Fractional CFO for subscription businesses

The decisions you make every month

Spend more on acquisition and growth looks better while cash gets tighter. Spend less and cash is safer while you miss the number.

Acquisition is the one you feel most. Engineering headcount is the same trade with a longer payback: a bet that a better product retains or converts enough to pay for the team. AI spend is that trade twice over: part of it is build, and part is an inference cost that sits in every active customer's margin for as long as they stay.

All of it is cash out now against revenue later, competing for the same cash. The numbers underneath are blended averages: one CAC, one LTV, one churn rate. Every CEO I work with knows they're rough. You set the dials anyway, because the alternative is not setting them.

And it's the same number underneath all of them

Almost every decision you make depends on the same thing: what your customer cohorts will be worth over time, and what they cost to acquire.

• How much to spend on acquisition, and in which channels.
• Who you can hire, and when.
• Which parts of the product are worth building next.
• When the cash runs out.
• How much to raise, when, and what story you tell doing it.

Get that one wrong and all five plans are wrong in the same direction at the same time. So it has to be a forecast, not a report. A cohort's real performance reaches your P&L three or four quarters after the decisions that depended on it, while your existing base carries the number and the new cohorts quietly underperform.

The number that would settle it belongs to no one

Your analytics stack reports attribution and ROAS. Your accountant reports the P&L. Whether the customers you acquired in March will pay back sits between the two, and it isn't anyone's job. So the question gets settled by whoever is most confident in the room.

What I do

Hi, I'm Trey Pruitt. I'm a fractional CFO for CEOs of $5M+ DTC, mobile app, and PLG SaaS subscription businesses. I build the forecast those decisions run on: the unit economics of your cohorts, modeled forward. Then I run it with you every cycle, so the calls you make have a number underneath them.

Three things come out of it first.

Where to set the dials

What the marginal dollar of acquisition spend actually returns, and when, by channel and by cohort rather than blended. You'll know within a quarter of a cohort landing whether it's paying back, while there's still time to act on it and before you commit the next quarter's budget.

The same model answers the ones next to it. How much retention or conversion lift an engineering team has to produce to pay for itself, and by when. Whether your AI cost per customer is growing faster than what those customers are worth. Same assumptions, one instrument, so the three can be weighed against each other instead of argued separately.

When the acquisition dial stops working

Paid channels get more expensive as you scale them, and the customers you buy at the higher price tend to retain worse. You pay more and get less, at the same time. That puts an end date on the current acquisition model.

You should see it three or four quarters out, with a ranked list of what moves it — retention, conversion, pricing and plan mix, a new channel — and how long each takes to land. Every one of those fixes is slow, which is why the warning is worth more than the diagnosis.

What your best customers actually cost you

If your product runs on AI, inference cost scales with how much a customer uses you. Which means your best-retaining, most-engaged customers can be your most expensive to serve. A cohort that looks excellent on revenue retention can look considerably worse on contribution.

Blended gross margin hides that completely, and so does treating AI spend as an operating line rather than a cost that belongs to specific customers. It changes real decisions: which plans to push, which segments to acquire, what to charge heavy users, and whether an engagement metric your team is optimizing is making you money or costing you money.

What that looks like in practice

• A cohort-driven model of your subscription business — acquisition, retention, expansion, churn — built on your own transaction-level data.
• Marketing spend connected to LTV and to cash, so scenarios come out as payback, lifetime value, revenue retention, EBITDA, cash impact, and runway.
• The model run with you each cycle: forecast refresh, scenario generation, gap-to-plan on what you're currently running, prep for board calls and fundraise conversations.
• A monthly read-out to your leadership team on performance against plan and what's changed in the forecast, so everyone is working from the same numbers.
• The rest of the CFO work alongside it: close oversight, board and investor reporting, cash and runway, the annual plan.

Who this is for

This works when all of the following are true.

• You're the CEO of a DTC, mobile app, or PLG SaaS subscription business.
• Revenue is $5M+ and growing.
• You sell monthly plans, published on your site or in your app. Annual as a discounted option is normal and fine.
• Paid acquisition is a meaningful part of how you grow.
• You have 18+ months of subscription history for the model to learn from.
• Someone already closes your books — a controller, a bookkeeper, or an outsourced accounting firm. I work above that function, not in place of it.
• You're the one holding the forecast today. There's no full-time VP Finance or CFO who already owns it.
• We work remotely, with occasional travel for board meetings and planning sessions.

And who it isn't for

• Under $5M in revenue. The data is usually too thin for the model to earn its keep.
• Annual contracts sold by a sales team. Reading a cohort early depends on monthly renewal decisions. Without them, there's nothing to read.
• One-time purchase, services, or transactional businesses. No cohort dynamics to model.
• No one closing the books yet. I direct and review the close; I don't prepare it. If that function doesn't exist, the right first hire is an accounting firm, and I'm glad to point you toward one.
• Wanting the CFO seat without the cohort work. The model isn't an add-on here. It's the reason to hire me rather than someone else. If you don't want it, what you want is a generalist fractional CFO. There are good ones.

Why me

20+ years as an in-house and fractional CFO in DTC and subscription businesses, across 30+ engagements in consumer subscription, mobile, and SaaS. $50M+ raised for clients in equity and debt, and two SaaS companies sold.

Calling how a cohort will pay back three months in, before it matures, is calibrated pattern recognition. The calibration comes from watching cohorts mature across dozens of businesses. The best analyst in your company has seen one company's cohorts.

So the forecast that nearly every decision depends on gets built without that read. And no growth team forecasts its own ceiling.

The other half is the instrument. The model isn't built from scratch on your dime. It runs on subscription forecasting software I've built and refined across these engagements, fitted to your business rather than generic. That's why it's running six weeks after your data is in hand, instead of the two quarters a platform implementation takes, and why the depth is there at all. Here's how the model works, and I'll show you it running on the discovery call.

That's the seat I fill, and I'm in it with you every month.

More about me