From Platform ROAS to Profit Truth: How M-Squared Rebuilt TeePublic’s Incrementality Playbook
CMO, Marketplaces (TeePublic & Redbubble)
Founder/CEO at M-Squared.club
When we started working with TeePublic, the business wasn’t short on marketing data. What it was short on was something far more valuable:
A single version of truth that Finance and Marketing could both operate on.
At the center of the tension was a familiar story.
Platform attribution—especially view-through—made Meta look like a hero. Internal UTM / last-click reporting made it look… not great. And if you’re running the company on contribution margin, that disagreement isn’t just annoying. It blocks decisions.
"You could almost think of it as finance versus marketing… and really, we're all one team. So how do we come to center ground truth?"
This case study is about how we did exactly that—by moving TeePublic from platform ROAS debates to incrementality and profit truth.
The Brand
TeePublic is a creator marketplace where artists upload designs and the platform turns those designs into hundreds of physical products. It sits alongside Redbubble under the Audactor umbrella.
Go-to-market leaned heavily on paid media:
- Meta (Facebook)
- Google PMAX
The challenge: the business needed to know whether those dollars were actually creating incremental profit—or just being credited for sales that would’ve happened anyway.
The Challenge
TeePublic’s measurement stack produced conflicting answers:
- Platform ROAS (view + click) said: “we’re crushing it.”
- Internal UTM / last click said: “we’re not.”
Sam summarized the real issue perfectly: it wasn’t just measurement—this was a Finance vs Marketing operating conflict.
"CFOs are overly calibrated on last-click ROAS… if you show them platform ROAS it becomes contention."
So the question we took on was simple and sharp:
What is the true incrementality of each channel—and how do we translate that into reporting Finance will believe?
Our Approach
1) We rebuilt the business into decision-grade P&Ls
Instead of judging “marketing performance” as a blended number, we separated TeePublic into:
- New customer P&L
- Returning customer P&L
Key fact:
-
TeePublic was roughly 50/50 new vs returning revenue
That matters because incrementality behaves differently across those two curves.
2) We ran incrementality tests where the risk was highest
We ran incrementality tests on:
- PMAX
- Meta
3) We turned the read into operating rules (not just slides)
The test result is the beginning. The value comes from what you do next: reallocation, optimization, and defining the efficient frontier.
Key Finding #1: The ROAS “Truth Gap” Was Real
Platform ROAS and UTM ROAS weren’t “two perspectives.”
They were two different realities—and the business needed a third: incremental truth.
Key Finding #2: PMAX Was Genuinely Incremental (and Worth Defending)
PMAX came back strong:
- Strong lift
- Strong contribution impact
- Multipliers were in a “reasonable” range
- At or above break-even contribution expectations
"Search is identifying users raising their hands… you want to capture that demand or a competitor will."
One nuance mattered: PMAX already had meaningful investment, so the conversation became less about “turn it up” and more about how much more scale is realistic before diminishing returns.
Key Finding #3: Meta Had Lift—But Spend Was Past the Efficient Frontier
The Meta read was uncomfortable but useful:
- The lift wasn’t strong enough to be contribution-positive at current spend levels
- The implication wasn’t “Meta is bad”—it was “Meta is only good up to a point.”
"I was very skeptical… Facebook was a strong acquisition channel in my past experiences. It changed the yardstick of success… and we realized we were past that efficient frontier."
But then the key shift happened: we changed the yardstick. “It changed the yardstick of success… and we realized we were past that efficient frontier.” They didn’t just slash spend. They pulled back strategically until efficiency improved—as you’d expect in an auction system.
Key Finding #4: CTV Exposed Vendor Truth vs Profit Truth
CTV looked like an obvious place to reallocate dollars—especially because earlier vendor reporting showed unbelievable numbers. One vendor reported something like an “18 ROAS” type outcome (i.e., not credible), so we tested it properly.
The first CTV test came back with:
- No incremental lift
- And a major red flag: after spend stopped, conversions kept “showing up” in reporting for weeks.
"The campaign continued to report conversions for another month and a half even though there was no spend."
A second agency test showed:
- Green shoots
- Still below break-even
- But close enough to justify an optimization roadmap
Sam’s stance was practical:
- You need places to put dollars if you’re going to grow
- Early tests won’t be perfect
- If there’s lift, you improve the levers instead of quitting
Three levers we identified to move CTV toward profitability:
- Creative consistency + DRTV fundamentals
- Test-geo CPM inflation + national CPM pro-forma
- Inventory control to remove low-quality delivery
Outcome: Profit Management With Less Debate
After M-Squared, TeePublic didn’t just get “better measurement.” They got operating clarity:
- PMAX became a finance-defensible, incremental channel to defend
- Meta spend moved back onto the efficient frontier
- CTV shifted from vendor narratives to a real test-and-optimize pathway
- Reporting moved from “ROAS arguments” to incrementality → contribution → scaling rules
