Runway Isn't a Number, It's a Model

I have sat across from a founder who told me, with total confidence, that they had five months of runway. The number was correct the day it was calculated. It was wrong three weeks later, because a customer churned, a contractor invoice landed a month later than usual and then hit twice, and a hire that had been "maybe next quarter" got pulled forward because a competitor made an offer to the same candidate. None of those three things was unusual. Together they turned five months into ten weeks, and nobody noticed until the number needed updating and nobody had built the thing that updates it.

This is not a story about a bad founder. It is the default outcome of treating runway as a number instead of a model. A number is static. A model is a set of assumptions with a mechanism for updating them. Startups that get surprised by cash are almost always running the former and calling it the latter.

Why cash ÷ burn breaks immediately

The calculation itself is not wrong, exactly — it is just answering a narrower question than people think it is. Cash divided by average monthly burn tells you how long you would last if the next N months look exactly like the last one. That condition almost never holds, for three reasons that show up in nearly every company I have worked with.

First, burn is not a smooth line. Payroll is usually the steady part. Everything else — annual software renewals, a legal bill for the SAFE round, a marketing spend that got approved in a good month and never got revisited, a tax payment — arrives in clumps. A trailing average smooths those clumps into invisibility right up until one lands.

Second, revenue is not a constant either, and treating it as one hides the direction it is moving. A company with flat revenue and one growing at 8% a month can have identical burn and identical cash today, and wildly different runway in real terms, because one of them is closing the gap and the other is not.

Third — and this is the one that costs people the most — the trailing average has no opinion about decisions you have already made but not yet paid for. An accepted offer letter for a hire starting in six weeks is not in last month's burn. Neither is the marketing spend you approved for next quarter's campaign. A model built only on history is blind to every commitment sitting in your inbox.

What a runway model actually needs

A model that survives contact with reality has four components, and the difference from a spreadsheet with one cell dividing two numbers is mostly about which of these four it includes.

1. A cash bridge, not a cash balance

Start from today's actual bank balance, not last month's close. Then walk forward month by month adding known inflows and known outflows as line items, not as an average. Payroll by headcount and start date. Contracted vendor payments by actual due date, including the annual ones that land once a year and get forgotten the other eleven months. Collections by expected payment date, not invoice date, if your customers do not pay net-zero.

2. A distinction between committed and discretionary spend

Split every line into what happens regardless of what you decide next, and what you could still choose to change. Payroll for people already employed is committed. A hire you have not made an offer for yet is discretionary. This split is the entire point, because it tells you which levers actually exist if the number gets uncomfortable. Founders who cannot answer "what could I cut this month if I had to" in under a minute do not have this split, and that is the same gap that shows up as a five-month runway becoming ten weeks.

3. Revenue as a trend, not a trailing figure

Model the trajectory — the actual month-over-month growth or decline rate, applied forward — rather than assuming next month equals this month. If you are shrinking, extrapolating flat revenue overstates your runway. If you are growing, the same flat assumption understates it and can make you cut a team you did not need to cut.

4. A scenario, not a single line

One number invites false precision. Three numbers — base case, a case where the thing you are worried about actually happens, and a case where the thing you are hoping for actually happens — tell you the range you are operating in and which direction the risk sits. The base case is what you report. The downside case is the one that tells you when you need to act, not when you find out you should have.

The single most useful number in a runway model is not the months remaining in the base case. It is the gap between the base case and the downside case. A small gap means you are one bad month from a real decision. A wide gap means you have room to be wrong and still be fine.

Burn multiple is the number that gets ignored the longest

Runway tells you how much time you have. It does not tell you whether the burn producing that countdown is buying you anything. That is what burn multiple is for — net burn divided by net new revenue over the same period — and it is the metric I see least often on a founder's own dashboard, even though investors ask about it constantly once a company has any revenue at all.

A burn multiple under 1 means you are generating more than a dollar of new revenue for every dollar spent — efficient by most benchmarks at any stage. Between 1 and 2 is normal and often fine for an early company still finding its motion. Above 2 is where the conversation usually turns to what specifically the spend is buying, and above 3 it becomes the first question in almost any investor meeting, whether or not you brought it up.

The reason this matters for a runway conversation specifically: a founder who is only tracking months-of-cash-left has no way to distinguish between burn that is buying growth and burn that is just burn. Two companies with identical runway can be in completely different positions — one is spending its way to a metric that justifies the next round, the other is spending its way to a harder conversation. The runway number alone cannot tell them apart. The burn multiple can.

How often to rebuild it

Monthly, at minimum, tied to your actual close — not a rough guess mid-month. If you are within six months of needing to raise or make a real headcount decision, weekly is not excessive. The model does not need to be elaborate to be rebuilt often. A twenty-line spreadsheet updated every Monday beats an elaborate model that gets touched once a quarter, because the entire value of the exercise is catching the moment something changed, not admiring the model's construction.

The other trigger, independent of calendar: any time a material commitment changes. An offer gets accepted. A customer gives notice. A vendor renews at a different price. Each of those is a five-minute update to the bridge. Skipping it is exactly how the gap between the model and reality opens without anyone noticing until it is large.

What this looks like in practice

The founders who handle this well are not the ones with the most sophisticated model. They are the ones who can answer three questions without opening a laptop: what is our runway in the base case, what would move it materially in either direction, and what would we cut first if we needed three extra months tomorrow. If any of those takes more than a few seconds of thinking, the model exists as an artifact but is not actually being used as a decision tool — which, for this specific purpose, is the same as not having one.

None of this requires finance software or a full-time hire. It requires treating runway as something you maintain rather than something you calculate once and repeat in board decks until it is embarrassingly wrong. The discipline is cheap. The alternative — finding out the number was stale at the same moment you needed it to be right — rarely is.