Joe Mandese, November 14, 2024

At a time when media inputs are becoming increasingly opaque, IPG Mediabrands Kinesso unit is doubling down on a scientific methodology utilizing high-speed experiments involving tens of thousands of tests to better understand which ones lead to better outcomes for clients, platforms, media and creative executions.

The hub of the new experimental model is aptly named the "Experimentation Lab," and is being led by Berna Tollefsen, a long-time analytics and experimentation exec at Citi, who joined Kinesso in May as vice president of experimentation.

"Experimentation is a scientific methodology to validate hypotheses," she explained in an interview on the eve of this morning's announcement, adding: "It can be as simple as testing different creatives, or it can bet a bit more complex, requiring us to test variations of multiple elements simultaneously."

While testing including variant (or A/B) tests of creative and media executions, as well as various forms of mix and attribution models has long been a practice on Madison Avenue, what makes Kinesso's new lab different is the discipline and comprehensiveness the new unit brings to the process -- specifically the ability to continuously field massive, large-scale experiments in near real-time, constantly testing hypotheses and reformulating them into best practices almost on-the-fly.

Kinesso first began utilizing that method shortly after its Mediabrands reboot and consolidation in 2023, especially its exploitation of algorithmic trading, which was part of the reason MediaPost named it Agency of the Year earlier this year.

But whereas algorithmic trading is focused on generating better marketplace outcomes for specific client campaigns and media buys, the new Experimentation Lab has a broader mission of identifying, testing and understanding the full spectrum of inputs and outputs that lead to better outcomes for the entire marketplace, and Interpublic clients overall.

In other words, it has a more meta view and purely scientific goal of generating insights, and learning that can be rapidly shared by Interpublic agencies, account groups and clients alike.

While that may sound abstract, theoretical and purely scientific, it is a resource enabling Interpublic agencies and clients to tap into a deeper knowledge set enabling faster adjustments to marketing and media strategies, and the ability to test new variants at hyper speed.

If that still sounds a bit abstract, Kinesso Global Head of Platforms & Intelligence Chris Schimkat gives a real-world example of how it can be applied.

"There's a particular industry where, within the first month of the year, it will spend a large majority of its budget," he says. "We have a hypothesis for our client that if you start spending in December, you're going to have way less competition and you're going to be top-of-mind by the time that January period rolls around. And you can actually spend less money to drive a higher result."

While Schimkat's example is a practical application, the overarching goal of the new Experimentation Lab is to harness the collective wisdom and insight from the tens of thousands of experiments being run across brands and industries to enable individuals to make more informed decisions.

Or as the lab's Tollefsen says, to "help marketers stay ahead of the curve."

https://www.mediapost.com/publications/article/401109/kinesso-ramps-up-scientific-model-opens-experime.html