CTR Optimization on Search Engine Results – Part 1

Browse By

Within the SEO space, almost every topic or niche is beaten to death – backlinking, keyword targeting, on page optimization, local search, citation building and social media/signals.

There is one area which really isn’t though, and that is CTR optimization on search results.

The reason for that, its an immensely difficult thing to measure and thus its very hard to infer a change that can be tied directly to updating the meta tags.

This makes A/B testing virtually impossible.

Here are the key challenges why this is so.

Lets say you have a set of meta tags/titles – we will call this ‘Instance 1’ which you are looking to optimize, i.e: get more people clicking on it and thereby improving your traffic without building a single new link or even optimizing anything on the page.

Very tempting, no?

The possibility that you can write a more creative and enticing set of tags which users are more likely to click.

This post will outline the steps necessary to get the data which you will need BEFORE you can do any actual optimization.

The only CTR data available is through the Google Webmaster > Search Index panel.

Challenge #1: In ability to get Keyword + Landing Search Report

The first challenge you will run into is that in the search console you will only get info based on Keyword OR the landing page, whereas realistically you need both, and separately they are quite useless.

    If you select the page based view, you will get the overall stats for the page:
  1. Total Impressions
  2. Total Clicks
  3. CTR
  4. Avg Position

The problem with this report is that it’s extremely misleading and actually harmful.

Firstly, for any given page – you will have dozens, hundreds or even thousands of KW’s and generalizing them into a single summary is basically averaging everything.

Your CTR on KW’s that are at the bottom of page 2 will be crap, even if its the best instance of meta tags ever written.

Your statistic is thus skewed, especially given the long nature of KW distribution.

If you go with the KWs view, you don’t see the actual page which ranks for that KW – so you actually don’t even know WHAT to optimize, even if you know that you do.

So the first challenge of anyone embarked on the path of CTR optimization is creating what I would call a Keyword + Landing Page report which would contain:

  1. Landing Page
  2. KW
  3. Impressions
  4. Clicks
  5. CTR
  6. Avg Position

You should see something like this:

1

This is a major first step, but unfortunately google don’t give you this data easily.

How to get it?

Hire a VA who would download the KW report and then manually create a page column and check which pages rank for the given KW. If your site has less than a 1000 pages, this should be a rather simple task.

OR
Write a script to pull API Data from Google

(Note: our SEO service will provide this report with one click after WMT access is authorized).

Challenge #2: Inability to ‘Weigh’ CTR versus Position

After you have creating the keyword + landing page report, you will need to really dig in to understand the inherent inaccuracy with the data.

For example, if you have a kw/page combo that ranks #1 – the CTR should be extremely high, I would say 20%/30% percent, (see statistics below).

‘Impressions’ as defined by google:

You get one impression for each time a URL from your site shows up in the SERP with the caveat that sitelinks are not counted as separate impressions. ~ 

So let’s say someone runs a search, you are result #10 on the bottom of the page, you get an impression, but realistically its quite likely that the user never even saw your result as its below the fold.

You will get 1 extra impression and no click, thereby reducing your CTR rate, even if you have the most enticing meta instance ever written.

Moral of the story: your CTR has to be weighed against the position.

Although there are no official publications from Google (shame), there HAS been some research done.

The first is the Study is done by Advanced Web Ranking and can be found here.

2

The Second one is done by AnalyticsSEO and can be found here.

3

The third one is done by Relevance and can be found here.

4

The fourth one is excellent and done by Catalyst Marketing and can be found here.

4-1

After taking into account these 4 case studies, we would like to great a summary table.

5

Note: there are many factors and segments that affect CTR such as branded vs. non branded queries, navigational nature searches, coupon based searches etc.

For the purposes of this post we will try and focus on ‘unbranded’ searches, since those are the ones that often require the real ‘fight’ for in SERPs and the CTR usually comes easier since search intent is very clear.

For the purposes of this post, we will perhaps over generalize but we will not factor in the relevance study, or the catalyst study, for the reason that they are much older and the data set is significantly different from the more recent studies.

We will match the AWR International Findings and the Analytics SEO to find the most median, ball park curve.

6

Another to note is that Analytics SEO has segmented their data into Branded and non-branded KW’s – assumingly AWR has not and vice-versa, AWR segmented into Int’l and US whereas we assume Analytics SEO did not.

Thus averaging the 2 based on common segments (Desktop & Mobile) should provide a ball park average of Global Searches (Int’l & US) and Branded & Non-Branded KW’s.

Obviously all this is very approximate, but without existing global standards for such research this is the best data we can work with.

And this is how it looks all together on a nice neat graph.
The thick red line is basically a very approximate, ball-park, CTR curve.

7

So actually there is a very clear curve which provides some useful insight into the CTR’s across US and International SERP’s aswell as mobile and desktop device.

Despite fluctuations, I find its still sufficiently useful to bench mark against.

Thus when looking at an CTR figure, you must look into what the ‘norm’ is (curve above) and how your CTR compares.

This would give you an idea.

(Note: our coming soon service would have the feature of weighing CTRs against norm).

Challenge #3: Inability to A/B Test

Fine, so you have figured out what KW’s are being clicked, which are not.

You have made some judgements on intent and have isolated pages which require an improvements in CTR.

You even have an idea why those particular query result viewing users are not clicking your meta instance, and you have crafted the most creative and juicy dripping new set of meta tags.

Now what?

You update the tags, but how do you know, has anything improved?

#1: Monitoring Traffic

That is no good, there could be a natural influx of search volume and within the same SERP and even worsened CTR your traffic could increase.

Won’t work.

#2: Monitoring SERP

Well, your page could have increased position DUE to the change in meta tags, but it could have also been due to a penalty on a competitor, or a backlink which just got indexed, or some on page change that affects the metrics.

Infact if your CTR stays the same in absolute sense as your position increases, your true, i.e: weighted CTR should be getting worse (since higher positioned demands higher natural CTR).

The Blue Print of True A/B Testing

So the ONLY way to really and truly measure whether a change in meta tags has made a positive impacts on your traffic is to have a record of:

Instance #1:

  • Meta Title = “Example Page”
  • Meta Description = “Example of description”.
  • From Date
  • Until Date
  • Weighed CTR across time span

And compare it to Instance #2:

  • Meta Title = “Example Page”
  • Meta Description = “Example of description”.
  • From Date
  • Until Date
  • Weighed CTR across time span

If the ‘weighed’ CTR increases in instance #2 versus instance #1 – then your A/B test wins.

The tricky part is that the DATA, MUST be pulled from the same time frame.

An important thing to note however is that simply looking at the summary of analytics on a page level isn’t enough.

More than likely, some keywords will rise, some will fall, thus selecting a balance of Keywords with the right ‘intent’ and a healthy level of searches is crucial BEFORE optimizing your tags.

You cannot optimize for every KW out there, some will rise, some will fall, so choose carefully.

Conclusion

Setting yourself up with the necessary set of data to be in a usable form and then measuring the effectiveness of your CTR optimization campaign is really tricky.

But do-able.

If you have read this far, you are probably interested in CTR optimization, and you will definitely be interested in the tool which we will publish by the 1st of november. On iseeogranic.com/service you can sign up to the mailing list and we will notify you once its launched!

Leave a Reply

Your email address will not be published. Required fields are marked *