This comes straight from Rand Fishkin from a Whiteboard Friday a couple of months back called “8 Old School SEO Practices That Are No Longer Effective”. He claimed that Title and Meta Descriptions will get higher CTR if they are written more like AdWords Ads, more natural, less keywordy, and more of a message.
Not that we 100% disagree with Rand on this one, but we decided to test it and call him out since he made this claim without presenting any evidence to support it.
In our last dossier we tested to see which meta title would outrank the other, a meta title optimized for a keyword or a meta title optimized for clicks.
Not surprisingly, the meta title optimized for keywords won.
To be fair to Rand, we ran this CTR test to see if perhaps the CTR gain would be so great that it would be better to lose one ranking signal (keyword placement in the meta title) but gain another (higher CTR).
If you didn’t see it, Rand used the example of Tobacco Pipes. He put together a keyword stuffed SERP listing and a more natural, message based SERP listing. We’ve put these examples below.
It’s virtually impossible to accurately test one against the other in a real world environment. So we created a virtual environment to test this using a fantastic (and FREE) piece of SEO software called SERP Turkey, by Tom Anthony.
You input your keyword, you upload the results page into SERP Turkey and you edit the SERP’s you want to test against one another.
You then use a service like Mechanical Turk by Amazon (http://mturk.com/) to get real people to click the ads in your test SERP and generate real life CTR data on each result.
This technique allows you to test different meta titles and descriptions, so that you can find the highest converting variations.
For this test, we set up 2 pages using the keyword Tobacco Pipes. There were 200 human testers.
Here is an example of one of the test SERPs from SERP Turkey:
Now we had to remove the variable of position because of course, position 1 gets the highest CTR. So we created 2 test SERPs and switched the test results between position 1 and 3.
The Mechanical Turk test took about a day to complete for 200 testers at an average hourly rate of $8.00.
So pretty good value, the total cost ended up being about $12 and we probably could have got it for less if we tried.
First test with the message ad variation in the #1 spot and the keyword stuffed variation in the #3 spot.
Test 2 switching the locations so the keyword stuffed variation is the #1 spot and the message ad in the #3 spot.
This myth is busted! In our test, the keyword stuffed result produced in total, the highest CTR with a total of 56.34%
To 100% prove this, we should probably re-do the test and see if we get the same results, but we don’t think that’s really the point here. There are a number of lessons from this.
As we repeat every month, don’t believe everything you hear, even if it comes from Rand Fishkin.
The jury is out on whether or not keyword stuffed Titles and Meta descriptions are better than ad like, message based ones. Test it yourself using this method and find out what’s best in your specific case.
As time has gone on, we felt the need for Clint to revisit these tests again. In this video, he gives his opinion on if naturally written title and meta tags get a better CTR.
Test 35, this is a good one. Does naturally written title and meta tags, get a better CTR than keyword stuffed in the SERPs? It’s kind of interesting how we knocked this out. So I want to try it out and just show you what’s going on. So let’s get into it.
Back when this test was completed, there was a big push to worry more about CTR in your title tags than optimization. And we’ve talked about this in some previous videos of the title tags tests that were done. But the short version is, back then or even now today, there is a big process, I want to say. Like we were writing our title tags with just keywords of mine. So, blue widgets, or wooden pipes, tobacco, wooden pipes and another pipe, custom wooden pipe. So there’s three keywords inside of our title tag and the idea was that, we’re actually hurting our click through rate because we weren’t making, I don’t know, clear, concise, National Enquirer type title tags for all of our stuff.
And so we wanted to test that out and see which one actually got a better CTR, which one answered the results of the searcher a little bit better. Something this fancy clickbait-y type or just the keywords that say – hey, this is what we are looking, or what this page is about.
So we use a service, it’s called SERP Turkey, you can still find it, just Google SERP Turkey. It’s on Tomanthony.co.uk is the domain, you’ll see it, and then you just put your keyword in here and then you follow the instructions. And what you’re doing is creating an SEO test and then you take that SEO test to Mechanical Turk, and you hire people to look at it and follow the instructions and you set up essentially, a CTR. Pick the one that’s best for you, pick the one you are searching for this keyword, pick the one that’s best for you. And it’ll offer the selections and then you figure out the variations, and how you want to go about doing that.
That’s actually pretty cool, especially if you’re doing some long term stuff, evergreen pages and you can actually test those two variations and see which one’s going to get you the better CTR.
Nowadays, I would recommend you actually use kind of a combination of make sure your keywords in there and then make sure it’s clickable at the same time. And we talked about that in another video and I gave an example of that.
But what we got out of this one is the message ad variation at the number one spot and the keywords the variation in the number three spot. That’s kind of how it ended up. And then this one is the keyword variation is the number one and the message was in the number two. That’s the ranking perspective. Now in the CTR perspective, the highest CTR was actually the one to get the keyword stuffed. So the wooden pipes, custom wooden pipes, handmade wooden pipes. That version actually got the higher CTR versus the less wordy optimized one for CTR, as it were.
Could it be that the copy just sucked? Could have been. But it just kind of gives you an indication. And also, we looked at the behaviors of users and looking at the title tags and determining that what that page is about in a quick glance and then you do that through the title tags and sometimes, you can miss the boat if you’re trying to be too cute with those things – with your title tag. So think about that.
Try SERP Turkey, it’s still out there, you can use the Mturk test model and test some different keywords especially if you’re building templates. If you do page templates and you’re making sites that target say 30 or 40 cities, and you need a template so that you don’t have to rewrite 30 or 40 different pieces of content and you just test that the title tag and Mturk. See which one is going to get you the better CTR and then pop that in there. And hopefully, if everything goes properly you’re doing your on page right, your schema, and your technical SEO, and your title tag is there, you’re going to get into the top five, maybe top 10, and then your title tags, it’s already proven to be the one with the highest CTR, could actually carry you up because it shows up and it’s attracting people in the clicks themselves, will carry the page up. And you don’t need as many backlinks, which is really really good.
Alright, so hopefully you find that useful. Maybe we’ll do another test with SERP Turkey and just kind of show that off a little bit. I think it’s a really cool tool and really useful.
If you want to read more about SERP Turkey and testing CTR in your Titles and Meta descriptions, there is a really great blog post written by Tom Anthony in the Moz blog here with more detailed instructions on setting up tests: https://moz.com/blog/split-test-gather-ctr-analytics-serps
Wasn’t that fun? We also have tests on the actual effect of CTR on rank. Check out our test articles for more details on that!