SEO Split-Testing Lessons from SearchPilot: Emoji in Meta Descriptions
Start here: how our SEO split tests work
If you aren't familiar with the fundamentals of how we run controlled SEO experiments that form the basis of all our case studies, then you might find it useful to start by reading the explanation at the end of this article before digesting the details of the case study below. If you'd like to get a new case study by email every two weeks, just enter your email address here.
For this edition of the #SPQuiz, we asked our Twitter followers what they expected to have happened when we replaced numbers with emoji in meta descriptions. The response was mixed:
No answer gained an overall majority of votes, but the largest group thought that this change wouldn’t have a significant impact on organic traffic. Only one in four of our Twitter followers thought that this change would be negative. In this case, it turns out that the minority was correct. As you’ll read below, this change was negative for organic traffic in this case, a fact that 75% of people got wrong!
The Case Study
The use of emoji can be a divisive topic in SEO, especially in the context of trying to get them shown in search results. On the one hand, they can be eye-catching, and an easy way to get people to notice your result and potentially click on it. On the other hand, they can come across as gimmicky, and some searchers may be put off from clicking because it looks unprofessional.
Before we get into the details of this particular test, a bit of background information on emoji and how they can be used for SEO purposes. Emoji are characters that form part of the Unicode standard for text, alongside letters, numbers, punctuation and all other characters in a wide range of writing systems.
Since they are just Unicode characters, they can be used within HTML elements the same way as any other character. As such they can be included in title tags, meta descriptions and even structured data markup. In general, Google will pull these characters through into the search results as it would for any other character. Note that the emoji will be displayed in the default font of a user’s browser, and as such may be displayed differently to how you might expect.
In this particular test (launched in July 2020), we wanted to test replacing the number of items listed in a website’s category page meta descriptions with the equivalent numerical emoji. For example, 125 would be replaced with 1️⃣2️⃣5️⃣. Below is an example of how this would look in an analogous website’s Search Engine Results Page (SERP) snippets.*
The hypothesis behind this was that the emoji would add some colour to the SERP listing, in a vertical where a lot of competitors have very similar results. This would ideally attract more clicks from the SERP. Given that we couldn’t be sure of how searchers would react to the different snippet appearance, and there was a good chance that something this unusual could have a negative impact, we tested this to find out the impact on organic traffic.
Below is a chart showing what the impact was:
The overall impact of this SEO split test was a 5% drop in organic traffic for the variant pages compared to their expected performance. This result was reflected across all device types, with a 9% decrease on mobile, and 4% on desktop.
Given that this result was so overwhelmingly negative, we rolled back the change. We generally assume that meta descriptions have no direct impact on rankings, so we can be confident that this result was due to users being put off from clicking on the results that contained emoji. This may have been because the emoji appeared ‘unprofessional’ in this vertical, or that they made the snippet more difficult to read. Alternatively, some users may not immediately read the emoji as numbers, as they can be mistaken for bullet points at a glance. This result could be different in different verticals, depending on user expectations and competitors.
In general, changes that directly impact the search results (such as titles, meta descriptions, and structured data for rich snippets) can have an unpredictable impact, given that users will see the changes and behave differently. Humans are hard to predict! This is why testing these kinds of changes is so vital.
*Some browsers, primarily Firefox, may see these characters displayed slightly differently. For this test, Firefox accounted for a negligible proportion of traffic, so the result was still significantly negative when just considering browsers where it displayed correctly.
Get case studies by email:
How our SEO split tests work
The most important thing to know is that our case studies are based on controlled experiments with control and variant pages:
- By detecting changes in performance of the variant pages compared to the control, we know that the measured effect was not caused by seasonality, sitewide changes, Google algorithm updates, competitor changes, or any other external impact.
- The statistical analysis compares the actual outcome to a forecast, and comes with a confidence interval so we know how certain we are the effect is real.
- We measure the impact on organic traffic in order to capture changes to rankings and / or changes to clickthrough rate (more here).