Testing Testing 1, 2, 3
Results for ad copy testing can vary account to account. With Google stating that a good click through rate is around 3%, how big does a change in click through rate have to be to indicate a truly successful experiment? Click through rate will vary between campaign types - Brand campaigns typically have a higher click through rate than more generic campaigns and Display campaigns typically have the lowest click through rate of all.
With what defines a good click through rate having so many factors I’ve decided to simply follow this mantra: "A good CTR is one that is consistently improving and being optimized." (PPC Hero). With that in mind, what can we do in order to ensure that our ad copy is not getting stagnant and that we are continuously striving to improve our click through rate? The simple answer is testing!
A Capital Offence?
There are many factors that can be experimented with in ad copy testing in order to try to improve your CTR. A regular subject of debate at Periscopix HQ is the potential differences that using sentence case or title case can make to an ad’s performance, so using some real client data I am going to examine whether there is any discernable difference in CTR between the two styles of ad writing.
Firstly, some quick definitions; sentence case is when only the first word and any proper noun in the ad copy is capitalised. Title case is when every word in the ad copy is capitalised, and many consider this to be grammatically inaccurate but potentially more attention grabbing. Can Sentence case vs. Title case in ad copy testing really affect your click through rate? If you have strict restrictions on your ad copy & don't want to change your message, but want to see if you can make your ads stand out, character case A/B testing can be a great way of experimenting. You could be forgiven for thinking that character case is a small and maybe even inconsequential design choice, but opinions on which is best are divided! Some think Title Case makes your ad stand out better, some think it's 'shouty'. Some think Sentence Case makes your ad easier to read, others think it makes it easier to ignore. I'm going to show you the best way to set this up to see if this quick and easy experiment can have long term effects on your accounts' performance, as well as have a look at some data to see if this test has had a conclusive impact on ad performance.
To start with you may want to test this on one ad in your ad group, if it is successful you can always roll it out across other ads at a later date.
A/B Testing Recipe: –
- Weigh out an ad group with a high amount of traffic and a work out which are your better performing ads, as this will be the most efficient way to get the data you are looking for.
- Select the ad that is performing the best. (If you are running a mobile specific ad with the same message as your other ad, you can repeat the process for this ad too.)
- Duplicate this ad twice and pause the original, so you have a pair of new identical ads to use in your test and the original to revert to at the end unharmed if needed.
- In one ad from the new pair, change the ad copy to all Title Case.
- In the other ad from the same pair, change the ad copy to Sentence Case.
- Pause all the other ads in the ad group – leaving only the duplicated ads live.
- Using the Labels feature in AdWords, label these ads as tests, with the date. You'll then be able to easily go back and see performance before and after you set them up.
- Your ads should then look something like this.
- You'll need to ensure you have your ad rotation settings set to "Optimise for clicks". The "Optimise for clicks" setting allows Adwords to give preference to your ads that are expected to get the most clicks, based on your historical CTR.
- Let the ads rest in your account until you have accumulated a good amount of data. . This will vary according to how much volume your campaign gets but you probably don’t want to look for results till both ads have received multiple hundred impressions.
We implemented this AB test at the beginning of January 2014, and have been some interesting results so far, and where they are not conclusive, it is interesting to have a look and see if we can use this data to inform our overall account strategy.
The green line on this chart shows the date we launched, and we can see that initially the Title Case ad outperformed the Sentence Case ad. Where we saw a drop in CTR heading into February, both ad formats dropped at the equivalent rate, indicating that this drop in CTR was not due to any input from the ads themselves. It is interesting to see that we have seen more consistent performance on Title Case ads, indicating that these are in general being received consistently better than their Sentence Case counterparts.
However, if we look at top level data from these ads spanning a 6 week period – across all devices; we can see there was little overall difference in metrics. We have seen that on mobile Title Case has had a slightly higher overall CTR, this may be due to standing out on a smaller screen or due to being in a slightly higher ad position. When shown on Desktop & Tablet there is almost no difference between the two overall.
So That Settles It?
Well, not exactly, there are a number of different ways we could go about expanding on these test results. Through adding this test across other differing types of ad groups, we would be able to draw wider conclusions on the behaviour of Sentence Case vs. Title Case. This, coupled with a longer testing period would allow for a greater differentiations in the data, particularly if your accounts’ traffic levels are subject to seasonal trends.
It is tempting to say that letter casing makes no difference to ad performance based on these results however there are many external factors that prevent the drawing of such definitive conclusions from this test and this data is only really relevant to this client in their specific sector. However continued ad testing does help to inform us of our best messages, and this type of A/B testing – combined with varied message & landing page testing allows us to understand how our target audiences are perceiving our ads and without continued testing we could miss out on potentially exciting opportunities, and we don’t want that!