February 11, 2020

Incrementality Testing Is Not Another Buzzword: Discover the true impact of your affiliate program

Tom Rathbone

There is a lot of discussion in the performance marketing space these days about incrementality testing and its usefulness. We know that sifting through hype and buzzwords can be frustrating for brands and marketers who are just looking for sound advice — which is why we wanted to take an opportunity to weigh in on the conversation. We take a firm stand in support of incrementality testing and have seen it offer our clients’ performance marketing programs more agility and better chances for success. 

Indeed, incrementality testing is not just a buzzword, it’s a useful tool to help brands break through the status quo and realize the full potential of their performance marketing programs. And it’s also a tool that affiliates and brands alike can benefit from. 

What is incrementality testing and what’s the point of using it?

Incrementality testing, at a basic level, is essentially a form of A/B testing to figure out which campaigns or publishers are providing true incremental value to your business. Incrementality, as a reminder, refers to the value or sales generated by the marketing efforts that would not have otherwise been generated. So, if we say that an affiliate’s efforts were incremental to the business, we mean that without those efforts, we would not have gotten the sale. Likewise, if we say that a particular campaign within an affiliate program was truly incremental to the business, then we mean that the sales that resulted from a customer’s contact with that campaign would not have happened otherwise.

Running incrementality tests is important because one of the most common ways that brands lose profit is by paying for sales that they would have gotten anyway.

The reason to regularly use testing to optimize your campaigns and other marketing strategies is so that you can understand the full impact of your affiliate marketing program and the publishers therein. If you aren’t looking for ways to improve your program, and questioning things that aren’t working, you’re not likely to find new answers and new ways of doing things.

For so long, the industry has played by essentially the same set of rules. For example, the standard attribution model in performance marketing is last-click, which does not necessarily bring the kind or results or publisher behavior a brand might be looking for. 

Brands can either accept this as the way things are, or ask for it to be better and more agile. And for those who don’t accept the status quo in performance marketing and want results to continue to improve, incrementality testing is key. If you want to truly understand what’s valuable and what is not in your marketing efforts, you’ve got to test, try and test again. Once you change things up in your marketing and try something new, you’ll want to measure back against that to determine how successful those changes were.

How to test for incremental value

Like any good test, in order to get the most accurate results, you’ll need to isolate the variable you want to test while keeping others the same. Same goes for incrementality testing. Say you have 10 channels and you want to test the results that changing one variable may have on any one of those channels. In order to run an accurate incrementality test, you’d want to isolate one channel variable (the one you’re testing) and keep all others static. For example, you could send one promotion just to the affiliate channel but not to any other channels and see what kind of lift you receive (if any). Then for the next test, you could push another promotion through another channel but not to the affiliate and see what those results are. You’re looking for the impact that these activities have in isolation; this determines if the activity has incremental value.

As another example, say you wanted to test the results that doubling the commission rate of one particular publisher may yield. In order to get accurate results, you want to be sure that all other variables (such as other publisher commission rates and/or other channel variables) remain the same when you make this change. Once your data is in, you’ll be able to measure the impact that this change has against the performance of the rest of your marketing. This is how you understand if the change in that variable (in this case, doubling the commission of a particular publisher) was incremental or not.

The two most important things to remember in incrementality testing are:

  1. Do your best to keep all other variables that you are not testing for as static as possible so that the changes of the variable being testing are clear.
  2. Repeat incrementality tests again and again over time. You cannot rely on the results of one test, as it may be an outlier. Only with aggregated data from repeated tests over time can you determine incrementality with 100% confidence.

Using incrementality testing, you can determine which partners, campaigns and channels add true incremental value — the kind of value you want to pay for. This in turn also helps your brand attract the kind of partners that are best suited to providing that value. It starts with expanding your way of thinking about the affiliate channel and innovating accordingly. Incrementality testing is one of those innovative tools that can help move the needle forward in a big way.


Have questions for Tom about the benefits of incrementality testing? Contact us at any time. We look forward to chatting!