Always be testing. Any digital marketer should be very familiar with this mantra.
- A/B Testing
- Ad Rotation
- Keyword Optimization
- I’m trying to run but my legs won’t work
- Multivariate Testing
Most of these are probably related.
But even if you can set up and run these tests in a semi-coherent state, you may not be getting the best results from your optimization efforts. I can’t count the number of times I’ve spoken to a client who was “already running some tests” within their account only to find out that they weren’t even quite sure what they were testing in the first place. That’s probably because I don’t keep a detailed list of such occurrences, but if I did I’m guessing the tally would be at least 10, and maybe as high as 14.
The Real Step 1
The problem usually occurs right at step one and that’s because most people’s step ones should actually be step threes or step fours. Step one is not writing different ads to test, nor is it setting up different ad groups so you can ensure even traffic distribution. Step one is deciding what you want to test. It is determining your Key Performance Indicator.
You should never be three months into an ad test and wondering which stats will make one ad stand out compared to another. Is this statement familiar? “Well, ‘Ad A’ has a better click-through rate, but as you can see here, ‘Ad B’ got much more on-site engagement.” Afternoons have been lost to such poor planning, but you can prevent this from happening again. You can determine your key KPI before you launch your test! Incredible perhaps, but it’s true.
Define Your KPI
Is CTR your main concern? Well then let it be known! Don’t spend hours debating the conversion rate of your ad if your test was purely designed to get a handle on what messaging your target audience likes the best. And the same is true for whatever test you decide to undertake. Knowing your KPI before you outline your test allows you to be very specific with your variable(s), and to be much more confident in your results. So please, for your sake and for the sake of anyone who may have to suffer through that testing meeting debating the statistical validity of a keyword test that doesn’t know if it prefers impressions or conversions: decide your KPI during step one, not step 15.