Multi-Touch Attribution

What is multi-touch attribution?

A way to determine which marketing effort drove a customer to make a purchase or take a specific action. This information helps marketers and app developers optimize their marketing efforts and ensure their ad spend is going toward efforts that contribute most to their success.

Multi-touch attribution is often considered the Holy Grail of cross-platform mobile advertising attribution. To see why, it’s best to back up a step.

What is attribution?

Attribution is the process of giving credit to a particular source for an event or action. For example, if you click on a Facebook ad to install a cooking app, your download would be attributed to the Facebook ad. This attribution is important for a whole host of reasons; but the primary one from a marketer perspective is that marketers want to know how successful each of their channels are in convincing users/customers to take particular actions. That way, they can invest more heavily in high-performing channels, and work on improving or reducing spend on low-performing channels.

Unfortunately, though, a customer’s path to a particular action is rarely so simple or straightforward. Imagine the same example described above, in which you install a cooking app after clicking on a Facebook ad. It’s possible that this is the first time you ever saw or heard of this cooking app, and so that particular ad was the sole thing that caused you to download it. However, it’s far more likely that there were multiple touchpoints that led to you downloading the app. For instance, you might have seen a display ad for the app on a cooking blog, got sent a referral to the app by a friend, and then ultimately decided to download the app after seeing the Facebook advertisement. Here, which channel to give credit for is far less clear. After all, you might not have downloaded the app if you hadn’t received a referral link from your friend. Or perhaps you would have, but not if you hadn’t seen the blog display ad earlier. Regardless, it’s almost certainly the case that more influenced your download decision than simply the Facebook ad.

Benefits of multi-touch attribution

Multi-touch attribution aims to give marketers a clearer sense for all the touchpoints that led to a consumer action. In the example described above, a solution that provided multi-touch attribution would show that you had viewed a display ad on a cooking blog, received a referral link from your friend, and then ultimately downloaded the app upon clicking on a Facebook ad.

But even this isn’t enough, since it’s unclear to what extent each of these touchpoints impacted your ultimate downloading decision. With a large enough sample size of consumers downloading the cooking app after various marketing touches, it would be possible to run sophisticated regressions on the data to determine that, say, the blog display ad was 30% responsible for your decision, the referral link 20%, and the Facebook ad 50%. However, given how many potential touchpoints there are, marketers rarely have the luxury of gathering this much data; and even if they do, it rarely results in attribution this clear. Instead, they make use of heuristic models that assign weights to touchpoints based on intuitive understandings of their consumers.

For example, a marketer for a highly transactional product or service might believe that a consumer decision is influenced mostly by the last touch. In this case, they could assign most of the “attribution weight” to the final touchpoint before the consumer takes an action. Alternatively, a marketer for a big-ticket item might believe that driving consumer awareness is important, but so is that final touchpoint. In this case, they might opt for a “W-shaped” multi-touch attribution model, in which most of the “attribution weight” is assigned to the first and last touchpoints. There are myriad other multi-touch attribution models out there, including exponential decay, first touch, last touch, W-shaped, etc. But all of them serve the same goal of trying to understand consumer behavior in order to make optimal channel investment decisions.