Series A scale-up: SaaS ROI & cash runway model

Where the money goes after a Series A

The failure mode after a Series A is rarely running out of customers. It is an organisation chart that outruns the revenue used to justify it. Eighteen engineers at a hundred and sixty-five thousand fully loaded is a payroll line that arrives every month before anyone opens a cloud console, and at a hundred and forty thousand of starting MRR that single line is well over one and a half times revenue. Total operating cost is close to double revenue in the first month. That is not a warning sign by itself — it is what a funded scale-up looks like on purpose — but it does mean the plan is a bet that growth closes the gap before the six million in the bank runs down, and break-even sits three years out.

Headcount is not one input among several here; it is the plan. Every engineer added beyond the eighteen pushes break-even out by roughly two months and takes something over a million dollars off the ending cash position. The model tolerates four more hires than it has. The fifth is the one that ends it: at twenty-three engineers the cash trough drops through zero in the fourth year, and at twenty-four the company simply does not make it to month sixty. Nothing else about the business changed in that comparison — same churn, same pricing, same acquisition efficiency. Six hires, made one at a time over eighteen months, each individually defensible, is the whole distance between a plan that works and one that does not.

Retention has a similarly narrow tolerance and gets far less scrutiny in the same meetings. At two point four percent monthly churn the model finishes with cash. Take churn to three and a half and it does not — the last year is spent below zero. One additional point of monthly churn, the kind of drift that shows up as a slightly worse quarter and gets attributed to seasonality, is the difference. In between, every tenth of a point pushes break-even out by several months. This is worth doing slowly on the control above, because the curve is not intuitive: the damage is not proportional to the change, it accelerates.

Two cost lines compound quietly in ways that founders consistently misjudge, and they are the opposite of the ones that get audited. Variable infrastructure at forty-eight dollars per thousand of MRR is invisible monthly and enormous cumulatively — over the five years it grows nearly fivefold while payroll grows about a quarter, because it is tied to revenue and payroll is only tied to inflation. Setting that figure to zero would add roughly a third to the ending cash balance, which makes it the largest single efficiency available on this profile. Per-seat licensing is the other, and it deserves the opposite conclusion: at two hundred and ten a seat across an organisation of roughly thirty-two seats, the entire licence bill is around two percent of five-year operating cost. It is the line most often taken to a cost-cutting meeting and the one least worth the meeting.

The unit economics on this profile are strong enough to be read sceptically. A high LTV:CAC alongside a fifty percent re-investment rate usually means the same thing: the company could profitably spend more on acquisition than it does. Push the re-investment rate from fifty to ninety and the ending ARR more than doubles, at a cost of roughly a third of the ending cash. That is not obviously the right trade — it depends entirely on whether the next round is priced on revenue or on runway — but it is a real option that is being declined by default rather than by decision. Under-investment is a genuine cost. It just never appears on a statement, so it never gets defended in a board meeting.

Read the whole thing as a timing problem rather than a viability problem. The economics work; the question is whether they work fast enough. The cash trough lands roughly a year before break-even, and every hire made in the meantime moves both dates in the wrong direction at once. A Series A company that wants to stay in control of its own next round should decide, explicitly and in advance, what the trough is allowed to be — and then hire to that number rather than to the hiring plan the round was raised against.

What to move, and what happens when you do

Engineering headcount

The plan has room for four more engineers and no more. Add a fifth and the cash trough goes through zero in the fourth year. Set this before you touch anything else; every other control is being read against it.

Monthly revenue churn

The tolerance is roughly one additional point. Move it in tenths rather than whole numbers to see how quickly break-even slides once churn passes three percent.

Cloud cost per $1k MRR

The fastest-growing line in the model, because it scales with revenue rather than with time. Taking it to zero would add about a third to the ending cash — the largest efficiency on the page, and the one nobody schedules a review for.

Re-investment rate

Raising it from fifty to ninety more than doubles ending ARR and costs roughly a third of the ending cash. Worth running once to see the size of the option you are currently declining.

SaaS cost / seat / mo

Included for calibration. The whole licence bill is around two percent of five-year operating cost. Halving it barely registers against a single additional hire.

Series A scale-up: five years at the default inputs