Benchmarking Organic CTR To Measure Your SEO Experiments
Welcome to the final part in our series on organic click-through-rate (CTR) optimisation, good to see you back!
In Part 1 we discussed the (albeit tenuous) relationship CTR has with SEO, and the importance of optimising around click-throughs as well as rankings to achieve search dominance, whilst Part 2 introduced a variety of experiments you can run to methodically test ways to improve organic click-through-rate.
Not read the other parts yet?
But now that you recognise the importance of including CTR optimisation as part of your SEO strategy, how do you know whether your changes will actually have an impact?
As with any experiment, the first step is to have reliable base rates to benchmark your changes against. From these you can measure positive or negative fluctuations in CTR to learn whether your efforts have been successful.
This concluding article on click-through-rate optimisation will cover various ways to benchmark your existing organic CTRs to measure whether the work you’re doing is making a difference as you begin implementing your master plan.
Click Through Rate Benchmarking
Before you dive head first into optimising your pages for clicks, it’s important to take a benchmark of your current performance.
Doing so will not only give you a standard rate to measure future improvements against, but will also provide an overview of what level of CTR you can expect to achieve based on your industry, the type of page being optimised (product, blog post, etc.), and the level of competition you’re up against in search.
There are already multiple methodologies for taking CTR benchmarks, each with their own strengths and weaknesses, so before introducing the method we like to use, let’s take a look at a couple.
Aggregate CTR Benchmarks
At the beginning of this series we introduced a brilliant site for looking at global CTR averages across millions of different search queries: the Advanced Web Ranking CTR History.
The tool aggregates data for searches across a variety of queries, including branded vs. non-branded search, long-tail queries up to 4 words and over, searches in which ads appear, and searches across specific verticals.
Although it’s a great data set to get a rough indication of how different ranking positions correlate with different CTRs based on the type of search conducted, looking at such a wide data set majorly lacks any form of granularity or specificity to your site.
For example, many ecommerce businesses will struggle with average click-through-rate benchmarks purely based on the nature of their industry. Sharing SERPs with the likes of Amazon or Ebay is likely to lead to lower CTRs than expected benchmarks, unless you have a niche product or an incredibly strong brand, even after optimisation work.
Furthermore there are now so many extra variables in search. With featured snippets, map packs, people also asked, and more influencing CTR to your pages, taking aggregate CTR benchmarks will never provide a true reflection of your actual CTRs, making it impossible to measure any improvement efforts.
Page Level CTR Benchmarking
To get a more specific look at your own site, you’ll need to use Google Search Console. In the Search Console performance report you can see side-by-side comparisons of your clicks, impressions, CTR and average position for any given page on your site (or your site as a whole) up to a maximum of 16 months in the past.
Page level benchmarking takes the data from the Pages tab, plots it on a graph, and adds a trend line to show a benchmark CTR across all pages of your site. Anything above the line is over-performing, whilst any page under the line is a potential target for optimisation.
At first glance, this can seem like an effective way to monitor for pages which need support, especially for large sites which might filter the benchmarks down to only look at product pages, or categories for example.
However, once you find a target page, take a look at the ‘Queries’ tab. This tab shows all the searches that have driven impressions and clicks to that particular page. For most sites, there will be potentially hundreds, if not thousands, of different queries which have driven impressions over a maximum 16 month period, and unfortunately, Search Console only shows 999 of these.
You’ll quickly see that each query has wildly different click-through-rates and positions, and some might not even be relevant searches but random phrases that have triggered impressions with zero clicks.
With the Page level report providing an average from every single one of these queries, your Page level Benchmarks will be massively skewed, and you’re still not going to get a granular view of your true CTR benchmarks.
Query Level CTR Benchmarking
Similarly to page level, query level benchmarking uses data available about your site through Google Search Console, but instead of looking at your pages as a whole, it uses the same method on query level click-through-rates.
After exporting your data, plotting a graph and trend line, and seeing which queries have a sub-optimal click-through-rate, you can begin to filter for good targets to work on. However once again we run into the same problems as page level benchmarking:
- You’re limited to 1000 search queries across your entire site
- You’ll be faced with huge volumes of irrelevant outlier queries skewing the data
- Similar queries will have variable positions, impressions and CTRs, and might even have different pages ranking!
Alongside the above problems, you must also understand that the figures displayed in Search Console are all averages.
Over a 16 month period your rankings will likely have changed, hence impressions and clicks would naturally change, but the data you’re seeing is an average of everything for each query whether it appeared in position 2 or position 15.
This makes it a bad benchmark to use as a basis for improvements. As you optimise for CTR using techniques described in the previous article, you may also experience fluctuations in your SERP positions i.e. making a title change that results in dropping one place.
However as you're testing CTR improvements, you need to also have a benchmark for CTR in that lower position, as well as your previous one, to understand whether the change you made had a positive or negative impact on clicks, not just the rank. Now it’s obvious why only having an average across all positions is fairly useless.
“So how in Google’s name am I supposed to get accurate benchmarks?” I hear you cry.
Well, let’s rewind this a little and go back to the basics…
Keyword Specific CTR Benchmarking
Keywords are the basis of any SEO campaign. They’re the high volume, high value searches that you want to be capturing, they’re the keywords you’re tracking rankings for, they’re the ones that drive leads to your business resulting in revenue and a great ROI from investing in organic search.
So why wouldn’t you focus your CTR improvements, and benchmarking methods, on those all important keywords which drive everything else you do in SEO?
Taking benchmarks against a specific keyword, whether it’s above or below aggregate, page level or query level benchmarks, will provide far more value as it focuses you on the real objective: driving more site traffic from the searches your target audience are making.
Keyword specific benchmarking also counteracts a lot of the other problems we’ve faced:
- There’s no outlier data from random searches
- It enables us to take measured benchmarks at different positions over the entire time frame
- It’s ultra-specific to what’s achievable in that particular search environment.
It’s pretty much the best method to calculate how your click-through-rate has been performing to date, enabling you to accurately measure the impact of any changes.
How To Take Keyword Specific Benchmarking
Using a combination of Google Search Console, Google Analytics and Excel / Google Sheets, it’s fairly easy to take your own benchmarks, but it can be a little long-winded if you plan on doing it for a lot of keywords at once. This used to be far easier when you could export historical data directly from Search Console. Let's hope Google add that back to the new Search Console experience soon...
Before we get started, head over to Google Analytics and make sure you have Search Console connected with Google Analytics. Then navigate to ‘Acquisition > Search Console > Queries’:
Before we can start analysing our CTR benchmark data, we’ll need to set this report environment up to show what we need. First, we’ll make sure we’re only viewing data for our chosen keyword. Open the ‘Advanced’ filters on the report and include results exactly matching your keyword:
Once we have just our chosen keyword showing up, we need to change the metrics displayed to compare CTR and average position:
Now we’ll add a few more filters to make sure we have the data we need. Start by changing the ‘Date’ filter to take the past 16 months worth of data (or as long as you have available):
Finally to ensure we only get results from our target country, set your secondary dimension to ‘Country’ then head back into the ‘Advanced’ filters and include results from your country:
With all of that done, let’s Export our data set. It’s up to you whether you want to use Excel or Google Sheets as they’re fairly similar, but I’ll be using Excel in the next few steps.
In your spreadsheet, you should have three columns: Day Index, CTR and Average Position. Start by removing the decimal places from Average Position, then order the entire table by Average Position from Smallest to Largest:
Next, use a formula to get the average CTR for every position group. Once you’ve done this for all position groups, you’ll be left with a table that shows the average CTR for for each Average Position:
You now have an average CTR for your chosen keyword according to which position your site appeared in within the past 16 months, which is far more than you had before.
Like I said, it’s a little long winded to do for lots of keywords at once, and it’s still not a perfect representation, especially if you’ve only got a few data points for a particular position compared to others, but it does provide a useful benchmark for measuring your CTR optimisation experiments, which brings us onto the final section...
Measuring Your CTR Changes
Using your newly minted CTR benchmark data, you should now begin implementing your CTR optimisation tests - if you already started them, make sure you only take benchmark data up to the date before you started testing to avoid muddling the results.
Without getting too in-depth with confidence levels and statistical significance for measuring our tests, you’ll need to leave your experiments to run for a while. How long will depend on the volume of your chosen keyword: a page ranking in the top 5 for a 20k searches per month keyword will need to run for a short time than one which receives only 200 searches per month. Ultimately I usually take a finger in the air method to this and over time you’ll get better at estimating when to stop your experiments.
Once you’ve run them for long enough that you’re confident you’ll have something to measure, repeat the benchmarking process to get your new average click-through-rate data set. Compare this new data against your benchmarks to see if your experiment resulted in a positive improvement to CTR.
Now just rinse and repeat, ad infinitum, as you experiment your way towards total click-through-rate domination!
Now, Off You Go
I hope you’ve found this series on organic click-through-rates and SEO useful.
You’re now prepared with the knowledge you need to start experimenting with CTRO techniques, and measuring your successes against past benchmarks, in order to squeeze the most you can out of every search your website appears in.
Like most areas of SEO, it’s not something that can be done overnight, and requires continuous commitment to testing to see success, but armed with knowledge from this series, you should have some awesome ideas on where to get started.
I’d love to read your comments on the series, and to know about what other topics you’d love for us to cover on the Evoluted Think Tank, so please leave a message below!