Integrated measurement approach.

The challenge

A major U.S. retailer already measured its marketing with care. Its analytics team ran randomized controlled trials (RCTs) to isolate the causal impact of specific campaigns, and those tests were widely used and trusted. But each method answered its own question. The retailer wanted them to answer together. It set out to find one framework that brought all of its measurement methods into a single system, with higher marketing ROI and better implementation efficiency, while every piece of proprietary data stayed inside the company. This case study describes how that framework was built, and what each of the three partners did.

Good tests, but no unified marketing measurement

The RCTs gave clear answers about specific campaigns. What was missing was a link between those tests and marketing mix modeling, which measures channel effectiveness across the whole mix. Without that link, each method stayed a separate answer. The retailer asked MASS Analytics to work with its in-house analytics team and add Marketing Mix Modeling knowledge and capability to the RCT techniques the team already used, not to replace them.

Four things the retailer needed from the framework

• Tests and models had to work together. RCT results were tactical and widely used. The retailer needed them to enter the models as test results, not sit beside them.
• Higher return across the whole framework. The aim was better marketing ROI from every measurement method used together, not from one method at a time.
• Data had to stay inside the company. No internal data could be shared with an outside partner, so the retailer’s own team had to collect the data and build the models.
• The team had to be able to run it. The in-house analysts needed the know-how, not only the output, so training and support had to continue through the whole project.

The framework brought three parties together, each with a clear job. The table below sets out who did what.

The solution

MASS Analytics proposed unified marketing measurement: one framework that brought several measurement approaches together, instead of replacing any of them. The method was to build regression models that took RCT results in as tactical test results. Member-level test results then helped shape the marketing mix model results, by triangulating the upper and lower bounds of the coefficients. Other complementary methods were also considered for what they said about media effectiveness.

Three layers of support kept the team independent

The engagement had three layers. Technology transfer gave the team the software it needed to run projects in-house. Knowledge transfer meant continuous training sessions, so the analysts could run the software and add MMM know-how to the analytics skills they already had. Continuous consultancy meant frequent communication throughout the project, with MASS Analytics acting as a sounding board. Technology transfer matters, but it was only part of the story. Empowering the retailer’s analysts through knowledge transfer was crucial to the success of the project.

Four agile phases behind the unified marketing measurement

The project ran as an agile, iterative MMM process in four phases that could overlap. Data collection and understanding identified what data the project needed. Data processing prepared it to suit the nature of the retailer’s business. Model building ran with weekly consulting from MASS Analytics, plus help with model verification and testing. Reporting and recommendations used MASS Analytics as a sounding board for final results, with help on dashboard and simulation tools.

Results and Impact

The partnership produced unified marketing measurement: one framework that carried RCT results, modeled results, and other methods together. Four outcomes mattered most, and none of them required the retailer to hand over its data.

Unified marketing measurement without sharing internal data

Data confidentiality was a goal from the start, and it held. The in-house team collected the proprietary data across channels and built the models itself, with continuous help from MASS Analytics. No internal data was shared with an external partner, and complete data protection was maintained. That also removed the need for confidentiality procedures with outside parties, which shortened the project delivery time.

Faster delivery and a team that can run it

Efficiency was the second gain. The agile workflow brought faster MMM turnaround, shorter refresh and update times, and cost savings. Knowledge transfer was the third. Internalizing the MMM function adds strategic, long-term value, because the retailer’s own team now holds the capability. The result is a unified marketing measurement framework worth normalizing across other industries.