पाठशाला Pathshala · हिसाब Hisāb, Unit economics · Lesson 22 · Build
When the unit economics say stop
Negative contribution can be a plan or a slow death, and the difference is a written threshold. Set the numbers at which the company stops buying volume and fixes the unit, before a round decides for you.
Pathshala, The Founder Library · 11 October 2026 · 7 min read
Every company that loses money on each order has a story about why that is temporary. Some of the stories are true. The way to tell is not conviction or the next board meeting but a set of thresholds, written down while the numbers are still comfortable, that say in advance when the company stops buying volume and starts fixing the unit.
This lesson sets those thresholds. It separates negative contribution that is a plan from negative contribution that is a habit, explains why raising more money is the wrong first answer to a breach, models the choice between growing and holding, and ends with the monthly review that keeps the decision from being made by a bank balance.
Negative contribution is a slope, not a state
[Contribution](/library/contribution-margin-first-number-to-know) is what one order leaves after the costs that exist only because it was sold: goods, courier, gateway, packaging, returns and the discount that closed it. When it is negative, every additional order deepens the loss, and growth multiplies the problem rather than diluting it. Fixed cost can be outgrown. Negative contribution cannot.
Yet listed companies run businesses below zero on purpose. Swiggy’s Q4 FY25 shareholder letter reported Instamart’s contribution margin at minus 5.6 per cent of gross order value, down from minus 4.6 per cent the quarter before, while its order value grew 101 per cent year on year. The same letter told shareholders it expected contribution break-even in three to five quarters. That is what makes negative contribution a plan: a number, a direction and a date, said in public, funded by a balance sheet that has agreed to carry it.
A seed or Series A company is in a different position. It has a number. It may not have a direction, because nobody has plotted contribution per order month by month. It almost never has a date. So the first discipline is to treat contribution as a slope: what it was six months ago, what it is now, how many rupees an order it improves by each month, and when, at that rate, it crosses zero.
Three thresholds, written before you need them
The fix rate. Decide the minimum improvement in contribution per order each month that the plan depends on, and measure it on a three-month rolling basis so one good month does not hide two bad ones. A company at minus ₹45 an order that needs to reach plus ₹40 inside a year must add about ₹7 an order every month. If the rolling figure falls below that, the plan has already slipped.
The crossing date against the runway. Divide the distance to positive contribution by the fix rate and you have a month. That month must fall at least six months before the cash runs out on the current plan, because a company needs those months either to raise on the strength of fixed economics or to cut to profitability without a round. Paul Graham’s test in Default Alive or Default Dead? asks the same question of the whole company: with expenses held constant and revenue growing at its recent rate, does it reach profitability on the money it has left? A company that cannot answer yes at the unit level cannot answer yes at all.
The cost of each new rupee. David Sacks defined the burn multiple in April 2020 as net burn divided by net new annual recurring revenue, and wrote that burning three times or more to achieve growth is a sign that product-market fit may not be what it appears. Transactional businesses can use the same ratio on net new monthly contribution. Set a ceiling, compute it quarterly, and read the full method in the lesson on [burn multiple](/library/burn-multiple-and-the-discipline-of-efficient-growth).
The rule that binds them: when any one threshold is breached for two consecutive months, the company moves to the pause described below. Two months, because a single month can be a festival, a courier strike or a price war. Not three, because by the third month the decision is usually being made by the runway instead of the founders.
Why raising more is the wrong first answer
The natural response to a breach is a round. The unit is nearly there, the reasoning goes, so more money buys the months needed to finish. Graham described what happens next in The Fatal Pinch: founders overestimate their chances of raising more money, because their spending has grown since the last round, investors now hold them to a higher standard, and the company looks more like a failure than it did when it was too early to judge. In the default-alive essay he names the condition precisely: default dead, plus slow growth, plus not enough time to fix it.
Investors read the same thresholds. A company that raises with contribution still negative and no improving slope is asking someone else to fund the experiment it chose not to finish. A company that paused, fixed the unit and then raised is asking for money to multiply something that works. The second conversation is shorter, and its valuation is set on better numbers. Treat the round as plan A for growth and never as plan A for survival.
Grow now, or hold and fix
The honest case for growing through negative contribution is that scale itself fixes the unit: courier rates fall at volume tiers, purchase prices fall with larger orders, fixed fulfilment cost spreads across more orders. Sometimes that is true. More often the growth itself takes the fix back. The volume is bought with discounts. The new cities come with worse rates and longer routes. The team that should be renegotiating contracts and cutting returns is busy shipping orders.
Take a Hyderabad grocery delivery company at 30,000 orders a month, losing ₹45 an order before fixed cost, with ₹40 lakh a month of fixed cost and ₹8 crore in the bank. Holding volume, its team believes it can add ₹8 an order each month and reach plus ₹40 in eleven months. Growing at 12 per cent a month, it keeps only ₹3 of that ₹8, because the growth is bought with first-order discounts and two new cities. The figure runs both paths.
On those numbers the growing company runs out of cash in its fifteenth month and by month 24 has lost more than ₹2 crore on orders along the way; the company that holds stays solvent throughout. Now set the fix lost while growing to zero, which gives growth every benefit of the doubt. Growing wins comfortably. That is the whole argument in one control: the decision turns on whether growth helps or hurts the fix, and that is an empirical question the company can answer from its own last six months, by comparing the improvement in cities it entered recently with the improvement in cities it has run for a year.
Growth is a multiplier. Pointed at a unit that loses money, it multiplies the loss; the only question is whether the unit improves faster than the multiplier works.
What pausing actually means
Pausing is not stopping the company. It is a short list of specific actions taken together. Stop paid acquisition in every channel whose customers do not repay their acquisition cost inside the payback window, and keep the channels that do. Freeze expansion: no new cities, categories or SKUs until the core passes its thresholds. Retreat to the profitable core: rank cities, segments or products by contribution and close the bottom of the list. Hold hiring outside the people working on the fix. Point the team at the unit: courier and supplier renegotiations, a price rise, a minimum order value, a cut in returns, a mix shift toward prepaid orders.
Indian e-commerce has done this in public before. In February 2016 the grocery start-up PepperTap, which had raised close to $50 million and had been aiming for 75 cities, closed operations in six cities and kept eight, among them Delhi, Gurugram, Pune, Hyderabad and Bengaluru; its chief executive said the company had decided to focus on depth rather than breadth. The lesson is in the timing. A retreat chosen with a year of runway left is a strategy. The same retreat chosen with two months left is a wind-down.
Set the conditions for restarting growth at the same time as the pause. Contribution per order positive for three consecutive months; a fix rate that held while volume grew modestly; the crossing date and the runway back in the right order. Then grow again, one city or one channel at a time, and watch whether contribution holds.
The monthly stop review
On the fifth working day of each month the founders and whoever runs finance meet for thirty minutes with one page. The page carries contribution per order for each of the last six months, by city or channel; the three-month rolling fix rate against the rate the plan needs; the projected month of positive contribution at that rate; months of runway on the current plan; and the burn multiple for the quarter to date.
Each threshold is marked as passed or breached, with the number of consecutive months in breach. If any line shows two, the pause starts that week and the founders tell the board what they paused and why, before the board asks. If none does, the meeting records the numbers and ends. Write the thresholds into the board minutes once, when the company is comfortable, so that the decision in a bad month is already made.
Nothing here is legal, tax or investment advice. The Hyderabad company is illustrative; your own six months of contribution data is the only model that can decide this for you.
Sources
- Paul Graham, Default Alive or Default Dead?, October 2015 — The test: profitability on the money left at constant expenses and recent growth; the fatal pinch as default dead plus slow growth plus not enough time.
- Paul Graham, The Fatal Pinch, December 2014 — Founders overestimate their chances of raising more money.
- David Sacks, The Burn Multiple, April 2020 — Net burn divided by net new ARR; burning 3x or more as a warning sign.
- Swiggy Limited, Q4 FY2025 Shareholder Letter, May 2025 — Instamart contribution at minus 5.6 per cent of GOV, GOV up 101 per cent, contribution break-even expected in three to five quarters.
- TechCrunch, Fearing current investment climate, India’s PepperTap scales back its e-grocer service, February 2016 — Six cities closed and eight kept; close to $50 million raised; a target of 75 cities.