How to split test your ads in AdWords

Split testing your ads is one of the easiest ways to improve the performance of your campaigns. Find out how to A/B test your ad text in less than 2 minutes.

1. Adjust your campaign settings

First click the ‘Settings’ tab from your AdWords campaign. Scroll down the page till your reach the ‘Ad rotation’ option under the ‘Advanced Settings’ heading. Your ad rotation selection determines how AdWords will serve your ads within each of your ad groups.

Set ‘Ad rotation’ to ‘Rotate’ to show ads within each of your ad groups more evenly. This will give any new ads you create the same chance of appearing in the ad auction as your existing ads. If you leave ‘Ad Rotation’ on Optimise, the default setting, AdWords will show ads with higher Click Through Rates and Quality Scores more often than other ads in each of your ad groups. This can work wonders when your campaign is optimised, but will prevent you from running effective A/B tests on your ads.

2. Select an ad group

Next, click the ‘Ad Groups’ tab and select an ad group to create your ad in. Whenever possible, choose an ad group already receiving a large number of impressions. This will help you draw conclusions faster.

3. Test at least two variations

You can test any number of variations on your ad to see which one performs best for your business. You should test at least 2 variations of your ad text and no more than 4 - the more variations you test, the longer you will have to wait before drawing conclusions. Test only for small differences in your ad headline, ad text or display URL. Keep your variations focussed on one element at a time.

4. Monitor your ad variations

Check your ad variations on a regular basis to see which one works best for your business.

• First focus your attention on the Click Through Rate (CTR). A high CTR means your ad is good at driving users through to your landing page. A strong CTR will also improve your Quality Score, and help bring your bids down over time.
• If you have setup Conversion Tracking, you should also monitor Conversion Rates. A high Conversion Rate indicates that your landing page is delivering on your ad’s promise. This should ultimately be your measure of success.

Click the screenshot below for a working example (the advertiser's name has been hidden). The only variation is in the ad headline. The first ad mentions a price point, while the second refers only to 'Latest Deals'. The second variation has out-performed the first on both grounds, with a higher CTR and Conversion Rate:

5. Choose your winning ad

I often get asked how many clicks to wait before ending a split test. The rule is that your tests will always be more reliable the longer you let them run.

• When comparing ads on the basis of the Click Through Rate, I generally build up at least 500 impressions before coming to a conclusion. However, my preference is for several thousand impressions where possible.
• When comparing ads on the basis of Conversion Rates, I focus my attention on the number of clicks each ad has received. If generally feel comfortable drawing conclusions once a page has received over 250 clicks. However, you may not necessarily need to wait for as many clicks if the difference in performance is large enough early on.

Once you've chosen your winning ad, think about another test you could run against it. Test, test and test!