Landing Page Testing
1. Decide what keywords you want to be involved in the test, and select these in the interface:
Above, I’ve chosen to select only brand keywords. This is a pretty simple example, but you could choose to select keywords in any way you want. Carrying out a test on only your top 10 keywords is another example.
2. Select "Start landing page test":
From here, you can choose whether you want your landing page test to be carried out on all devices, only on mobile, or just on desktop and tablet.
3. Fill in the details of the test:
Give the test a unique name with a clear naming convention. Specify the landing pages you want to send traffic to. You can choose to either split traffic evenly between these, or decide on a custom traffic split.
4. Click "Save", and your test is ready to go!
5. To see the results, go to the "Landing page tests" tab:
Here, you can see results split out by each landing page. You can customise this report to add the metrics that are most useful to you. It’s worth noting that you can’t see engine metrics such as cost and clicks here, however you can see visits - the DS version of clicks - which are tracked by the DS clicktracker.
Keep checking back here every so often until you’ve collected enough data and are ready to switch to the top performing landing page. A good way of seeing if the test has been running long enough is to check the “is result” column. If you see a green tick here, it means that the results are statistically significant.
6. Finally, click "Stop test", and choose which landing page you want as your keyword URL:
You can either chose one of the landing pages involved in the test, or revert each keyword URL to what it was before the start of the test.
Benefits To Landing Page Testing In DS Over AdWords
1. No need for drafts and experiments:
Testing landing pages in DS means you don’t have to mess around with campaign drafts. These can be a pain, especially if you need to make changes to the campaign during the course of the test, as you have to make the change in both the original and the duplicate campaign.
2. Testing across campaigns is possible:
You can easily test all generic keywords, for example, even if they are housed in separate campaigns.
3. Testing on certain devices is possible:
You can easily create a test for just mobile traffic, which is especially useful if you’re looking to optimise your activity for mobile users.
1. You will need to have either Floodlight or Google Analytics tracking set up:
No engine metrics will be shown in the Landing page tests tab, including clicks, costs, and conversions. While advertisers who do not have either Floodlight or GA conversion tracking set up will still be able to create these tests, you won’t be able to measure their performance by looking at conversion data.
2. You must have keyword-level URLs:
While DS now supports ad-level URLs, in order to carry out landing page testing, you must have your URLs set at keyword level.
3. You won't be able to make changes during the course of the test:
It isn’t possible to add or remove keywords from the test while it’s being carried out. Similarly, you won’t be able to edit keywords that are being used in the test. This is because any of these changes may skew the data. So, it’s worth double checking everything is correct before you start the test in the first place - otherwise you may need to set it up again.
Ad Copy Testing
1. Navigate to the ad group that contains the ads you want to test, and select the ads tab:
Bear in mind that you can’t specify the ads you want to test - any test you create will include all active ads in that ad group. You also can’t create tests across multiple ad groups.
2. Select "Create ad copy test":
Here, you can choose to test the ads on all devices or only on mobile.
3. Choose the ad rotation setting to apply once the test has ended:
In order for the test to be fair, the rotation setting will be changed to “rotate indefinitely”, but once it’s over, you’re probably going to want it to change back to optimise.
4. Create the ad copy test:
Give the test a unique name, then decide what your goal is - either CTR, revenue per visit, action rate, or transaction rate. Finally, decide how statistically significant you want the results to be (from 85% to 99%). We would recommend choosing a statistical significance of at least 95%.
The more statistically significant you want the results to be, the longer you’ll have to wait for the results. Exactly how long will depend on the level of traffic you typically receive.
5. Navigate to the "Ad copy tests" tab, then click on the test name to check results:
You can customise the report to view your desired metrics. Next to each ad, there will be a symbol displayed which denotes how the ad is performing compared to the other ads in the test. A grey +/- sign means the results are not statistically significant, a green arrow means the ad performed better and a red arrow means the ad performed worse.
“Observed value” depends on the goal you set. If your goal is CTR, this will be the average CTR of the ad in question (this will match the figure in the CTR column).
“Confidence interval” is the CTR (or other goal) range that the ad has achieved. Once the test is complete, depending on the statistical significance threshold, you can be X% certain that an ad will achieve a CTR within this range.
6. Select "Stop test" when enough data has been collected:
Here, you can opt to either pause ads that you have specified, have DS pause the poor performing ads, or keep all ads running.
Benefits To Ad Copy Testing In DS Over AdWords
1. The dedicated ad copy tests page makes it easier to determine the test results:
Having the symbols in DS means it’s really easy to quickly see which ad is performing best, or if the test needs longer to run. This tab also makes it easier to see results of previous ad copy tests, which are difficult to find in AdWords.
2. The statistical significance threshold takes the guesswork out of deciding how long to run the test for:
With DS, you not only know that results are statistically significant, you know how statistically significant they are, meaning you can make decisions with more confidence.
3. You can easily test ads on mobile only:
While this is possible on AdWords, ad copy testing on DS makes this much easier. With more and more searches taking place on mobile devices every year, this is particularly useful.
1. Only change one element of the ad per test:
Testing 3 completely different ads is fine, but you won’t have many learnings to take away from the test. Instead, it’s best to only change one element slightly, trying different synonyms of “buy now”, for example. This way, you know exactly why the top ad out-performed the others.
2. Don't make any ad changes during the test:
Any changes that you make to ads that are involved in a test will skew the test results. It’s worth bearing this in mind because the test will continue to run, even if an ad is edited.
3. The test doesn't start immediately:
Once you create an ad copy test, it won’t start to collect data until 12AM the following day.
Although these might seem like fairly basic tests to run, they can have a big impact on performance, and reveal useful insights about customer behaviour. Using DS to carry out these tests means you can set them up in a matter of minutes and see the results with a click of a button, giving you more time to spend on the complex stuff.