Testing a new creative or landing page well is actually not easy at all. Make sure you don’t make these basic mistakes when testing. I’ve written down 4 problems that we often see when evaluating creative tests.
Problem 1: Lack of data
This is the situation when you deploy a test and even after a month you don’t have enough data to evaluate it.
What about that? At the beginning, you need to choose where you will test and calculate how long it will take you to have enough data for evaluation. And how much is actually enough data.
A similar problem is testing on a small sample.
An example would be rcs database an e-shop with 2000 conversions per month and you give a recommendation tested on 20 conversions. I know why this problem arises – people don’t want to deploy a test on larger campaigns so as not to spoil the performance. But you can test gradually. First test on a smaller sample and when the test works, you roll it out further.
Problem 2: It is evaluated against the wrong metric
Typical situation – a PPC specialist enthusiastically reports positive test results, our CTR increased by 30%. Then we look at the campaigns thriving industries or sectors for foreign together and find that the conversion rate has dropped by half.
What about that? I recommend evaluating creative changes by CTR, but always look at conversion rate as well. If you’re testing a landing page or different offers, evaluate by conversion rate.
Creative change is not significant
You AB test banners where b2C phone list you have changed only a small text that is almost invisible. Or the background color – beige vs. old pink.
What about that? Just think in advance whether the changed element is significant enough.