Written by Nick Swan. Published on 13, August 2020
This guide takes you through the steps on how to create an SEO testing using SEOTesting.com, and how to interpret the results once they start to flow in.
If you are concerned about clicks and rankings to a page, it makes sense to run an SEO test whenever you make a change to an individual page or your entire site.
You compare the SEO numbers from before the change was made, to the SEO numbers after the change.
By getting the numbers from Google Search Console, SEOTesting.com automatically collects this data for you, allowing you to easily see whether the changes have had a positive or negative impact on rankings, clicks, click through rate etc...
Before running an SEO test, it is good to bear these best practices in mind:
1, Go ahead and make the change to your page, and resubmit the page for Google to recrawl through Search Console.
2, After selecting the site in SEOTesting.com you want to create a test against, click the SEO Tests link on the left hand menu.
On the SEO tests screen you’ll be able to see any Live tests you currently have running as well as the tests that have completed.
From here click the ‘Create a new SEO Test’ button.
You’ll now be on the Create Test screen.
3, Give the test a meaningful name and description of what the hypothesis of the test is and what on the page or for the query has been changed. Recording this within the test will be hugely useful when you are reviewing test results or looking back on the test in 3 months.
4, You can setup 3 types of tests based on page, query or the entire site:
5, If you want your SEO test to be based on data from a specific country, you can add the 3 character country code.
6, You can select the date the change was made. Often, this will be today’s date. But it is also possible to set the change date in the past. This is useful in a couple of scenarios:
7, Finally you can select the test period to be 2,4 or 6 weeks. We always advise to run a test for as long as possible to get as much data as you can.
8, Upon clicking the Create button you’ll be asked to confirm the start and end dates for the control and test periods. Once you click ‘Save and Start Test’, in the background SEOTesting.com will start collecting and processing data.
The second way to create an SEO test is from one of the SEOTesting.com reports. Whenever you look at a report containing queries or urls there will be a little flast test tube icon next to the keyword or page.
Clicking on this flask icon will take you to the Create a new test form, but with the values filled in so the test is based on the respective query or url.
When the results of the test period start to come in you will get a set of traffic light indicators to show whether SEO key metrics clicks, average position, impressions, click through rates, and number of queries queries are better when comparing the test period to the control period (before the change was made).
Often if a change is good for SEO, a page may start ranking for a lot more keywords. This means the impressions the page gets in the search results rises dramatically, but the clicks do not rise at the same rate. This could lead to a red indicator next to Click Through Rate, but overall you would consider the changes made to be positive.
As a general note, if you are seeing more green than red in the indicators - dig into the graphs but the test is looking good.
You don’t need to run a test for the whole duration if the results are indicating a big drop in results since the change was made. In this scenario I would revert the change - and stop the test.
Here are some examples of the kinds of things you can look to change or improve on a page and then measure the results.
Improving your page titles and meta descriptions can really improve the click through rate you get from Googles search results. As CTR is considered one of the input signals for Google's machine learning algorithms - this is something worth working on. If you have something where the page titles are all a similar format - such as products on an e-commerce store, it would be a good idea to test a few page titles before rolling out a change to all your products.
Is a page low on quality content? Try improving a page by writing more in-depth content and including useful images. Once a page has been improved ask Google to re-crawl it and test the results.
Does your page have a poor Google Page Speed Score? Page Speed has been confirmed by Google as one of its core ranking factors. Why not try improving the speed on one of your pages and measuring whether it improves the page's average position. If so it would probably be worth improving your Page Speed scores across your site.
May be you are unsure of which keyword a page should be targeting? Try testing different variations and compare results. Different keywords may rank similarly, but have different click through rates - so it is always useful testing these things.
If you have spotted a new external link to a page on your site, creating an SEO test can help show whether the link has given a rankings boost.
This guide has given you a quick intro into how to setup an SEO test using SEOTesting.com. If you'd like to learn more about best practices, business cases, and a bigger list of ideas to test, please checkout The Complete Guide to SEO Testing: Best Practices, Experiment Ideas, Tool Walkthrough.