SEO tests are a critical component of any SEO strategy. After all, after you've implemented site changes in an attempt to boost search traffic, you need to have a proper understanding of whether or not those changes had a positive effect.
In this post, I want to show you different types of SEO tests you can run, and a simple technique that will introduce you to the concept.
For those who are unfamiliar with the concept, let's quickly cover the basics: what SEO testing is, and why you should do it.
What is SEO Testing?
SEO testing is the process of determining which website appearance choices led to better performance on the search engine results pages (SERPs). The results can then be used to determine if the respective changes should be implemented further, or undone.
When I say SEO testing, I'm referring to the "optional" areas within SEO where one configuration, design, or phrasing of the page may cause a web page to perform better in search results (that is, gives your web page higher rankings).
Testing can also tell you more about where to best spend time and other resources to grow your program further. This is not to be confused with SEO considerations while performing site tests such as A/B tests for SEO.
Why Test in SEO?
Here are why you would perform this particular type of an SEO Test:
- Justify projects – Often SEO projects are labor intensive for development or marketing teams. Consider a broad SEO project like "update blog posts” or "increase page speed." While both projects may be good for SEO and potentially for your search engine rankings, it’s hard to justify the internal cost at face value. With testing, you can break it up and "test" on some pages (like one blog category or one page speed improvement) to prove the success before rolling out to larger projects.
- Save time – With an SEO test, a configuration or tactic can get a proper evaluation to see if it's worth continuing or scaling the project site-wide, allowing resources to be spent on other projects.
- Buy-in – Testing helps get more done. A well-measured SEO test that shows a market opportunity, and growth is more likely to get internal support.
A Simplified SEO Testing Technique in 6 Steps
1. Identify something to test.
This may seem obvious, but you first have to decide what to test for! There are many ranking factors that search engines use in their algorithm to determine rankings, and may on-page SEO components that make a page unique. So, the subject of your test can be an adjustment to a page’s content, meta values, design, or a performance update to improve page speed, for example.
2. Group the landing pages receiving the variable.
This is ideally pages from a unique business category or page template, e.g. “hats” or “category pages.” The URL structure should help you group your pages respectively.
3. Find keywords associated with the landing pages.
Shoot for 2-5 keywords per page if possible. Tools like the Research Grid and Search Analytics are great for finding the top keywords associated with each page.
4. Implement the variable on the Landing Page in the test group.
5. Review KPIs.
Review key performance metrics such as “Average Rank” by the test Keyword Tag and site metrics like Traffic, Conversion Rate, Bounce Rate, and Time on Site by the test Page Tag.
Note how the test groups (keywords and pages) trend against the controls to find evidence of your SEO test taking effect. This step may be the most difficult in SEO, as the directions of these KPIs can be nuanced.
6. Decide if the test impacted SEO positioning.
If CVR or CTR improved without rankings it may be worth leaving the “test” element in place or continuing the strategy. Ideally the Average Rank of the test Keyword Tag also improved compared to the control. We can only speculate on correlation but these are the kinds of trends we want to see! Be sure to note any seasonality factors that could affect traffic or rankings before confirming impact too. For a tighter statistically approach, you can build on these metrics with these steps from Kissmetrics among other models.
Techniques for Testing SEO Impact
The following are common SEO optimization areas that make for good tests with specific techniques for how to approach.
It should help SEO to increase the quality and quantity of internal links to a target landing page. To see if this impacts your rankings and is a tactic worth optimizing, follow the steps above. (Not to be confused with backlinks, which are links from another site that point to yours.)
As noted in a previous post, testing title tags can be a highly effective test. Title tags have the dual benefit of improving rank and click-through rate if the listing jumps out to users.
Similar to title tags, meta descriptions can have a huge impact on CTR. Remember, after a Google search is made, users can see the metadata of a page as a part of a SERP listing. This can either draw in their click, or keep them scrolling.
Ideally, you want to capture attention and offer value, and SEO testing can help with this.
In fact, case study research shows the overall positive trend of CTR resulting from adjustments to the metadata over time. These changes wouldn't have been implemented if it weren't for tracking and testing the results.
You can run this test, too. seoClarity users can use Search Analytics, and non-users can run the analysis with Google Search Console data.
Update the content on a page to meet a searcher’s needs. A good test is to experiment short versus long formats of the content, or simply adding more relevant or semantically related keywords.
There are endless ways to enhance the experience of a website. From the variations in presenting content, to colors, and format. Try mobile improvements and page speed if it’s possible for only a portion of pages to receive the enhancement.
But remember, the user experience begins with proper usability! (A site audit can help with this, although if you deal with smaller website SEO, you can use individual SEO tools, like an SEO checker, to check for usability issues).
Winning Answer Boxes
This is another test your content marketing team can run. Featured Snippet, or Answer Boxes, continue to show for many target keywords and testing tactics to win is a worthy undertaking.
The adjustment here would be to select keywords that are already ranking, say, in the top 5 for the test before making content or technical updates to optimize for this listing.
We've conducted updated research on the Answer Box that covers key tactics to follow for winning the SERP feature (if that's a part of your digital marketing strategy).
Adding structured data to pages is a solid SEO project and easy to test because it can be applied using JSON that doesn’t disrupt the main code of the page.
Put all the pages receiving the structured data into a Page Tag and find keywords for a test group (ideal keyword here should rank less than 7, in the running for the Answer Box). Keywords should also be selected that show any associated Rich Snippets you may be targeting with structured data such as Price, Ratings or Reviews.
Pages receiving this test and winning the associated Rich Snippet should receive an increase in CTR as a benefit of a structured data test and outperform their expected CTR based on their ranking position. Again, this is another opportunity to boost the user experience – starting right on the SERP!
You can see that there are many components that you can test for. See what approach our internal experts recommend you follow, here: The One SEO Test You Should Be Running.
Proper testing and measuring the impact of SEO enhancements is critical to any SEO campaign. They are useful for building new projects and learning the impact of site optimizations, and also site updates happening outside of requests from the SEO team.
Capturing SEO improvements and projects as "tests" can go a long way in building a high performing SEO program. Like a tree falling in the forest: If an SEO breakthrough occurred, but there was no testing approach in place to perceive the impact, did it really happen?
Rest assured knowing that your various SEO decisions have an impact on search visibility.