5 min read
Incremental Measurement: The Next Evolution in Attribution
By Sam Malissa
Without attribution, you wouldn’t know which of your marketing dollars were working and which were just vanishing into the digital void. But attribution can be tough, since you’re trying to track activity across so many different channels, and it’s hard to say for sure which interaction was the one that nudged the customer from prospect to purchase.
That’s why some of the more sophisticated brands are using incremental measurement, a method that compares results from a test group and a control group in order to zero in on which marketing tactics are the most effective.
In our two previous posts on attribution, we covered how the more common methods of Multi-touch attribution (MTA) and marketing mix modeling (MMM) can be more accurate and effective with a strong data foundation and by incorporating customer lifetime value. This time we look at why MTA and MMM alone aren’t enough and the benefits of incremental measurement to take attribution to the next level.
The Standard Ways Tell You What’s Happening, But Not How Much It’s Worth
MMM and MTA each have their advantages and their limitations, leading many marketers to use them together to cover the gaps. MMM provides a high-level view of how a mix of marketing corresponds to sales over a longer period of time, which gives useful visibility into trends but is often too broad to give granular insights.
MTA tracks digital interactions across platforms and channels, which is critical in a world where so much happens online. But it leaves marketers to wonder how their efforts are performing in other offline areas, whether that’s in print, in person, or on the airwaves. And even information from online interactions can be limited, due both to restrictions on cross-site tracking cookies and the fact that many ad impressions are being served within Facebook and Google, which also have restrictions on tracking.
But the biggest missing piece is that neither MMM nor MTA give you a clear picture of how valuable any given step is in the customer’s path to purchase. They let us know that customer Bob made a purchase after interacting with touchpoints X, Y, and Z but we don’t know which of these did the most to get Bob across the finish line. Combining MMM and MTA gives marketers a decent sense of which media Bob is interacting with, but not necessarily which are the most effective.
Using Incremental Measurement to Find the Marketing that Works
This technique aims to answer whether a specific piece of marketing actually leads to value. You do this with good old fashioned A/B testing: set up a treatment group that sees the marketing and a holdout group that doesn’t, then measure this difference in conversions (or sales or signups or times you ring the gong in the call center).
That difference quantifies the value of the marketing. If you find that one touchpoint is much more valuable than the others, you can put more resources there and give the underperforming marketing elements a little vacation.
The principle can be extended beyond individual advertisements to entire channels. For example, we worked with a brand that wanted to measure the impact of non-branded paid search (which is very upper-funnel and notoriously difficult to assign credit). We helped them identify many pairs of lookalike markets throughout the US based on search volume, region, and competitive saturation. Then they ran a test where they turned off non-branded paid search in one of the two lookalike markets, and kept it going in the other. This let them isolate the impact of that investment.
In the course of A/B testing messages and channels, you’ll also start to identify segments of your audience who are more receptive to marketing, and others whose purchasing habits remain more or less the same whether they are marketed to or not. You can then focus your efforts on the receptive audiences and get more of that bang for your buck everyone is always talking about.
Three Steps to Powering Up Your Marketing Analytics
Following our previous posts on attribution, think of incremental measurement as the last step in a crawl-walk-run progression.
First you get your data foundation into shape so you’re only working with reliable customer data.
Then add predictive CLV when doing your marketing attribution, to get a better understanding of which tactics and channels work with the most valuable audiences.
Finally, incorporate incremental testing to find out what different elements of your marketing are doing the most work and focus on those in an upward spiral of improvement.
With all three of those in play, you’ll be well on your way to the top of the analytics game, a master at finding the precious signal amid all that pesky noise.
Want to talk more about strong foundations or evolved attribution? Get in touch.