Author: Ryan Jones
Your potential customers are constantly putting your website to the test. They’re evaluating your brand based on relevance, content quality, user experience, and so on. They’re also comparing your site to your competitors’ to see where they should spend their money.
Diving into your web analytics can help you get a snapshot of your performance, but it doesn’t always help you understand which of your optimizations are moving the needle, or how well your content could perform in organic search. For that, you’ll need to conduct SEO testing.
In this blog post, I’ll help you get started with your own SEO testing by covering:
What is SEO testing?
SEO testing is the process of optimizing elements of your website in order to obtain data or a particular outcome in your reporting. You can test to optimize traffic, keyword rankings, or on-page metrics, like bounce rate.
You can also test as a way to correlate your SEO recommendations/implementations to business revenue, or simply to identify potential website issues.
SEO testing allows digital marketers and business owners to find and improve features and content on their website in a systematic way, enabling them to navigate Google algorithm updates by consistently improving content, and achieving long-term success away from the general spiel of writing new content, building backlinks, and maintaining technical SEO health.
Why SEO testing is crucial for online businesses
Any brand that relies on its online presence as a revenue driver and competes against other brands in the search results can benefit from SEO testing.
SEO testing becomes even more indispensable the more your business relies on organic traffic and the greater your competition becomes.
Conducting SEO tests allows marketers and business owners to monitor changes on their website to see what has the biggest impact so they can close the gap on competitors and get buy-in for larger projects.
While regular SEO testing can be very advantageous for all businesses, it is crucial for businesses in an agency or enterprise setting, perhaps even more so than for small businesses. Having a strategy-defined SEO testing plan can allow you to:
Create a data-informed strategy and help you stick to it
Learn what you can change on your website to have the most impact on the SERPs
Determine which tactics get you closer to your website’s goals, so you can put more time and budget into these activities
Verify what works and what does not, so you can take those lessons into future SEO campaigns (or client accounts, if you work in an agency setting)
When I talk about a “strategy-defined SEO testing plan,” I’m referring to making changes (testing) as aligned with your strategy, which should be defined by your company goals.
For example, if one of your goals as a company is to grow organic traffic by 20%, you may test meta titles and meta descriptions in order to directly bring more traffic in. If one of your goals is to improve the on-page time from 50 seconds to 1 minute and 40 seconds, then you might make changes to page copy or improve CTAs.
A strategy-defined SEO testing plan is not about making changes for the sake of change—it is about planning and testing in order to move closer to your company goals.
Getting started with SEO testing: The fundamentals
It’s one thing to change an element on your website and track the result of that change. This—in the most simple terms—is SEO testing. But, you can go further.
In this section, I will guide you through how to start SEO testing from scratch to move your website forward.
Formulate an objective to keep yourself on track
First and foremost, it’s critical that you identify your objectives and corresponding metrics. Are you looking to:
Increase organic search traffic to your website?
Keep visitors on your site for longer?
Improve your conversion rate?
Correlate a new page layout with a change in average order value?
Knowing what you want to influence with your optimizations and tests helps you stay on track and work in lockstep with your business (as opposed to operating in a silo).
There is a belief in some SEO circles that traffic and ranking are the main goals—that is incorrect. The main goal should be making your business more money. This means that SEO testing does not necessarily need to use search KPIs, you can use business metrics, too.
This is especially true for smaller businesses. You could think about tracking form conversion rates, page load times (for better user experience), and so on.
Whatever KPIs and objectives you set, make sure to record them within your strategy document. This will help ensure the people in charge of your SEO testing know exactly what the test is measuring when they come to track, analyze, and report on the results.
If you’re looking to uncover opportunities for testing (or simply find the highest-value area to start with), you could conduct a content audit and form some testable hypotheses based on it. For example, is there a blog post on your website from 2017? Could it benefit from being refreshed? This is a good idea to test out so you can have some idea of the ROI on a wide scale content refresh campaign.
Track data carefully so you can better interpret results
You’ll likely monitor the impact of your tests over a variety of metrics.
For example, let’s say you plan to refresh a blog post on your website and let the test run for six weeks (to collect data). To truly measure the outcome of your efforts, you’ll need:
Data on the performance of the content during the test
Historical data on the performance of the content during a corresponding time period before the test
Comparing these two data sets is how you’ll determine whether the test was considered a success or simply provided you with useful data.
The data you’ll need to collect is based on the KPIs you’ve designated for the test. If you want to test to see if the content refresh improves search traffic to the page, you need these numbers (from your analytics tool of choice). If you want to experiment with UX metrics like bounce rate, then you need to record this data, too.
Once you have all your data, visualize it with graphs and charts to compare the period before the test to the period during and after the test to find out how your performance changed.
At this point, I need to mention that tests that do not yield a positive result are not failures. Every test is a learning opportunity. If you have a test that has not yielded a positive result, go back and find out why. Have you missed certain information in the blog post? Is the URL not SEO friendly? There could be several reasons. The main outcome you need after processing an SEO test is a lesson. You can take these lessons into all future SEO tests.
Tools for SEO testing
There are several tools you can use when you start SEO testing. Each tool has its particular strengths and weaknesses, making it better (or worse) for certain types of businesses/use cases.
In this section, I will go through some of the main SEO testing tools (including a free option), their advantages, and how you can use them.
SEOTesting is a popular SEO testing tool that allows users to complete a wide array of SEO tests, including:
Single-page SEO tests
Group SEO tests
A/B tests and split tests
URL redirect tests
This comprehensive range of test types makes SEOTesting suitable for all business sizes and marketers of all skill levels.
However, it’s worth noting that SEOTesting does not allow you to make website changes within the tool itself. You will need to make these changes manually through your CMS or via your development team before logging it as a test in SEOTesting. So, if you are working with a client and find it takes a long time to get basic website changes approved and implemented, this might not be the right testing tool for you.
SEOTesting is competitively priced and intuitive to use, making it a great option for small businesses and agencies looking to complete regular SEO tests.
SearchPilot (formerly DistilledODN) is an enterprise SEO testing tool that focuses specifically on A/B testing for SEO.
SearchPilot has some features that SEOTesting does not have, which might make it a better fit for larger businesses that wish to deploy SEO testing in a systematic fashion across websites (e.g., large agencies).
SearchPilot allows you to make changes within the tool itself, meaning you do not need to create the pages independently from the tool, which can be a big timesaver. It also uses a neural network algorithm to analyze changes for you, letting you know whether a test has been successful or not directly within the tool with no outside input.
This is an enterprise-focused tool, which might not make it the right choice for smaller businesses or individuals who do not have a large budget to dedicate to SEO testing.
Google Search Console, Google Analytics, and built-in reporting
SEO testing, in its purest form, is making a change to your website and analyzing the results. When we think of it like this, we have all been doing SEO testing for a while. If you do not have the budget for paid SEO testing tools like SEOTesting and SearchPilot, you probably do not need to use them.
Simply hypothesize a change you wish to make and how long you want the test to run for, take the data for the time before the test, make the change on your website, and let the test run. You should be able to get a good, actionable feel for how performance changed based on the data within these tools and the guidance in this blog post.
To illustrate my point, let’s take a look at an example of SEO testing using nothing but Google Search Console.
An example of SEO testing (with free tools)
Let’s take a look at what this process might involve, step-by-step.
For this example, we will use a meta description change as our test. For ease (and because you may have access to a tool like SEOTesting or SearchPilot), I will explain how to conduct the test manually.
01. Start by exporting the raw data from Google Search Console for the period before the test.
You can do this by heading to the appropriate Search Console report, adding a filter for the page you wish to test, and clicking the “Export” button in the top-right corner.
02. Once you have all the desired data downloaded into a spreadsheet, implement the change you wish to test and submit the page for re-indexing. Once this has been done, let the test run.
The duration of the test depends on your website. If you are testing a page that does not gather a large number of clicks per day, you should leave this test to run longer to ensure statistical significance. A shorter test will likely work fine for other pages with more clicks per day. For most tests, I recommend leaving it to run for at least 14 days.
03. After the testing period ends, repeat the first step by downloading all the raw data from Search Console for the test period. This will include (but is not limited to):
Clicks per day
Now, you can open both spreadsheets (or merge the data into one) and compare the “control” period to the “test” period to evaluate whether the test was a success.
This is a very manual process, but it is one just about anyone can follow. You do not need to use any external tools other than Search Console to start SEO testing for your business.
SEO tests to try on your own website
Now that you know more about what SEO testing is, why it is important, and how to do it, all that’s left is some inspiration. Here are some potential tests to try out for yourself as you integrate SEO testing into your workflow.
Single-page SEO testing
As you might have assumed from its name, single-page SEO testing is great for small-scale tests on single pages. The page type does not matter. You could test your website’s homepage one minute and an eCommerce product listing page the next.
Great examples of single-page SEO tests include:
Title tag testing
Meta description changes
Changes to text within a single-page template
This is a very useful test type for a variety of scenarios, but especially if you work at an SEO agency and are looking to get buy-in from clients for larger-scale work.
For example, you could conduct a content refresh of a page, show the client the increase in clicks and ranking, and get more buy-in for more content refreshes. In other words, more work for your business.
Group SEO testing
Group SEO testing is similar to single-page SEO testing, but allows you to test a single group of pages. For example, your blog pages as a group or a particular topic cluster.
Some common examples of group SEO tests include:
Template changes to blog posts
Changes to specific sub-folders
You can also apply the experiments listed within the single-page SEO testing section as a group test. The only difference is that these test types run across multiple pages at once.
As an example, let’s say I want to change the layout of my blog posts to include a smaller hero image. I could download the data for all of my blog posts as a group, make the change, and then review the data after. I’ll be able to easily see if the test was successful.
A/B and split testing
First off, an important note: A/B and split testing are often used interchangeably, but there is a slight difference:
A/B testing refers to two web pages competing against each other.
Split testing ensures traffic is equally split between the control variant and the test variant.
That being said, A/B and split testing often work for similar experiments, so I am including them here.
A/B (or split) testing is effective when you want to decide which direction to go down for a particular group of pages or are trying to get buy-in for larger projects. Examples of this include:
Changing layouts of product listing pages or product detail pages
Mass changes to large areas of your website
Mass image optimization
Improving page speed
You can see the theme here: A/B and split tests are great for assessing the impact of large-scale changes before implementing the change site-wide.
Analyzing your SEO test results
If you take nothing else from this article, understand this:
SEO testing is useless if you do not fully analyze the result. The main takeaway is that you learn from every single test—not necessarily that the particular test was successful.
The first step is to analyze the results against your pre-test KPIs. Were you running an SEO test to increase clicks to a page? Did clicks to the page increase across the test period? If not, it’s time to figure out why.
It’s also important to look at data from multiple angles. Continuing with the example above, did you find that your test did not improve clicks? It would be easy to look at this and consider it a failed test, but that might not be the case.
Did your overall ranking improve for the page? If so, you could still consider this a successful test. It might have been a case in which the search volume for your queries reduced over the test period. Were there any outside factors that may have impacted the test? Was there a public holiday? Look at the whole test and its surrounding context, not just your KPIs.
Once you obtain all the data and have analyzed it from multiple angles, it’s time to start making decisions. Do you consider the test result successful? Have all stakeholders been notified?
If so, you can push the change live across the rest of your site (if this was the goal behind the test). Don’t need to make a decision right away? For example, if you ran a single-page test and updated a meta description, you could simply save the data, as it may be useful when writing meta descriptions in the future.
Saving data is massively important. It will help you optimize your SEO testing process over the long term. You can see what works and what does not work for your site, apply this to future tests, and become more successful over time.
Today’s SEO tests may be tomorrow’s key to success
Most SEOs and site owners know about the importance of:
Amazing technical SEO
But, what happens when all of that seems to be the best it can be? What do you do next? Planning and executing a comprehensive SEO testing strategy for your business or your clients’ sites might just be the path forward you need.
It can allow you to overtake competitors, get a better understanding of how the search algorithm works for your chosen keywords and topics, and make better decisions for future clients or your own brand.
Ryan Jones - Marketing Manager at SEOTesting Approaching a decade in marketing, Ryan has worked in-house and agency side. From scaling an eCommerce business from £400K to over £1.4M in annual revenue to increasing conversion rates for small, family businesses, Ryan loves making marketing work for the masses. Twitter | Linkedin