Friday, Jan 28, 2022

Data Strategies that Make Multi-Touch Attribution (MTA) Work for Your Organization

The former chairman of the largest media agency in the world once said at a conference I attended, “Marketing has to look a lot more like digital… it ..

title
Abstract image of intersecting blue curves

The former chairman of the largest media agency in the world once said at a conference I attended, “Marketing has to look a lot more like digital… it can’t go the other way!” Such a simple and brilliant insight. But what does a digital world mean for those involved in modeling advertising effectiveness? I believe it means that MTA (Multi-Touch Attribution) is an integral part of the marketing future.

Marketing Mix Modeling (MMM), the big regression models of aggregated marketing data versus sales, is still in broad use for media planning. Yes, these models have value, but they do not answer all of the progressive marketer’s questions. MMM is backward-looking (using over three years of historical data, usually). MMM is not very granular; it uses media channels and promotion types as causal variables but doesn’t get to the level of this publisher as opposed to that one, six-second video vs. 15 second, etc. Typically, the consumer is literally not in the equation and so MMM offers no insight into targeting strategies.

Related

Plan B for When Models Fail

Enter MTA. It ingests user-level data as exhaust of digital and addressable marketing where ad serving is linked at a user level to conversion outcomes. MTA is forward-looking, analyzing conversions as they happen – as ad impression serving and exposure unfold. MTA is extremely granular. You can have millions of data points (IDs) that are exposed to advertising in a wide variety of ways that can all be represented as variables. The workhorse of MTA is logistic regression (using maximum likelihood approaches) that is happy to model dozens – and even hundreds – of independent variables.

According to surveys conducted by the MMA – as a consultant, I act as that trade association’s MTA expert and advanced analytics lead – MTA has been adopted by about 40% of marketers. Now, if MTA provides great advantages, why isn’t the adoption rate higher? Because complex data stitching is needed that makes MTA hard and complicated work. The data challenge is much harder than the analytics!

In order to ameliorate this burden, the MMA created the MTA DataMap (PDF available here) that I had a lead role in creating. This map gives comprehensive guidance on the kinds of marketing activities that are linkable, and what technology will work best to achieve the linkage.

Yes, MTA is hard work, but best-in-class marketers will make the commitment.

Header Image: Richard Horvath, Unsplash

The post Data Strategies That Will Make MTA (Multi-Touch Attribution) Work for Your Organization first appeared on GreenBook.

--------------------------------------

By: Joel Rubinson
Title: Data Strategies That Will Make MTA (Multi-Touch Attribution) Work for Your Organization
Sourced From: www.greenbook.org/mr/levelup-your-research/data-strategies-that-will-make-mta-multi-touch-attribution-work-for-your-organization/?utm_source=rss&utm_medium=rss&utm_campaign=data-strategies-that-will-make-mta-multi-touch-attribution-work-for-your-organization
Published Date: Fri, 19 Nov 2021 12:00:13 +0000

Read More