<aside> 🚨 New here? Subscribe to my newsletter to receive articles similar to this one: alexsavy.com

</aside>

I’ve been writing a lot about Reddit marketing lately because that’s what I’m deeply focused on these days (the next article will be about Reddit Marketing).

But I’m still heavily involved in SEO and AI SEO/GEO/AEO, and I’m constantly running experiments.

Today, I want to share one of the tests I’ve been running over the past year and the results were fantastic, especially considering that I didn’t update the content or build any links to the pages or domains.

I ran this across multiple pages and 5 domains to make sure the data was accurate.

I wish these tests were faster, but with SEO, if you want reliable results, you often need to wait.

So, let’s dive in.

About the experiment and why I decided to run it

Google heavily relies on user clicks in the SERP and their behaviour to understand the usefulness of individual pages.

This means that if people consistently click on your page over others in the SERP, Google should rank your page higher.

This concept is widely known in the industry as CTR manipulation.

It used to work very well, but last year when I sent traffic to influence rankings for specific keywords, the results were much harder to see. They were mixed, while in previous years the impact was fast and obvious.

Since Google clearly relies on click data, I did not believe the tactic suddenly stopped working.

My thinking was that if Google assigns a type of rating to pages, then older pages probably already have a strong established rating and it may take a long time to move it. That means we need a new page with a fresh rating. But new pages also take time to grow.

So the question became: how can we publish a new page and make it not so new?

The first answer that came to my mind was to use rel=canonical and merge the pages.

This way, the URL is new, but all the link equity flows through the canonical to the new page, and then you can focus on improving user behavioural signals.

So what I did:

I ran a series of tests on five domains by changing page URLs, merging them with canonicals, and sending traffic through specific keywords to the new pages while also making sure that I received branded traffic to the same new pages at the same time.

The first thing I did was drive “brand + keyword” traffic.