Cody Plofker, CEO of cosmetics brand Jones Road Beauty, doesn’t care if his tests show that a particular channel isn’t working. In fact, he’s happy.
“You want to know something is going to go wrong,” he said. can Optimize. “
That’s not how most marketers are wired or how their tools work.
To them, measurements are typically a mashup of platform reporting, media mix modeling, MTA vendors, and the occasional geolift test, all of which claim to show: gradualness And no one completely agreed.
“There’s too much noise and disruption,” Plovker said. “So I keep thinking, do I trust this or do I trust that?”
That’s a good question.
Another way to look at MMM
A little more than a year ago, Jones Road Beauty began working with ad measurement startup House to gain a more consistent view of how its media performs across channels and over time.
Shortly thereafter, the company became an initial design partner for a new Haus product called Causal MMM. Causal MMM is a media mix model built on the results of real experiments, not just correlations.
Most incrementality measurements focus on narrow questions. In other words, if you change this one thing about your campaign, how much additional lift would you see compared to a control group that wasn’t exposed?
Causal MMM collects multiple test results in one place and uses them to form media mix models. House’s chief strategy officer, Olivia Colley, said that rather than relying on historical data or patterns of correlation as traditional MMMs do, it treats the brand’s own experimental lift at a given spend level as a point on a curve.
“Experiments address the problem of causality as the basis of MMM,” Colley said. “The results solidify the data in a way that we can no longer guess.”
mix shift
The great thing about Jones Road is that this approach is neatly integrated into the testing roadmap.
For example, Plovker’s team is already experimenting with Demand Gen. Demand Gen is a product similar to Google’s AI-powered PMax that serves video and display ads across YouTube, Discover, and Gmail. They tested at both $7,000 and $10,000 per day, so they already have a revenue curve for that channel.
When House’s model suggests increasing spending on YouTube to, say, $15,000 per day, that recommendation doesn’t come out of a black box. This is based on actual increases that Jones Road has seen at lower spending levels.
This is a departure from the way most MMM vendors operate, and is “very ‘just believe’,” Plovker says.
“You’ll hear voices saying, ‘If I change the mix this way, this is what I’ll get,’ but the reality is we don’t know for sure and there are too many variables,” he said.
Causal MMM makes the process more “scientific,” Plovker said, because the recommendations are simply the next experiment on the roadmap, rather than a blind bet.
Recalculate numbers
This same logic applies to Profker’s biggest channel, the meta.
For years, Cody thought he understood the impact meta had on performance. the core of his spendingBut he never actually ran any incremental tests on it. When he finally did, the results were solemn.
While Meta remained accretive, Jones Road found itself overspending and much of its revenue coming from repeat customers rather than new acquisitions. “We were spending beyond our efficiency,” Plofker said.
That’s why it’s important to have a test-and-retest culture, especially in large, high-spend channels, Cory said.
Jones Road ended up pulling some of its spending on Meta’s Advantage+ shopping campaign and redirecting more of its budget to mid-funnel campaigns. What started as a single test became a multi-month roadmap of tests that reshaped how brands thought about the role of meta in the mix.
Causal MMM also helps identify channels that should not get any budget on Jones Road. For example, Plovker tested Google Brand Search and found that it was not incremental.
“I was a little upset when I found out I was spending $3,000 a day on something that wasn’t driving business, but it’s good to know,” he said. “That’s $3,000 a day. I don’t have to spend it at all anymore. Otherwise, I could put it into something that would do better, like Demand Gen.”
reality check
Jones Road isn’t alone in grappling with the difficult questions of where to spend time and when to stop.
But Collie said it’s hard to know what to do when traditional MMMs and platforms tell a flattering story that they “know what marketers want to hear.” But those stories don’t always hold up to scrutiny.
On top of that, there’s what Kory calls “that” thing. multicollinearity In statistics, multicollinearity occurs when multiple independent variables are highly correlated, making it difficult to understand the influence of each independent variable.
In practical terms, this means that when brands increase spend, they often do so across multiple channels at the same time, during periods of increased business anyway, such as in the run-up to Black Friday. They believe their marketing has worked, but the increase in sales is really due to the season.
“Unfortunately, we often have to bring our brands back to reality,” Colley said.
But Plovker loves a good reality check.
“We always try to run live tests,” he said.
#Jones #Road #Beauty #type #MMM #reset #media #measurement #AdExchanger