SEO Split-Test Result: Does Bolded Text Help Your SEO?
Before you start: what do you know about SEO split-testing? If you’re unfamiliar with the principles of statistical SEO split-testing and how SplitSignal works, we're suggesting you start here or request a demo of SplitSignal.
First, we asked our Twitter followers to vote:
56% of our followers guessed it right! The result was no significant change.
Read the full case study to find out why.
Recently, Google's John Mueller confirmed that bolding important pieces of text in your content can help your SEO performance, as it allows Google to better understand your content. This statement was made during a Google SEO office-hours hangout from November 12, 2021. To understand what the content on a web page is about, Google looks at several things to figure out what is being emphasized. Headings are a great way to do this, as are things like bold text on a web page. Google said it can add extra value to a certain extent because it's a clear sign of what the page topic is about.
Naturally, these statements sparked our interest, and we decided to put this to the test for a major Canadian software company to see if bold pieces of text could indeed improve their SEO performance.
HTML contains several elements for defining text with special meaning. Note that the <b> element and the <strong> element are not the same. The <strong> element represents the text of certain importance. The <b> element does not provide such special semantic information since HTML4 styling information of <b> element is deprecated. Most browsers will still draw bold text when you use the <b> element, but if you are using the <b> element with no semantic purpose, you should instead use the CSS property font-weight with the value "bold" to make text bold. If you use <b> elements with a semantic purpose, you should use <strong> elements to make text bold.
We hypothesized that bolding important keywords on blog pages could add extra relevance to what the blog is about. We wanted to validate whether bolding important keywords on the page would have a positive effect on organic traffic, so we did. The Test
We used SplitSignal to set up and analyze the test. 489 blog pages were selected as either variant or control. We used <strong> elements to bold and highlight the keyword/topical tags on the page. The Results
The result was inconclusive at the standard 95% confidence level. After running the test for 28 days, we saw a 1.5% increase in clicks, with a confidence level of 81%. This means that there is no statistical certainty that the result can be traced back to the bolding of important text on the blog pages of this site. To put it plainly, we cannot say with confidence that there is an effect for this optimization. Why
As mentioned at the beginning, proper use of semantic HTML can give a little more meaning or value to a page, resulting in a better understanding of the content for search engines. Looking at the outcome of just one test, we cannot say that this is never the case. We believe that this is a classic "it depends" situation.
Google says that while bold text can give a clear sign, it usually aligns with what they think the page is about, so it doesn't change much from that perspective. That also seems to be true when we look at the results for the website in question.
What we know from experience, however, is that the results of the same test setup can differ per website. Semantic HTML can make content more meaningful, so it might be worth testing this for your website.
Since this topic has really interested us, the outcome of this test and Google's statements about semantic HTML encourage us to roll out variants of this test on multiple sites. You can expect more on this topic in future case studies.
That said, it's never a bad idea to experiment with SEO optimizations on your journey to give both users and search engines the best possible answer. Good luck!
P.S. Do you know of any other Google statements you'd like us to test? Let us know!
Have your next SEO split-test analyzed by OrangeValley Agency.