Dose #203: The Ad Buying Math Most Subscription Brands Get Wrong

Matt here with your weekly Subscription Prescription 💊

Ad-buy is a must for any growing brand, but many metrics will lead you astray on growth. Last week we talked about Gruns and it’s 3:1 LTV to CAC ratio. This week we sit down with Trevor Crump to go deeper into why that makes more sense than ROAS-driven growth.

This week’s newsletter is also a full podcast interview, so you can listen or read at your convenience:

Matt here with your weekly Subscription Prescription 💊

This week I sat down with Trevor Crump, co-founder of Bestie Media, to talk about paid acquisition for subscription brands. Trevor has been in the e-commerce space since 2016 and has worked with brands at every stage, from pre-revenue launches to $350M companies.

The conversation went deep fast. And the core theme kept coming back to the same thing: most subscription brands are managing their ad spend by looking at the wrong number.

Here is what I took away.

Lesson 1: Your ROAS target might be strangling you

ROAS, or return on ad spend, has been the default efficiency metric for e-commerce advertisers for years. If you're at a 4x, you're doing well. Below 2x, something is wrong. That's the conventional thinking.

The problem is that ROAS is a blunt instrument. And for subscription brands specifically, it misses the most important variable: what a customer is actually worth over time.

Trevor walked through a concept he calls the Meta Optimized Market. Here's how it works.

Your total addressable market is the full pool of people who could theoretically buy your product. But Meta isn't going to show your ads to all of them. The platform narrows that pool based on your efficiency constraints. The higher your required ROAS, the smaller the window Meta operates in.

If you need a 4x ROAS and your competitor is willing to operate at 2x, they can reach twice the audience you can. They win more auctions. They scale faster. You stay constrained.

Grüns is a perfect example of this. Their founder has been public about maintaining a 3:1 LTV to CAC ratio, and pulling back when it dips below that threshold. That's a profitability discipline, not a ROAS discipline. There's a meaningful difference.

Takeaway: If you have strong subscriber retention and a known LTV, the question isn't "what's my ROAS?" It's "what can I actually afford to spend to acquire a customer profitably?" Those two questions lead to very different spending decisions.

Lesson 2: Do the math before you set the constraint

Trevor's framework for building an acquisition model starts with three questions.

First, what are your actual costs? Not just COGS, but shipping, fees, and the variable costs that come with every order. If your product retails for $100 and your variable cost is 30%, you're working with $70 before you even think about marketing spend.

Second, what is your retention rate month over month? If you're losing 12% of subscribers each month, you can map out exactly what a cohort looks like over time. Month one they're worth $70. Month two, factoring in churn, that drops. But the subscribers who stay keep contributing. You can model when you get paid back and what the total value looks like at 3, 6, and 12 months.

Third, what is your cash flow situation? This one catches brands off guard. You might be technically profitable on your acquisition math and still be cash-strapped because you had to write a $100,000 purchase order that hit before the revenue came in. The payback period model only works if you have the runway to survive the gap.

Trevor shared an example of a food subscription brand they started working with that had strong unit economics: $200+ AOV, $30 CAC, solid retention. They were barely spending on ads and growing slowly because they hadn't done the math on what they could actually afford to spend. When Trevor's team modeled out the LTV and showed them that a $150 CAC would still yield strong profitability by month two, the decision became obvious. They nearly tripled spend. The first 30 days were uncomfortable. By month 2.2, they were more profitable than they had ever been.

Takeaway: Build the model before you set the constraint. Understand your costs, your churn rate, your payback period, and your cash position. If you've never done this, even a rough version in a spreadsheet will change how you make ad buying decisions.

Lesson 3: Discount customers are not the same as real customers

This one is not new, but it came up in a way that I think is worth repeating.

Trevor has worked with enough brands to see a clear pattern: customers acquired through percentage-off discounts tend to be lower value over time. They came in for the deal. They're more likely to leave when the deal ends.

The shift he recommends is moving from straight discounts to value-based offers. Instead of 25% off, try a buy-one-get-one-50%-off structure. The effective discount is similar, but the customer who takes it is spending more upfront. They're more committed. They see more value in what they're buying. And that tends to follow them through the relationship.

He was on a call the day before we recorded with a brand doing over $40 million a year. His recommendation to them was to never run a percentage-off offer again, including over BFCM. Three years of data told him the numbers just don't work. That's a strong position to take, but when you've seen the LTV data across enough cohorts, it stops being a guess.

Oats Overnight is a brand that has taken a similar stance. They pull back on discounts during peak sale periods and focus on value instead. The customers they acquire that way are simply worth more.

Takeaway: Audit your recent acquisition offers and look at the LTV of customers who came in through discounts versus those who came in at full price or through value-bundle structures. The data usually tells you what to do.

Bottom line

There's a version of paid acquisition that looks disciplined but is actually limiting. You hit your ROAS target. Your CPCs look good. But the brand isn't growing the way it should and you're not sure why.

The answer is usually that you're optimizing for the wrong thing. The number that matters for subscription brands is not ROAS. It's what a subscriber is worth, and how much you can afford to spend to get one while still running a profitable business.

Do the math. Build the model. Then let the model guide the constraint, not the other way around.

Until next Tuesday, that’s your Subscription Prescription. 💊

 - Matt Holman 🩺

The Subscription Doc