It’s becoming more and more common these days for brand-side marketers to ask about incrementality as each dollar of their marketing budget becomes more precious. And no channel gets this focus more than affiliate marketing (perhaps because lots of marketers still think of affiliate as the land of discounts and coupon codes).

Measuring affiliate incrementality in a cross-channel mix is neither simple nor linear, but undertaking the challenge is a smart strategic move that often leads to unexpected insights – and needle-moving optimization opportunities.

In this post, I’ll tackle how to measure affiliate incrementality, including:

• Tools and data sources to use
• The steps to take with the insights you get from the analysis
• Missteps to avoid in your incrementality measurement process

Let’s start with the data and analytics.

Tools, data sources, and caveats for affiliate incrementality measurement

Before I dive into the tools and data you’ll need to undertake incrementality measurement, a few notes.
First, you can’t truly study the incrementality of a channel in a vacuum; you need to analyze how one channel influences all conversions, no matter which channel captures the conversion event.

Second, when you’re looking at a range of time periods, it’s critical to consider seasonality. For instance, don’t include heavily promotional periods or holiday surges in one piece of the study and not another.

Third, it’s critical to remember that not all conversions are created equal. New-customer conversions are almost always more valuable than repeat-customer conversions, so set yourself up to analyze those segments separately.

With that understood, let’s look at your sources of data and how to use them. At DMi, we have access to:

• Affiliate platforms like Impact, AWIN, Partnerize, or CJ, for analyzing attribution paths and publisher-level performance
• Web analytics (GA 4, Adobe Analytics), for cross-channel assisted conversion and path-to-purchase analysis
• Third-party attribution platforms (e.g. TripleWhale, Northbeam, or Rockerbox) for modeled lift and cross-channel synergies.

You’ll notice that’s a good mix of platform and third-party data sources, which is important for cross-verification purposes.

No matter your tools and data sources, you’ll also need to plan for customized methods of analysis including holdout/lift testing (geo splits, audience splits, or suppression tests) to directly measure the lift affiliates provide compared with a control group. Truly advanced incrementality testing, like we offer at DMi, can also incorporate custom algorithms – for instance, weighting publisher activity by z-scores of new customer contribution or by diminishing returns curves.

The last thing I’ll say here is that running these models and referencing these data sources is an ongoing, iterative process that needs to start as soon as possible, not a one-time snapshot. Plan to run a full incrementality analysis at least quarterly so you’re not missing out on any major shifts. It’s a big uplift that you may want to outsource to an experienced partner; we’ve worked with a lot of brands who have dipped a toe in the incrementality measurement waters and quickly discovered they needed to bring in support to measure and take action effectively.

Speaking of taking action…

Expected takeaways/next steps from affiliate incrementality measurement

The two most common outcomes of affiliate incrementality measurement are reallocation at the publisher level and reallocation at the channel level in a brand’s holistic media mix.

Publisher-level adjustments often come down to strategic commissioning adjustments. Note that “reallocation” doesn’t mean “cutting”; for publishers that are under-indexing on incrementality, test commission decreases or structural overhauls before cutting the publisher altogether.

As for channel adjustments, incrementality tests that consider each channel in a portfolio usually show opportunities for more efficient allocation – for instance, bottom-funnel channels like brand search and remarketing are usually less incremental than click-based measurements tend to indicate.

Misconceptions to debunk before leaping to conclusions on incrementality

There are three schools of thought I often need to correct throughout the incrementality measurement process.

First, and I’ve touched on this already, but it’s worth repeating: it’s not the right move to simply cut publishers that measure as less incremental than others. For instance, a last- click coupon publisher that’s not driving “incrementality” may be critical for your conversion rate. There is nuance to the value of lower-funnel, conversion-assisting strategies that may not come through in incrementality testing.

Second, ROAS does not equal incrementality, especially when you’re referencing metrics reported directly from a platform. ROAS will look great on last-click, bottom-of-funnel publishers, whereas incrementality will take cross-channel lift and top-of-funnel influence into account.

Last, many marketers think of incrementality as a binary, yes/no decision. Publishers, publisher types, and marketing channels will all inevitably show some level of incrementality and some level of inefficiency – and, importantly, these levels will shift over time. The outcome of an incrementality analysis is not going to be “keep these publishers and channels running, and turn these off,” and the outcomes that result won’t necessarily last beyond the next analysis you run.

Wrapping up

Since I just threw a lot at you, it’s worth re-stating the benefits of affiliate incrementality measurement: you’ll be better prepared to invest your marketing dollars where they’ll have a more powerful impact.

The significance of smarter budget allocations, both within affiliate and considering affiliate within your greater channel mix, is bigger than ever given competitive and macroeconomic pressures. If you’d like to start a conversation about how to get traction, we’re always happy to chat.