Watch The Video
Transcript: How to ACTUALLY Do Split Testing on Google Ads (AdWords) Text Ads
We already know how Google wants us to A/B test our adverts, because when you set up a new campaign, it goes through the process and it shows you how to create multiple ads, which is part of the process, but is that the right way to approach it? How do you test your ads effectively? How do you know what’s working? How do you know what’s not working?
How do you know what success is when you complete an ad test? Is it more conversions? Is it a better click-through rate? Is it a mixture of both? Well, in this video, I’m going to break down how you should be running your A/B tests on Google Ads, the scientific way, which isn’t necessarily the Google way. Coming up.
Hey guys, Darren Taylor of thebigmarketer.com here. My job is to make you a better marketer. Now, if you’re new to the channel, I’m Darren Taylor, a digital marketing trainer and consultant. I specialize in search engine marketing, SEO and PPC. If that’s of interest to you, you should subscribe to the channel, and check out my other content as well.
In this video, we are looking at the age old question of A/B testing, which is something specific to Google Ads, but also specific to marketing in general. First of all, it makes sense to quickly understand what an A/B test actually is, when it comes to Google Ads. An A/B test is simply running two ads within the same ad group, against the same keywords, to see which one will perform better for people searching for your services.
When somebody goes to Google, they type in a search term that triggers your ads on Google. You serve different variations of your adverts, and then you decide when you get the data from users, which one performs the best, and you go forward and use the one that performs better, and then you do it all over again. It’s a wash, rinse and repeat of testing. You keep iterating your ads and improving them, in order to get to a point where you improve your ads completely.
We know what testing ads is in Google Ads, but what is it that Google asks you to do? How does Google ask you to test your ads in Google itself? Well, if you’ve set up a new campaign recently, you will know Google asks you to create three ads per ad group. The reason for this is, because they want you to create two standards text ads, and one responsive search ad.
You’ve got three ads running against the same ad group, and similar keywords as well. You can have an A, B, C test for your text ads, but is this the right approach? Well, for a start, responsive search ads are already a bit of an issue on my part, I think they perform worse than standard text ads. I’ve seen this in the data from many, many campaigns, I’m sure you have too.
I’ve made a video about this very subject in the past, so I’ve linked that up for you in the description as well. Honestly speaking, I think responsive search ads aren’t a great thing in general anyway, but then you’ve also got the two standard text ads to run alongside your responsive search ads. That is how Google wants you to set up your ad tests, but let’s go into why I think that’s wrong.
If you head over to Google, and you create a new campaign, a new ad group, and you create three variations of your ads, and you write completely different headlines, descriptions, and all different types of ad content across your three ads, you run a test and you start getting data, and you can clearly see one of your ads has a better click-through rate, or is driving more conversions than another version of your advert.
Okay, you found your winning ad, brilliant. You’ve picked that ad and you’ve put that into your campaign as the main one to use, but of course, with Google Ads you don’t stop testing. You want to run another test to see if you could improve things even more, but what’s the problem here? How do you know which part of the ad resonated with the user? You don’t know specifically what’s worked and what hasn’t worked.
If you’ve written two completely different text ads in terms of three different headlines on here, three different headlines over here, two different descriptions here, two descriptions over here, they’re completely different ads, and when one’s working and one isn’t, you don’t know which element of the ad is performing. What should you do instead? Well, you need to go back to when you were at school and did science, and you need to do what’s called a fair test.
Now, a fair test means, you change only one element at a time, that’s only one variable in the test. Therefore, you can measure specifically, what’s working and what isn’t. In this occasion, what you shouldn’t be doing is creating an ad with different descriptions over here to the ones over here. Keep the descriptions the same, and start testing headlines initially, because that’s going to give you the data you need to determine whether or not your headlines are working.
When you complete that test, then you can go and create another test to test your descriptions as well, because that way you would know specifically, which headlines are resonating with the user, and then after you’ve done that one, which descriptions are working for the user. It’s really important you separate the two, because otherwise, you’ll be lost, you’ll be writing ads, and not even sure whether or not you’re changing the right part, because you’d run an A/B test, you get a successful ad out of the test, you write a new ad and you’ve made it worse than the previous ad, which means you’re going backwards and not forward with your testing.
Another thing you really need to be aware of when testing your ads on Google Ads, is if Google decides the winner of your test too early. This can happen, Google can decide when you set ads to optimize automatically based on performance, that they decide that ad A is better than ad B. The problem with that is, if it’s done too early, you don’t actually have enough data.
Even though Google thinks it knows already at the very early stage, which one performs better, sometimes it gets it wrong, and I’ve seen this many times in campaigns. In that case, what you should do if you’re seeing this in your campaigns, is that you should override Google and rotate the ads indefinitely for a period of time. Therefore, you’ll get even data for both of your ads.
You can run a test and see specifically which one is performing better, and then from that point of view, you can remove or pause the one that’s underperforming, and create a new ad in its place. That’s really important to note, because Google can sometimes decide too early, so that’s something to be aware of. Now, you know exactly how to run a fair test in Google Ads to know which parts of your ad is working, and resonating with the person searching for your services.
Now, after you’ve done this, what does success look like? What is the marker for being a successful ad? Now, it depends on the performance of the ad across two specific metrics, which should be quite obvious to you by now. The two metrics are click-through rate and conversion rate. Now, these two things are really important for your ads, because of course, people see your ads on the search results page.
The one with the best click-through rate means it’s driving the most traffic, which sounds great, because of course, if you’re driving traffic, we know click-through rate is one of the major parts of your quality score, meaning, you’re going to get cheaper traffic costs and a better, more efficient campaign. However, that’s not always the case, because if you’re getting a fantastic click-through rate, tons of traffic coming to your ads, and you get zero conversions from that specific ad, there is a mismatch between the messaging in your ad and what you’re providing the user, despite the good click-through rate. That is a problem in itself.
When you look at results like that, you know the conversion rate is something you need to pay attention to more than the click-through rate. You should be taking a lower click-through rate, if you’re driving no conversions. You should be trying to make sure your ads are working from the users’ point of view, what their expectations are, to what you provide on your website.
You may think you’re doing a great job optimizing your ads, if your click-through rate’s improving and improving, but if your conversion rates aren’t staying consistent, or they’re going down, then there’s an issue there, you need to make sure you’re matching your content with your search. That, in a nutshell, is exactly how to create a fair A/B test in Google Ads, and understanding what specifically within your ads is working, and what specifically isn’t working.
If you liked this video, please leave a like below. Let me know in the comments, if this is the way you’ve been testing your ads, or whether you’re going to try this new scientific method of making sure your ad tests are working successfully. More important than that, don’t forget to subscribe, check out the other content across my channel. I’ll see you guys on my next video.