We love a good old SEO experiment here at Indago Media, it’s not that we don’t trust what Google says – actually it is. So many times over the years and especially recently what Google says contradicts what is actually the case.
Whether that is because they don’t actually know themselves what affects rankings in Google or whether they just don’t want to the SEO community to know. I think it’s a bit of both since rankbrain was announced it was basically Google saying we have no idea anymore.
Over the last couple of years, there has been talking that CTR was a ranking factor (directly or indirectly) and I was even at Mozcon in 2015 when Rand did an experiment on stage.
Google rhetoric since that day is that CTR isn’t a ranking factor.
So we decided to run a test. There are sites out there which allow you to purchase real humans to do a search and click on your result. This is important, it needs to be real people and not bots – Google is clever enough to know the difference.
During the Xmas break, I was bored one evening and thought I would start the test.
Obviously, this wasn’t tested on either our site or one of the clients. We have test sites for these test. We don’t want to hurt the long term rankings of our clients because we did a failed experiment.
So we picked a keyword where the site was ranking in position 3 in Google. The key term has around 500-1000 searches a month and isn’t what I would call a commercial term. Something which is going to drive sales.
The term is “product x vs product y”.
So our average position in the 30 days running up to the start of the experiment was 3.7 (taken from Google Search Console).
We paid for traffic for two weeks, about 30 people in the UK to search for the term and click on our results. The searches are then allowed to exit the site – just not bounce back to Google.
So during the two weeks, we had gone from 3.7 to 2.7 which is a whole position increase.
What was interesting is that our rankings for “product y vs product x” also increased from 3.9 to 2 and this has more search volume associated with it.
We then started ranking for a lot more of keywords that are on the page.
The traffic stopped last week so here are the final results.
“product x vs product y” – gone from 3.7 to 1.8
“product y vs product x” – gone from 3.9 to 2.1.
We will update this article in a month to show you the impact after the track has stopped.
But what I can clearly say is that CTR is a ranking factor.
How can this help you?
Firstly – we wouldn’t recommend you buying traffic – firstly it’s quite expensive and it isn’t going to convert and if the search volume is large your going to need some serious volume to shift the needle.
There are things which you can do to improve CTR without actually buying traffic.
- Add “Offical Site” onto home page title. There was a tweet a while ago from a Google employee which said that on Adverts simply adding the word “Official Site” to the end of the add-on brand campaigns can have a massive impact on CTR. Add this term to your home page (going to be a future experiment) to improve CTR
- Speak with your PPC team. These guys understand the importance of CTR and are always doing experiments. See what keywords, phases etc you can steal, I mean borrow from them to use. At a previous company, I had a certain budget for testing PPC text to be feed into meta descriptions, it was to improve CTR but at the time I didn’t realise it would affect rankings.
- Rewrite meta description – this is the first real description to sell your site and to get visitors to come. Use it well – don’t mislead people to what they might expect on the page – but defiantly sell your site and why they should visit you. We do have a guide for improving meta descriptions that we have been testing for a number of years across multiple brands and clients. We are thinking of releasing this in the future.
We want to re-run this test but for a commercial keyword with higher search volume.
We haven’t historically shared experiments and results in the past on our website, but going forward we will be sharing more results – both positive and negative results.