Multi-Touch Attribution

MTA is the most accurate attribution method when complete and accurate person-level impression data is available. When it’s not, MTA-RCT is a new variant that can measure inside the walled gardens.

How Multi-Touch Attribution Works

  1. Create a history for each individual person of media and marketing exposures and purchases made
  2. Attribute sales to activities most correlated with purchases
  3. For each purchase, assign partial credit to every media and marketing exposure that contributed to it
How Multi-Touch Attribution Works
Proven Analytics

Proven Analytics

Marketing Attribution uses attribution methods scientifically proven to perform better than the alternatives. The Digital Media Consortium Ross Link led while at Nielsen in 2013-2016 in partnership with Google, Facebook, Yahoo, Krux, and 11 major advertisers tested the accuracy of a variety of MMM and MTA models. Across 1.4 billion digital impressions and 45 billion TV impressions, the best measurement approach depended on data granularity and availability.

Scalable Solution

Our cloud-based Attribution Engine can handle whatever it needs to. Highly scalable and highly portable, it can process billions of impressions and can run on the Amazon cloud, Google cloud, Microsoft cloud, or even on-premise infrastructure (e.g. for Netezza data stores).
Scalable Solution
MMM Coverage vs. MTA Depth

MMM Coverage vs. MTA Depth

Marketing mix modeling (MMM) and multi-touch attribution (MTA) have pros and cons. MMM is generally able to cover virtually all sales channels, and all media and marketing investments. But it sometimes doesn’t have detailed enough data to measure very targeted, smaller-scale digital media. MTA has very detailed data, and can do better than MMM on very targeted media. But MTA generally can’t cover all sales channels (e.g. not cash transactions) or all media and marketing investments. If a marketing exposure can’t be linked to an individual or household, MTA can’t measure it, at least not in the way MTA was designed to.

Integrated MTA and MMM

For these reasons, most organizations that use MTA also use MMM, presenting opportunities for conflicting results. We use Bayesian priors, the established, proven statistical technique for combining information from multiple models, to ensure MTA and MMM tie together and provide consistent, actionable recommendations.
Integrated MTA and MMM
MTA-RCT Data Creation

MTA-RCT

Our new MTA-RCT model leverages randomized controlled trials to accurately measure ROI by individual digital campaign, even within walled gardens. Utilizing hundreds or thousands of randomized controlled trials, the model gets an accurate, unbiased read on each campaign, while controlling for the impact of other campaigns. No tagging is required. Privacy issues are minimized. Walled gardens can be measured, while controlling for ads within other walled gardens. MTA-RCT is currently in Beta test. Please let us know if you’re interested in participating in our Beta program.

Contact Us For More Information

Tools Team Contact
Information
Home Tools
Solutions
Marketing Mix Models Multi-Touch Attribution Forecast Tracking Customer Journey Analytics In-House Modeling Software
Company
Team Contact