3. Both equal?

If I did not miss something, I believe that they are pretty equal. For example, they set cookies which are not deleted automatically when you log out of your personal search.

[incrediblehelp]

John I was speaking more towards the value of the click. My guess is personalized clicks may add more algorithmic value than a regular click.

[espmartin]

Originally Posted by incrediblehelp

John I was speaking more towards the value of the click. My guess is personalized clicks may add more algorithmic value than a regular click.

In as much as Google - with "personalized clicks" - can use that data in more coherent
ways. I mean, since with clicks that come from a known variable (Your Personalized
account) - they can use that data against other known variables. But if your clicks are
not "trackable" to any one account, they are therefore "random", and therefore in my
opinion - less valuable. Well, not so much "less valuable", but used in different algorithms.

[crankydave]

Originally Posted by espmartin

In as much as Google - with "personalized clicks" - can use that data in more coherent
ways. I mean, since with clicks that come from a known variable (Your Personalized
account) - they can use that data against other known variables. But if your clicks are
not "trackable" to any one account, they are therefore "random", and therefore in my
opinion - less valuable. Well, not so much "less valuable", but used in different algorithms.

I can see this as being just the opposite.

Afterall, the basis of PR at least in part, is the likelyhood that a completely random surfer will click on a link. "Random" being the key.

Known variables simply give them more reasons or means to exclude data.

Let's also not forget that click data a seriously skewed by the way they order results in the first place. They could easily take a relevant page from 200+ in the SERP's, move it to #1-#2 or #3 and the click data would change significantly.

My own personal opinion is that if it is used at all, it's done so to exclude pages algorithmically rather than increase rankings. Pages that are somehow being found "not relevant enough" for particular queries, ie MFA pages. Very much in the same way meta descriptions are not used to increase rankings within the search results pages but are used for other things.

Let's also not forget the role of what the description snippets play when it comes to click throughs and who controls what snippets get displayed and how click data *could* help them improve this.

I know this is way oversimplified but...

If they were looking at description snippets they were pulling from they page they could...

-Display a snippet with the text immediately following the query
-Display a snippet immediately preceding a query
-Display a snippet that both precedes and follows a query

If the click through rate changed drastically, they'd have an *indication* of what the random surfer found more relevant.

Dave

[mjtaylor]

Originally Posted by crankydave

Afterall, the basis of PR at least in part, is the likelyhood that a completely random surfer will click on a link. "Random" being the key.

Well, maybe. Adam Lasnik of Google has commented that Google's algorithm has come a long way since the PR patent and that there are now many other factors (such as searchers behavior, perhaps?) that go into the rankings.

I wouldn't hazard a guess as to whether a clickthrough is more or less valuable based on whether it's from a known searcher (toolbar user, for example) or otherwise - I think as cass-hacks suggested above, that Google can see the larger patterns and that's what matters.

[Duncan Pollock]

I suppose I'm showing a lot of nerve if I tend to disagree with some our major posters, but I'm reluctant to decide that click throughs and SERPS have much of an effect on one another. Admittedly, I'm thinking of a click through coming from an AdWord rather than simply originating with an organic appearance; and the lack of any/much relationship between the two of them has been reguarly denied/dismissed by Google themselves, never mind many of my WPW colleagues.
However, even is there any effect of one on the other, I think we need to decide that we're dealing with Apples (PPC and c/ts) on the one hand and Oranges (SERPS and c/ts) on the other. Either bunch of fruit may well be influenced by the number of "related" click throughs it generates, but pushing one upwards (or, I suppose, downwards) because so many (or so few) visitors are clicking on the "unrelated" Google response to a search query is surely no more than a coincidence.
Duncan

[mjtaylor]

Originally Posted by Duncan Pollock

I suppose I'm showing a lot of nerve if I tend to disagree with some our major posters, but I'm reluctant to decide that click throughs and SERPS have much of an effect on one another. Admittedly, I'm thinking of a click through coming from an AdWord rather than simply originating with an organic appearance; and the lack of any/much relationship between the two of them has been reguarly denied/dismissed by Google themselves, never mind many of my WPW colleagues.
However, even is there any effect of one on the other, I think we need to decide that we're dealing with Apples (PPC and c/ts) on the one hand and Oranges (SERPS and c/ts) on the other. Either bunch of fruit may well be influenced by the number of "related" click throughs it generates, but pushing one upwards (or, I suppose, downwards) because so many (or so few) visitors are clicking on the "unrelated" Google response to a search query is surely no more than a coincidence.
Duncan

Well, you are right that Google has repeatedly said that PPC clickthroughs have no weight on organic listings. And you can take them at their word, if you choose. I reserve judgment.

However, we are talking about the clicckthroughs on organic listings here. If the original poster was not clear when he used the term 'clickthrough' the post on the article he refers to is:
"There’s a theory in some SEO circles that the click-through rate of organic listings can contribute to the ranking of that listing."

Cheers, MJ

[espmartin]

There’s a theory in some SEO circles that the click-through rate of organic listings can contribute to the ranking of that listing.

One thing is almost agreed upon; that Google can and does keep track of click-through
on websites. And what would that type of data tell you? "User path" - where a user goes
on a site. More specifically, their "landing" and "exit" pages, and what pages specifically
in-between that got them there. Anyone who's familiar with Google Analytics can see how
this is important data.

If that is something we can agree to, I think it will take this discussion into the correct
"funnel", in my opinion, that the OP was shooting for.

If that is the case then (will SOMEONE agree with me please?), I theorize that Google will
use that data, and specifically which page is the most "landed" page - and use it among the
many other ranking algorithms they have accordingly. If you agree, then can you see the
importance of accessifying a website in relation to SEO?

[crankydave]

Originally Posted by espmartin

If that is the case then (will SOMEONE agree with me please?), I theorize that Google will
use that data, and specifically which page is the most "landed" page - and use it among the...

Where I will agree is that they utilize the click through data.

I disagree with the notion that it is a "positive" or "measured" metric that can increase the rankings within the results.

Dave

[Webnauts]

Originally Posted by incrediblehelp

John I was speaking more towards the value of the click. My guess is personalized clicks may add more algorithmic value than a regular click.

That makes big sense to me.

[espmartin]

Originally Posted by crankydave

Where I will agree is that they utilize the click through data.

I disagree with the notion that it is a "positive" or "measured" metric that can increase the rankings within the results.

Dave

So you think that the click-through data is utilized how, in some sort of "neutral" way that
does NOT effect a site's position, either in a positive (reads: higher SERP) or negative
(reads: lower SERP)?

[mjtaylor]

Originally Posted by espmartin

One thing is almost agreed upon; that Google can and does keep track of click-through
on websites. And what would that type of data tell you? "User path" - where a user goes
on a site. More specifically, their "landing" and "exit" pages, and what pages specifically
in-between that got them there. Anyone who's familiar with Google Analytics can see how
this is important data.

If that is something we can agree to, I think it will take this discussion into the correct
"funnel", in my opinion, that the OP was shooting for.

If that is the case then (will SOMEONE agree with me please?), I theorize that Google will
use that data, and specifically which page is the most "landed" page - and use it among the
many other ranking algorithms they have accordingly. If you agree, then can you see the
importance of accessifying a website in relation to SEO?

Makes a lot of sense to me ... lots.

Look: if Google is based on the concept of popularity by citation, it follows that Google would figure out how to algorithmically count the votes of actual web users. Why *wouldn't* it be logical to give the end user behavior a large vote, just as a link is a vote?

Think about it: if it were your search engine, wouldn't you? And isn't that a part of what personalization is about? Taking into account the user's surfing behavior. So, Google does it for gmail and other G users, you think they aren't using the same sort of information about general user behavior?

And if they aren't, why aren't they?

My answer: they would if they knew how ... and I am guessing they already do ...

[crankydave]

Originally Posted by espmartin

So you think that the click-through data is utilized how, in some sort of "neutral" way that
does NOT effect a site's position, either in a positive (reads: higher SERP) or negative
(reads: lower SERP)?

I posted a couple of possible uses in my post above the one you quoted Did you miss them?

Dave

[espmartin]

Originally Posted by crankydave

I posted a couple of possible uses in my post above the one you quoted Did you miss them?

Dave

Sorry, yes I missed your uses somehow, or selectively forgot them. Must have been the
cold coffee I was drinking (not an iced coffee, but just got cold )

Originally Posted by crankydave

...Known variables simply give them more reasons or means to exclude data...

My own personal opinion is that if it is used at all, it's done so to exclude pages algorithmically rather than increase rankings.

So your take has it that Google uses all the "user data" they take in (clicks and false clicks)
to negatively impact a site's "standing" with them? Maybe negatively is the wrong
description here...

But if "good" click-through data is found for a site (tons of traffic via successful
click-throughs from landing pages to sales pages - all originating from a Google search
page, for example), why wouldn't Big Google use that in an "increase a ranking" algorithm
for that site?