SEO Split Test Result: Will removing many H1 improve clicks per page?

Before you begin: what do you know about split-testing SEO? If you are not familiar with the principles of statistical SEO split-testing and how SplitSignal works, we suggest you start here oz request a demo.

---

first we asked our followers on Twitter love:

Here is what other SEO experts have to say about this test:

Boštjan TankoSEO specialist at APOLLO Insurance:

In my opinion, an article or landing page with more than one H1 is confusing for Google to figure out what the page is about.
In the case you added: https://khn.org/ is a different story, as their home page acts as a portal for different pages, so there will be no difference in this case, and it’s okay to have more than one H1.

Vladimir GertnerSenior Project Manager at Soft Road Apps:

Well, the element is there to define style. It does not necessarily affect the ranking per se, as it is not semantic. It is not used to add weight to an article.
John Muller also said that more H1 is not a problem when it comes to SEO. Since we can’t trust anything a man says, I would say we can expect positive change.
As far as actual official data are concerned, this should not be a significant change. Maybe 4% more considering the changes are made, and not negative at all.
However, if you use instead of the address tag , I would not expect perfect results. H2, for example, sends a message to Google that the article is designed the way it is designed, and G spiders check H2.
So if you did not include any address tag and use only instead of the title tag I would say is not ideal.

Find out if our followers were right by reading the full analysis of this test.

Case study

Hello everyone, welcome back to the new and interesting case study.

The site we are looking at is a non-English content aggregation site whose individual articles are not formatted correctly. The main problem is that instead of using properly nested address labels (H1, H2, H3, etc.), each the address is H1.

Hypothesis

We assumed that by removing the confusion about which address is the preferred address, we would notice an improvement in the number of clicks if we left the above H1 address in place and converted others into tags .

It seems simple; what do you think will be the result? Read on and find out!

The test

The SEO experiment was set up and configured with our SEO testing tool, SplitSignal. This is not a particularly large site and has just over 50,000 indexed pages, so we chose a total of 335 pages, 158 for the test control and 177 for the test version. How does SplitSignal divide pages?

On all 177 test pages, the upper H1 was left in place, and the rest were dynamically converted to tags . The test lasted 21 days for confidence level 99%whereby Googlebot re-searched 100% of the test pages during the testing period.

Results

After twenty-one days, we observed a statistically significant, NEGATIVE result for pages with converted H1 addresses.

We asked you earlier what you think will happen. Business SEOs on LOKOMOTIVA Agency came after the experiment and gave their explanation as to why they think this happened. However, we would love to hear your opinion in the comments on what you originally thought and do you think it could be something else?

Overall, the test pages received 2.4% fewer clicks during the test period, which means a loss of 2,680 clicks.

As we’ve said, it’s getting harder and harder to surprise us, and as a longtime SEO, this result would be what you’d expect given the knowledge of how often Google has to fix or assume development errors.

Ever since web standards have existed and search engines have tried to use algorithms, machine learning, machine vision, etc. to understand websites, web developers have not always followed these standards to the end. This leaves search engines with the unenviable task of trying to figure out what the hell we mean when we’ve done something in an unconventional way.

Fortunately, as we see in this test, Google has become pretty good over the years and has probably already ignored additional H1 addresses and assumed they were H2, H3, and so on.

H1 removal and their conversion into an inserted element such as a label , removed the advantage given to them by the title element, even though it was used incorrectly, and caused the loss of relevance of these titles to the entire current authority of the site. , resulting in a drop in average rankings and a causal drop in clicks on the page.

It would be interesting to see a similar test that would instead convert H1 to converted into appropriately nested titles, but since each article is different, this would be difficult to achieve without moving blind assumptions. forward. Nevertheless, the test could reveal more about the assumptions Google makes about this site.

What do you think?

---

Let your next split test be analyzed by SEO technical experts LOKOMOTIVA Agency.