I meet a lot of affiliate marketers who are still making a giant mistake when it comes to picking the best creatives while optimizing. That is why I decided to quickly write up this guide to hopefully help You once and for all to never make that mistake.
Using this method in your media buying strategy is a must if you want to dominate the competition. Remember, most affiliates overlook this and pick off heart after seeing some good numbers in Voluum data.
Let’s do an exercise together
Which Top 2 Creatives Would You Pick and Why?
Here’s our data (a very small data sample, but an expensive one that cost $987 on a RON buy). We don’t want to run another $1000 and see what happens. Instead we want to use the data we have and optimize to try and recoup our initial investment… so… which ones would You pick?
Chances are you picked creative 11543684 because it had most conversions, and maybe creative 11543682 because it had highest CR/CV.
Or maybe you picked 11525974 because it had a high CTR, and relatively nice CR and CV.
Whichever you picked, it probably isn’t the best one.
So how do we know which is the best banner to pick?
We need to use a statistically significant calculator tool based on the Exact Binomial and Poisson Confidence Intervals.
WTF is that you might wonder, well its some high end math stuff, don’t worry You don’t have to study up on it unless you really want to.
Please just click here to open the tool.
Go to the “Binomial Confidence Intervals” section. Put the number of conversions it’s had into the ” Numerator (x):” box. (In our case 3 conversions)
Put the number of impressions it’s had into the “Denominator (N): ” box. (In our case 2557 visits)
Hit The Compute Button.
This will tell us the Conversion Rate range that our banner can yield. As you can see in this banners case the lowest CR it can get is 0.0002, while the potential highest is 0.0034. We must multiple by 100 to get the %. So 0.0002 * 100 = 0.02% which is horrible.
Let’s type it in our spreadsheet, then repeat this for all the other banners and find the lowest probably % based on statistics.
This is our calculation based on our very limited data set. If you have more data, the numbers are going to be much better and accurate but we don’t want to spend more money blindly, and want to optimize so we want to pick the best to continue with.
Based on statistics using the extremely limited data we have, the Top 2 banners to continue with are 11525974 and 11543682
So we’d pause the rest, and continue collecting data with these two and then redo the calculation after we have more conversions.
After we’d redo the stat calculation between the two banners and find the clear winner, the next step would be to make similar banners to that and various variations to try and push the CTR up and our click costs down.
There’s a super detailed Guide on this getting into more detail inside the members area on STM, check it out here