Analytic Partners builds experimentation tool into GPS Enterprise platform
Paul Sinkinson: "Experimentation is crucial where historical data doesn’t exist yet. It lets organisations test strategies in-market and build evidence before rolling them out more broadly."

Analytic Partners has integrated its Agile Learning experimentation solution, which includes built-in decontamination, into its GPS Enterprise platform.
The company said data from its ROI Genome shows more than two-thirds of marketing experiments are contaminated, producing misleading results that lead to poor investment decisions.
Agile Learning was previously a standalone tool. It lets businesses test strategies in-market, understand what is driving results and reduce risk before committing at scale. Inside GPS Enterprise, businesses can run tests when they don’t have the time or data to wait for a model, then feed decontaminated results into their marketing mix models and scenario planning.
The solution identifies other factors that change during an experiment, such as promotions, stock shortages, weather, competitor activity and other media, and removes them, so the reported lift is the lift the activity caused. Where no comparable market exists, it builds a synthetic control.
Analytic Partners said in one go-dark test for a large retailer, decontamination uncovered $40 million in potential lost revenue that the raw result had missed.
The platform combines data from across the business with Analytic Partners’ proprietary intelligence layer, ROI Genome, which draws on more than 25 years of global performance benchmarks. Organisations can compare scenarios, forecast likely outcomes and test strategies through integrated experiments before making larger investments.
Paul Sinkinson, Managing Director of Australia and Asia at Analytic Partners, said: "Experiments have become marketing's favourite answer to a tough year, but most can't see what's distorting their results. A promotion, a stock-out or a competitor's push can all change the outcome. Every test gives you an answer, but only a decontaminated one tells you what to do.

The leading media trade publication in Australia.
Get our top stories straight to your inbox daily by signing up to our Newsletter
By providing your information, you agree to our Terms of Use and our Privacy Policy. We use vendors that may also process your information to help provide our services.
"Test results are increasingly being fed into marketing mix models. A contaminated test doesn't just give one wrong answer; it steals credit from other channels and hides the real reason sales moved. That's why our experiments and our models run on the same data, in one platform.
“Experimentation is crucial where historical data doesn’t exist yet. It lets organisations test strategies in-market and build evidence before rolling them out more broadly.”
More from Mediaweek

The leading media trade publication in Australia.
Get our top stories straight to your inbox daily by signing up to our Newsletter
By providing your information, you agree to our Terms of Use and our Privacy Policy. We use vendors that may also process your information to help provide our services.





