Ad advertiser recently contacted us and asked a simple question: If I want to test two ads and two landing pages; is this how I should setup my ad test:
There’s not a simple answer to this question as it depends on what you want to learn.
For instance, if you want to learn if Ad A and LP1 (landing page 1) lead to more conversions than Ad B and LP2; then that’s a valid test.
If you want to learn which ad and landing page combination works best; then you need this setup:
Some people immediately dismiss the first test since you’re really just testing ads to landing page combinations; however, there are times that is very useful. Consider these scenarios:
In these cases, you are testing a message across the ads and the landing pages so using one and and one landing page against another ad with its own landing page is valid.
Of course, you can go even further to ensure that one message is much better than another one by varying the benefits of the action within the ad and then seeing which landing page is best across multiple benefits; while still promoting the main message within the ads and landing pages such as:
You can then look at the aggregate data to the landing page to see which ad message and landing page message works best for you.
There was a local service based company that wanted to know how much using the location in the ads would affect their metrics; so they tested:
In this test result, the local ad has a much higher CTR than the non-local ad. However, its conversion rates were much lower. The ad set the expectation of finding local information post-click and it wasn’t there; so users were bouncing off the page.
What they should have tested (and did later on) was 4 combinations:
In this case, the local ad to local landing page was a very clear CPI winner.
Of course, testing within one ad group only gives you information about that ad group. If you want to see how these tests work across many ad groups before you create a lot of local ads and landing pages; you should use multi-ad group testing.
If you want to test a consistent message across your ads and landing pages; then create one ad for each landing page (or two ads for each landing page; but don’t use the same ad for both pages).
If you want to see what combinations of ads and landing pages preform best; then you need to test each ad and landing page combination together to see accurate results.