Just as nature abhors a vacuum, so too do marketers regard uncertainty.
Despite years spent learning the ins and outs of SEO, no marketer knows the exact weight of the more than 300 variables calculated in Google search results—including Google engineers. The company’s advice has been relatively consistent on the matter—create high-quality original content and get inbound links from high authority sites—but this still leaves a lot to the imagination.
There’s a vast difference between the unknown and the unknowable, however. Marketers can use testing to unveil the relationship between specific page-level variables and search rankings. Read on to learn more about:
It’s both risky and expensive to implement a change across your entire site, so it behooves you to evaluate the impact of changes before committing them in-masse. Using an SEO testing approach, you can change a single page or a group of pages and track the results. Overall, SEO testing gives you a competitive edge against more established players in your industry.
Using a framework for running SEO tests will allow you to base decisions on data instead of opinions.
These updates can help match current expectations while receiving more traffic, increased click-through rates, and higher keyword rankings. As a martech consultancy, we’ve found SEO testing to be amongst our most impactful offerings.
And this traffic growth helped influence a simultaneous growth in keyword rankings, seen here represented in the growth of total keywords that received first-page ranking.
These tests are particularly impactful because we’re often tweaking prominent page elements. These components are the first thing your users see before (hopefully) clicking to your site. In this case, first impressions matter.
Improvements in average SERP position across five SEO tests aimed at improving organic click-through rates of the respective pages.
The way SEO testing works is, you first form a hypothesis. For instance, “Due to Google’s apparent dislike of the overuse of <strong> tags, removing them from our keywords will likely benefit their SEO.
Then, you only change one variable at a time. For instance, one shouldn’t remove <strong> tags while making significant changes to body content. Any resulting changes would be impossible to attribute to the right variable.
SEO testing best practices throughout this process are used to ensure reliability. They allow us to get directional data about our test results that we can act on with confidence. Here are some SEO testing best practices to implement:
SEO rankings say something about how search algorithms interpret your site structure and content, and they certainly scale as traffic and other metrics improve.
Still, your end goal is better serving your users, so they remain on pages longer and are more attracted to click and interact with the website. Ultimately, clicks are a more meaningful measure of success than the numerical ranking doled out by Google.
You derive an additional brand awareness benefit by differentiating yourself from competitors. While this benefit may be hard to measure, sales generally follow familiarity.
You’ll also want to make sure that the pages you include in your searches aren’t continually moving up and down in rankings. This is a sign of volatility that could interfere with the results’ accuracy, potentially leading you to incorrect conclusions.
Moz’s Rand Fishkin recommends including pages that place between the eight and thirtieth result; too high, and the competition is too tough to isolate single variable changes. Too low and subtle changes will be hard to notice.
So long as you have a trustworthy analytics tool, you don’t need a dedicated testing tool to get started. But if you’re looking to do more than dip your toes in the water, or turn SEO testing into a service offering, you’ll want a tool to help establish and track tests.
One tool you can use is SEOTesting; it works with control groups and AB tests. Since it’s narrowly focused on SEO tests and experiments—instead of covering everything SEO, it makes the job much easier, helping users track changes, suggest complementary keyphrases, and perform historical tests. What’s more, it’s quick to learn, packs in clear visualizations of test results, and the team behind it has published a testing guide to help you dive deep.
It’s essential to ensure that your control group and your experiment group share the following characteristics:
Much like scientific experiments need experimentation to rule out the interaction of unforeseen variables, SEO tests require repetition, too. If you see that a given test is working, move it to other groups of pages or other keyword groups or subject matter.
Now that you’ve got the basics, it’s time to cover the specifics. The following eight tests cover a variety of SEO variables. Your mileage may vary.
Among the most popular and useful SEO tests, meta titles, and meta tags form the first impression your site gives to searchers, making their optimization crucial.
Remember, you’re not just writing for search engines. You want to rewrite meta titles and meta tags to be more click-worthy. Done right, and searchers may be more likely to select your page, even if it’s not the top result. The ensuing traffic boost may then lead to an increase in rankings as well.
Meta title updates for a round of testing, per our internal CTR optimization spreadsheet
We regularly run such tests for our clients and often see promising results. One test with a diet advice client resulted in a 70% boost in traffic.
Changing their format makes them more direct and compelling. Adding special characters like parentheticals, brackets, and em dashes, making them stand out from the rest of the results.
Title tags are the single-most-important variable in SEO rankings, trailed by subheaders and body copy, the URL, and the meta descriptions. As such, headline tests are among the most impactful SEO experiments possible.
How should you change your meta titles?
People understand stories on a deep level. Try more story-centric headlines to attract searchers’ attention better; any increased traffic from them can lead to higher rankings, creating a snowball effect.
For a range of reasons, removing links may benefit your site. It might keep users on your site, remove associations to low-ranked sites, or it may make your site stand out better.
On the other hand, including more highly-ranked resources that your readers find useful may make your pages more valuable, boosting your authority in Google’s eyes.
There’s reason to believe that comments and social share buttons may be taking a toll on your site. Pages, without any comments or shares, or very few, may signal a lack of credibility to searchers, thus decreasing their likelihood to share or comment, creating a negative feedback loop.
Google rewards sites that feature a diverse array of multimedia. Digital publishers, on the other hand, sometimes struggle sourcing and incorporating relevant and compelling visual elements.
After HubSpot optimized its content catalog, it saw monthly organic searches rise an average of 106%.
However, merely including more visual content may not get you the results you’re looking for. Such content should be relevant and add true value for those looking for specific answers to their search queries. Lean on unique, compelling, and relevant content. For more help, see our image SEO guide.
Internal links are a great way to help search engines understand the hierarchy of your content. That’s just one of the reasons why 42% of SEO experts spend so much time building internal and external links.
While adding relevant internal links willy nilly is better than adding no links at all, you should apply some methodology here. Cornerstone structure, also known as the concept of pillar content, is a highly effective means of arranging your content’s subject matter in an organized, intuitive manner.
Create pillar pages that broadly speak to a subject and interlink them with content that delves further into the specifics. SEO experts recommend integrating two to five internal links per page.
If your pages already use cornerstone structure, test by reorganizing the main navigation or featuring content pillars on your homepage.
Content is king. When all else fails, or when pages don’t perform well enough to conduct other tests, it’s time to upgrade core content.
Identify pages that perform poorly among the metrics Google values most highly:
Poor performance in these measures demonstrates that visitors to your site aren’t finding information relevant to their search. It’s an excellent opportunity to add content that may better answer a searcher’s questions and make pages more accessible for consumption.
Note that the optimal blog post length is between 1,000 and 1,500; bringing shorter pages into this range is in your best interest.
[add a data visualization image here — Straight Line analysis]
Some SEO tests appear to show that bolding keywords causes bad rankings. If you already have a habit of artificially strengthening content with such tags, now might be a great time to see what happens once they’re removed.
URLs can provide significant search indexing benefits, provided they’re concise and contain clear keywords. While the former best practices have been known, SEO experts have recently started suggesting the dates often structured into URLs may be getting in the way of better indexing.
The theory behind this one is similar to that of share tags and comments. Google and users compare dates on their search results and may naturally favor newer ones, regardless of whether such content is evergreen.
One noteworthy test, by Harsh from ShoutMeLoud, found that removing dates from URL structure helped his search rankings. Another test raised traffic 30 percent, according to SEO writer Nathaniel Towar.
Not all content is evergreen, though. For pages that have timely relevance, update them with new facts and updated context.
Note that should you try this method, you’ll first need to update your permalinks to avoid breaking links.
To recap, we covered five best practices for SEO testing:
And we also covered nine examples of SEO tests to run on your site:
The point of SEO testing isn’t exclusively about raising rankings and getting more clicks. It’s about gaining concrete knowledge of how audiences interpret your content, how well you respond to their curiosity, and what the web’s largest gateway values.
John Reinesch is a co-Founder and SEO Director at Chosen Data. Living on Long Island, he’s data-driven, loves a complex Google Sheet, and goes deep in his research. The resulting strategy files, automations, and execution quality have led to just as many snowballing charts about revenue from organic traffic as the amount of lifetime AppSumo deals he owns.
Please read our Comment Policy before commenting.