A subscription wellness brand was losing $2.7M a year. Its founder had started to believe the company was failing. The raw data said otherwise.
We rebuilt the numbers around subscription math, held course through a year of broken attribution, and set the plan the company is now beating: cash positive on $9.5M revenue.
2024
The call
I first built this brand's financial model when it was small. That model ran its planning for the next two years. Then the business outgrew it, which is what happens to every model.
When the founder came back to me in 2024, everything was bigger, including the problems. Revenue was $7.21M and the company was losing $2.70M a year, a negative 37% EBIT margin. There was a board now, with questions he struggled to answer.
What struck me most was his mood. He had carried the story that the company might be failing for so long that he had started to believe it.
2024
What the data said
Before forming a view I went into the raw data: transaction exports, cohort tables, subscription behavior, acquisition cost by channel, fulfillment invoices.
The picture did not match the mood. Under the ugly EBIT line, the machine was working. Cohorts were retaining and subscriber economics were sound. On corrected projections the company had a clear path to cash positive. So my first deliverable was unusual for a turnaround: a data-backed case that the doom was wrong.
The same dig found something smaller. The 3PL's invoices understated international shipping. Everyone believed a Canadian order cost under $45 to ship. The real figure was closer to $75, and Canada was one of several countries affected. Nobody had lied. The cost was hidden in how the paperwork flowed, and every international decision rested on it.
2024
The reframe
The deeper problem was the lens. The brand sells a physical product online, so everyone measured it like e-commerce: revenue, ROAS, blended margin, month over month. Most of its revenue came from subscribers. Structurally it was a subscription business, and those run on cohort retention, lifetime value and payback.
We rebuilt the company's KPIs around that. A blended LTV-to-CAC framework changed how the media budget was spent: once you know what a customer is worth over their life, you know what you can pay for one. CAC came down about 25%. The loyalty program and the money-back guarantee both came out of the cohort math.
The founder had a board expecting answers. We built the story together, and I sat beside him in the board meeting with corrected projections and the reasoning under every line. He could now argue his own numbers.
2025
The year the dashboards lied
In 2025 new privacy rules on ad targeting hit the brand's category. The platform stopped reporting conversions, targeting weakened and attribution broke. The analytics stack reported a paid-social CAC of $840, three times the usual $280. Taken at face value, the channel was dead.
Post-purchase surveys said customers were still finding the brand there. My read: the channel was wounded and working, and the measurement was what had broken. Blended CAC, total spend over new customers, averaged $440 through the restrictions. Painful, and a very different decision from $840.
We tested the read. I brought in a specialist agency to measure the channel's true CAC with marketing mix modeling. Meanwhile we cut spend we couldn't measure, replaced a media buyer who wasn't performing, and I backed the founder's push into Amazon.
The restrictions lifted in Q3. CAC averaged $282 in the testing window and reached the low $200s by December, better than before the crisis. We scaled spend back up.
The books show the year plainly. Revenue dipped 11% to $6.39M because we pulled spend on purpose. The EBIT loss shrank from $2.70M to $0.68M: three quarters of the loss, cut during a measurement crisis.
2026
The scoreboard
The 2026 plan: $9.48M revenue and $0.99M EBIT, cash positive at a 10% margin. Six months in, the company is at 95% of the revenue target and beating the profit target, because CAC keeps landing under plan.
My current argument with the company is one you might not expect from finance. They are underspending on marketing, and I'm pushing them to spend more. Every well-priced customer acquired today compounds for years. Knowing when the numbers say yes is as much the job as saying no.
The takeaway
Are the numbers true?
The turnaround came from better numbers. We replaced figures that had drifted from reality, then managed to them through a board meeting, a platform crisis and months of wrong dashboards.
Every engagement starts the same way: in the raw data, asking whether the numbers we steer by are true.
Let's compare notes on your numbers.
Bring your model, your last P&L, or the decision that's keeping you up. You'll leave knowing what I'd look at first.