What Is SEO Testing

SEO testing can mean many things; for many SEOs, this means making a change to one page or a series of pages and looking at the before and after data. While often directionally useful, this process has a lot of room for error. In the world of CRO, websites can test A/B versions of the same page by exposing a portion of traffic to the test and a portion of traffic to the control. Unfortunately, in SEO, this isn’t exactly possible when you want to see how Google or other search engines will react to a change. This is a challenge for the SEO industry, which leads to a lot of guessing, and we feel a lot of wasted time and resources. 

This is where the need for an accurate way to test SEO changes comes into play. What if you could make data-driven decisions about the impact of SEO changes before you actually make them? How would that change your approach to SEO?

SEO testing in its most accurate form is the practice of using statistical models to identify the true impact of changes you may make to a website that will either increase or decrease SEO traffic. In the case of Semrush, our tool SplitSignal employs a bayesian model and Google’s Causal Impact model to analyze results and tell you whether a change had a positive or otherwise impact.

SEO testing unlocks opportunities by helping SEOs to prove the direct ROI of SEO, as well as save developer resources by only asking dev teams to implement winning changes. SplitSignal does this by allowing SEOs and marketers to implement tests with a simple client-side implementation of a JavaScript snippet that allows you to make changes in the platform, not in the CMS.

Why SEO Testing Is a Great Fit for Marketplace Websites

Marketplace websites such as Amazon, eBay, Etsy, Upwork, and Grubhub have a series of unique challenges and SEO opportunities built into the business model. Firstly, these sites tend to have huge numbers of pages, often in the hundreds of thousands if not millions. Additionally, these sites often have large numbers of pages with a similar makeup or layout. Lastly, marketplace sites tend to get enough traffic to make analysis statistically sound, so marketing teams can feel more confident about the results of the test. 

With marketplace SEO, one simple change for a given element could mean a significant lift across millions of pages which then could lead to thousands of additional sessions or dollars in revenue. 

6 SEO Split Tests Marketplace Websites Should Run

Speak Your Audience’s Language

In one test, a leading marketplace for cannabis made a simple but powerful change. They changed one word, which took the name used primarily by the government and law enforcement for their products, and changed it to reflect one of the more common phrases that most people use. This led to growth in traffic through better connecting with the nomenclature used when searching for these products.

Below we can see that by week two of the test the variation group began to see a larger number of clicks. 

Brands, I know this is hard, but please try to resist the urge to get fancy in the way you talk about something if keyword volume shows you that people just don’t search in that way. 

Just call it weed. Nobody says where I can buy some “headband cannabis,” so why use that in your title tag.

By making this one simple change, this marketplace saw a lift of 2.7% in the test group. 

Make Titles Shorter

Short and sweet is usually a good motto to live by. Our data shows that this is no different when it comes to title tags. Studies over time and even recently have shown that titles that fall between 50 and 55 characters tend to get the most clicks. This means in many cases, using the full 70 characters (or 600px) available may not be the best approach. 

In this test, a leading marketplace for personalized gifts shortened the title tags for one of their most popular product types, t-shirts. The result was a 6.3% lift to the target test pages, which resulted in thousands of additional sessions. 

How many long title tags does your marketplace have? Probably thousands, meaning there is a huge potential opportunity for your site. 

Add “Free” to Your Meta 

While it’s true, you get what you pay for, who doesn't love free? In this test, that same personalized gifts marketplace added “Free Design Tool” to their meta descriptions. This test is particularly interesting because I’ve personally never come across a paid design tool on a self-serve personalized gifts site. That’s kind of like their THING! With that said, in the same way, our eyes gravitate toward .99 and by one get one free in supermarket aisles, the word “Free” is alluring in nearly any context. Even when the free item is assumed. 

The addition of this language led to a 2.8% lift in traffic to these pages. This is particularly notable since a meta description is not a ranking factor. CTR increases, in this case, are what led to more traffic. Savvy SEOs should never underestimate the power of a good meta description! 

Programmatically Optimize Header Tags

Well-optimized headers still seem to have a measurable impact on ranking. In one test, the cannabis marketplace and directory site added the word “weed” to the heading tag that stated “Buy [strain] Weed Near You” leading to a 2.4% lift in clicks from search. We think this helped the page in ranking better for terms like “blue dream weed” by having that exact phrase appear in an additional header. While keyword density is not a “real thing”, the occurrences of a target term on a page does matter, and this test proves that. 

Does your marketplace have generic headers that say things like “property details” or “find this product near you”? Consider programmatic solutions to enhance those headers with your target terms. Real estate websites often do this on their property detail pages. For example, don’t just use “Property Details” as a header across 1M pages. Instead, say “123 Story Lane, Belmont MA Property Details” as an example.

Near Me Optimizations

Many studies have shown that including “near me” optimizations can lead to higher rankings for near me searches. In this example, a marketplace for alcohol delivery added the words “Near Me” to CTA buttons. The effect was a massive 12% lift in clicks to the test pages. We think this led to growth in 2 ways. Firstly, this could be another case of the number of occurrences of a term having an impact. If “near me” didn't appear anywhere else, adding that target term in its exact match format very likely impacted rankings for those types of searches, leading to more traffic. The other benefit is one related to user experience. While we don’t have the data to support this theory, we believe that this optimization likely led to more engagement with the button by demonstrating that the tasty drink in question is available near you. 

A Need for Speed

In the post-amazon consumerism landscape SPEED is perhaps one of the most motivating factors, as well as one that leads to insanely high retention. This is why when one marketplace added ‘same day’ to their titles, meaning the product would be created the same day of the order, they saw a 9.5% lift. 

If you have made to order products OR simply can ship the same day for ready-made items, consider adding “Same Day” to your title tags and see if you also get a lift like this marketplace. 

Why These Tests are Perfect for Marketplace Sites

The above tests are perfect for marketplace sites for a few different reasons. Firstly, the ease of implementation. In each case, the test is based on a simple change that can be executed at scale with minimal developer time if any and does not require much page-level customization. Each page follows the same format, so the one change fits perfectly for a huge number of pages. In general, anything you can do programmatically to improve traffic is a good idea. Additionally, aside from SEO benefits, many of these changes help the usability of the page in question and could lead to higher conversions for all traffic, not just organic traffic. Lastly, they all had a significant positive impact. That impact means more users, more revenue, and more resources these SEO teams can allocate to other areas now that they’ve found winning tests. Are you looking for more marketplace SEO tips? Check out our blog from Aleyda Solis on winning marketplace SEO strategies. 

If you are looking for an SEO testing tool, be sure to check out SplitSignal. If you are looking for an agency to help guide you in the process of testing, contact me and the team at Stella Rising.