Google SERP positions and CTR

[Doodled]

This is all a bit inspired by an article here:
http://forums.seochat.com/search-eng...rs-383180.html

and it has got me questioning if there is actually much difference these days between Google#1 and Google#5 when it comes to CTR (I think the data in the article says there isn't - even though the article argues that there is). I'm thinking of carrying out some larger scale research into it but don't want to re-invent the wheel.

So the most recent research I have found is here:
Comparison of Google clickthrough rates by position - Smart Insights Digital Marketing Advice

Which is pretty interesting in itself - especially how CTR for Google#1 varies so widely by industry. However I have a suspicion that they may have misused the data to arrive at their 'averages'.

But before I go and do anything does anyone know of any other recent research into SERP position and CTR? With Google changing the way it presents SERPS any research has to be recent to be relevant...

Cheers!

[realityhack]

It is a very interesting question. With a lot of complicating factors.
In the old days you got the 10 blue. And maybe some side ads. It was easy to track and see that #1 got a lot more clicks per impression. This is well documented. As is the fact that having a PPC ad would effect your natural CTR at that time.

Now you have lots of different 'types' of searches. Some searches show video results. Others do not. Same for images. And some throw up map results.
Some trigger ads with ad-extensions, others do not. Sometimes ads appear above the first position, sometimes not.
Add in that many people now use adblockers, and more and more people open the top Y sites in separate tabs or do a lot of quick click and back browsing that was not as prevalent before high speed connections were as common.

The result is that you can no longer say that a google results page looks like X

I know for certain the eye tracking on the different 'types' of results pages will be different (even if I don't know exactly what it is for each).
Given that I seriously doubt that the click throughs have the same distribution for the various result page 'types'. So you might see position 4 doing surprisingly well when it is right above image results vs. when it is not.
I don't think it is valid to talk about position vs. clickthrough anymore without a context of search type (informational, transactional, etc) and the type of SERP returned.

[Chedders]

There are so many variables to this and I know when I searched for the same question I got quite drastic results so I decided to log my own.
All I can offer is what I have found for my site and the keywords I track.

Position	% Traffic
1	22.10
2	15.80
3	5.27
4	3.46
5	3.05
6	3.40
7	4.88
8	2.85
9	2.92
10	2.60

The one interesting blip in the above is position 7.
I cant totally explain this but maybe my data is still too small. It has only been collected only over the last 3 month period.

[Test-ok]

The one interesting blip in the above is position 7.

maybe it because of the 7 listings on certain results. (last one on page 1)

[SeoKungFu]

You're either number one, or not so much in the business.

Come page two and you're out of business.

[Doodled]

Good feedback!

I'm beginning to think that research itself may be flawed because:
1) It is very industry specific
2) As realityhack points out it can also be specific to the type of search

That means even if you reach the conclusion that #1 gives a 50% CTR this is misleading because it is such a generalisation and your real individual CTR could vary dramatically from that figure.

Better sleep on it!

[geniusauthor]

Please I got this message from google about my blog. I even
noticed that I'm not getting traffic again from google but other sources.

Here is the message and I don't know how to go about solving it.
__________________________________________________ _____

SingBravo: Googlebot can't access your site

Over the last 24 hours, Googlebot encountered 1 errors while
attempting to connect to your site. Your site's overall connection
failure rate is 100.0%.

You can see more details about these errors in Webmaster Tools.

Recommended action

Verify that the web service software for your site is installed and
running properly.
Verify that your firewall or server are not blocking googlebot's
access to your site.
Verify that all scripts that run on your site have proper permissions to run.
Verify that your site has proper permissions to access the pages of your site.
Using Webmaster Tools, find a day with a high error rate and examine
the logs for your web server for that day. Look for errors in the logs
for that day and fix the causes of those errors.
Your site may be overloaded. Talk with your hosting provider about
either reconfiguring your web server or allocating more resources to
your web server.
If your site redirects to another hostname, another possible
explanation is that a URL on your site is redirecting to a hostname
that has one or more of the problems listed above.

After you think you've fixed the problem, use Fetch as Google to
verify that Googlebot can properly access your site.
_________________________________________________


Please do what you can do to help me on this.

[Dice79]

Anyone that suggests that the CTR for Position 1 is the same as Position 5 is either smoking something or looking at a key phrase that gets minimal traffic.

I've actually built my own rank tracking system and it compares actual traffic, estimated traffic using estimates similar to what Cheddars threw up and rank (On a weekly basis for 8,000 keyphrases). Traffic by rank is of course different for every industry and every keyword, but the obvious trend is always the same. The higher on page 1 the more the CTR.

[Doodled]

the obvious trend is always the same. The higher on page 1 the more the CTR.

This is the bit that bugs me in trying to model anything. To explain using a statistically invalid(!) example:
- There are two pages at #1. One page has a CTR of 10%, the other has 80%. So the average CTR is 45%
- There are two other pages at #2. One page has CTR of 40% and the other has 46%. So the average CTR is 43%

So being #1 has a CTR of 45% but the probability of any page having a CTR of 45% at #1 is very low.

Being #2 has a CTR of 43% but the probability of any page having a CTR of 43% at #2 is very high.

This is all to do with the spread of results making the data at each # either reliable or not. Personally I have plenty of examples where pages below #1 perform better for CTR because at over 50% they must by definition be getting more click throughs than the #1 ranking page.

That said SERPs are a shifting sand and some of these pages could be on their way to #1 (i.e. that is their longer term natural position) which again makes for more noise in any attempted analysis.

This aside the question I am actually asking in my head is perhaps this: How confident can you be (or what is the probability) that #1 will deliver a higher CTR than #2, #3, #4, #5. Where confidence is the statistical measure of the word.

The answer would be something like this : There is a X% chance that ranking at #1 results in a higher CTR than ranking at #2. There is a Y% chance that ranking at #1 results in a higher CTR than ranking at #3.

Now if X% or Y% is between 50 and 70 you can do anything you want with averages, in reality the reliability of those averages is too questionable to make any realistic prediction.

And it is very possible that X could be below 50 but only in certain search types (say product shopping) or certain industries.

[Chedders]

Size of data sample is key here and to exclude brand keywords from the results as they tend to result in a much higher CTR. I also tend to disregard the highest CTR% and the lowest to attempt to find the average. You cant take just 1 or 2 keywords to accurately predict so the more data you have the better. I wish I could be of more help here as I love playing around with stuff like this but I only have 1 site and 50 ish keywords to monitor so my sample is unique to my site and not large enough for anything more meaningful to anyone else.

I would guess my figures above is fairly average and similar to many sites out there but there are too many variables for anyone to say for certain the exact figures. So I treat it as a best guess approach.

[KernelPanic]

10 years ago a number 1 spot pretty much guaranteed you the click. It's different now. Surfers are more savvy and they look before they leap. I would rather be #3 with a compelling title and description snippet and a surfer-friendly URL than at #1 with none of the above.

Of course #1 with all of the above still brings the best results.

[tstolber]

I think the layout of the page and where the fold typically falls for many people and the scroll amount lends its self to explain position 7 having a bit of a blip. Position 7 area is often at the end of the sponsored search and other data thrown in on the right hand side, this somewhat frames position 7 as the end of the useful data and draws users attention to this area when scrolling down the page.