By now, you've likely noticed how difficult it is to define a single, surefire way to rank well in an organic search.
Sure, there are plenty of concepts to help with achieving rankings and driving traffic, and we’ve shared countless of them on our blog already (like the internal links workflow for site authority, the key elements of SEO content writing, and these modern keyword research techniques).
Ask me for a single, ultimate strategy however, and I'd have to give you an array of answers based on your SEO goals. That’s because, at its heart, SEO has a predictability dilemma.
As practitioners, we can never be sure how a particular strategy or technique we use is going to work. We know how it’s supposed to work, but too many variables might affect the desired outcome.
Yet, our success depends on doing what works. So, how do you figure out what works for your site, industry, or audience?
Before I tell you what SEO tests our experts say you should run, let’s discuss the predictability dilemma a bit further.
The Predictability of the User Search Experience
Think about your work for a minute.
Even though you optimize content to rank in search engines, you have to rely on third-party data and information to do so with little assistance from Google.
Most of your knowledge comes from research studies that SEOs conduct, in fact. Those studies offer unparalleled insight into what makes sites rank well. Unfortunately, those studies are mostly based on assumptions validated with data to some degree.
So, to improve your search visibility, you have to follow what the industry has identified as best practices, while also aligning everything you do with what the user expects or is looking for.
While I'm not here to undermine the importance of SEO, I do want to highlight that, with little guidance from Google, the only way for us SEOs to know how to connect and engage users is by testing our ideas.
What is SEO Testing?
I define the concept as the practice of testing an SEO hypothesis to prove or disprove it and influence future actions.
Let me illustrate this with an example. Let’s say that you’ve noticed a situation that looks like the image below in your content’s performance.
Each of those pages attracts a significant number of impressions; yet, their organic click-through rate (CTR) leaves much to be desired. To solve this challenge, you formulate a hypothesis.
Hypothesis: The search listing is disengaging and misaligned with the user’s intent when looking for that information. As a result, most users dismiss this page as irrelevant, choosing to view other content instead.
To prove or disprove it, you run an SEO test (maybe you update the tag, update some on-page SEO basics like meta description, title, and keyword usage - you get the idea). You also mark that event in your analytics software and monitor for any potential changes in performance.
If, after a couple of weeks, you notice a drastic improvement, you can assume that the hypothesis was true. If the results don’t change, however, you know that you have to look for other causes that are affecting your performance.
In short, going through the testing process helps you:
- Improve your site’s performance by identifying potential problems that are holding it back.
- Expand your professional experience, as each test teaches you something new, even if it disproved your theory. The more you test, the better you understand how to drive results, and, in turn, make more informed decisions.
- Determine the best future actions to take based on what the users want.
What Can You Test in SEO?
There are no limits to what you can test in SEO. Because you can work on almost every on-site and on-page SEO element, you can also test their effect on your rankings.
Below are just some of the ideas for tests you could run:
- Testing the effect of expanding or updating the content on page’s rankings
- Improving page speed
- Writing more engaging title tags and meta description (while retaining their keyword targeting)
- Adding images or other elements that will hold a person on a page for longer and measuring the effect of engagement on rankings
- Improving internal links’ anchor text to position a page for more relevant keywords.
And, that's barely scratching the surface on SEO testing!
In fact, to show you how professional SEOs test their theories, we’ve asked members of our team to recommend some of their favorite SEO tests. Below are their recommendations for the one SEO test you should be running.
One SEO Test to Rule Them All (According to Our Experts)
Tyson Braun, Senior SEO Business Consultant, Professional Services
Personally, my favorite test is validating different meta tag data to improve the click-through rate. In my tests, I typically focus on the format that works best for a particular content. For example, I add numbers to the title tag or different values in the description to make the listing stand out.
I’ve written more about my approach to testing title tags here, in fact.
Julia Wisniewski Stewart, Senior Client Success Manager
For me, the one true test I run constantly is A/B testing the title tag length. Test what Google likes the most in the title tag, without changing it for you or replacing with what it considers a better tag.
Chris Sachs, Director of Client Success
While many SEOs have testing methods they prefer most (A/B, multi-variate, on/off), I've learned that enterprises need to be cautious with, when and how they test. All companies buy into the need for testing, but visibility can drop fast and recover slowly.
I see many large enterprises take these approaches to testing:
- Test in small batches. Pick one, few or a section of pages to test.
- Test non-strategic pages first. A full-blown content test on your #1 traffic page is a huge mistake
- Try isolating web browsers to test the smaller traffic population. If you know the majority of your traffic comes from Chrome or Edge, try directing traffic from Firefox to your test to mitigate user impact
Kris Leasure, Senior Client Success Manager
Something that's come up a lot lately is using schema and different content formats to try to win answer boxes. There's no real "do this and you'll end up in the answer box," but we've been encouraging clients to test out things like adding lists to their content to make it more answer box friendly.
Another thing I can think of that would be fairly easy to test is using Content Fusion for better rankings over time. Take the keywords/pages where you're ranking in striking distance and tag them for the test, then optimize each page in the tag with some suggestions from Content Fusion. It would be a bonus to use Traffic Potential to forecast the benefit of improving the content in the tag if there's a need to demonstrate the business case beforehand.