Meal subscription: why good unit economics brings no cash

“Six hundred subscriptions, two and a half million in revenue. An investor looked at my unit economics and said that with numbers like these I should be opening a second city. And that same week I was asking my chef to wait three days for his wages” — that is how Taras Melnyk, owner of a lunch subscription service in Lviv, opened the conversation.

If you run a subscription, you have heard this from yourself. The metrics are in place, the base is growing — and every month the account holds exactly enough to reach the next round of payments.

The root is almost always the same: the two main numbers of a subscription live on different dates. LTV — the total contribution of one subscriber for the whole time they stay with you — is 9 272 ₴ in Taras's case. CAC, the cost of acquiring one subscriber, is 1 882 ₴. A ratio of 4.9, and any investor will nod. Yet the margin of that very same business is 1.75%.

This is not a calculation error, it is an exact description of it: LTV arrives over nine months, while CAC is paid today. We break the subscription down into two units — one subscriber and one month of their life — and show why churn of 11% costs Taras 124 212 ₴ a month, 2.82 times more than his entire profit.

The founder's road: from ten lunches in a cooler bag to six hundred subscriptions

Taras started with ten lunches a day: he cooked at home, drove them around himself and took cash in the stairwell. His maths was simple — collect for the week, subtract groceries and petrol, the difference is mine.

Four years on there was a rented kitchen, a chef and four cooks, two managers on the chats and an outside courier service. The subscription came to cost 4 200 ₴ for 20 lunches, subscribers grew to six hundred, revenue to 2 520 000 ₴.

That is precisely when the clarity disappeared. Revenue grew sixtyfold, and free cash did not grow at all — on the contrary, for the first time there were weeks when it was not enough for the ordinary payments.

“When I did the deliveries myself, I saw every one of them: who did not open the door yesterday, who has been turning it down for two weeks. Now I have six hundred lines, and about each one I know a single thing — that it is paid for.”

Why “six hundred subscribers” says nothing about money

A subscription is not a sale, it is a promise to feed a person for 20 working days. The money for it arrives at the start of the month, while the costs stretch across all 20 days and do not depend on whether that person will still be with you next time.

While all six hundred sit as a single figure, the main thing stays invisible: how many of them have only just arrived and have not yet paid back the cost of acquisition, and how many will be gone before the next billing. The number of subscribers is not a base, it is the balance of two flows that nobody counts separately.

What the monthly report shows

Revenue of 2 520 000 ₴ is 600 subscribers at 4 200 ₴. Variable costs, that is groceries, packaging and delivery, are 3 180 ₴ per subscriber, 1 908 000 ₴ in total. The subscribers' contribution towards covering everything else is 612 000 ₴. The kitchen's fixed costs are 568 000 ₴. Profit of 44 000 ₴, a margin of 1.75%.

Working numbers for prepared-food delivery. Only none of them answers why, on revenue of two and a half million, the owner is postponing the wages.

Life before churn became a line

The bookkeeping Taras lived with for those four years will be recognised by any subscription service that grew out of its own kitchen.

• Subscribers were seen as a single figure. “Six hundred active” — with no split between those who have been with the service for a year and those who arrived this week.
• Churn was treated as a marketing metric. It was of interest in the advertising reports, not in the payment calendar.
• Marketing was judged by the number of new ones. Sixty-eight subscriptions in a month is a success, and nobody asked how many of them merely plug a hole.
• The delivery zone was not split into near and far. The courier drove out of town for the same subscription price, and in the books that address was no different from the next street.
• Uneaten lunches had no price. That people do not eat everything was known in the kitchen. How much it costs in hryvnia was known by nobody.
• Pausing a subscription was a courtesy service. The manager set it in two clicks because he felt sorry for the client, and it appeared nowhere as a line.

There is not a single item in this list about the kitchen or the vehicles. Six decisions were taken with no figure in front of anyone's eyes — which is why neither the chef nor the accountant could have spotted them.

The costs that do not depend on the number of subscribers

The first thing Taras did was write out everything the kitchen pays every month without change, and put the share of revenue next to it.

Cost item	Per month	% of revenue
Kitchen rent	95 000 ₴	3.77%
Chef and 4 cooks	214 000 ₴	8.49%
Managers	86 000 ₴	3.41%
Marketing	128 000 ₴	5.08%
Other	45 000 ₴	1.79%
Total	568 000 ₴	22.54%

What corridor these 568 000 ₴ leave

These costs take 22.54% of revenue, and not one line will get smaller if there are a hundred fewer subscribers. A contribution of 612 000 ₴ minus 568 000 ₴ gives a profit of 44 000 ₴ — a margin of 1.75%.

In a corridor like that every decision is expensive. One per cent of the price given away to a subscriber is 25 200 ₴ off revenue of 2 520 000 ₴, that is 57.3% of the entire profit. One per cent — more than half the earnings.

The main unit: one subscriber, not one subscription

A subscription is a line on a price list. A subscriber is a person with their own arrival date, acquisition cost and lifetime. The money of the service is decided by the second one.

Metric per subscriber	Value
Subscription price per month	4 200 ₴
Variable costs: groceries, packaging, delivery	3 180 ₴
Contribution towards fixed costs	1 020 ₴
CAC — cost of acquisition	1 882 ₴
CAC payback	1.85 months
Lifetime at churn of 11%	9.09 months
LTV — contribution over the whole lifetime	9 272 ₴
LTV / CAC	4.93

Four numbers here concern one person, and they are easy to merge into a single feeling that “a subscriber brings money”. Let us separate them. 4 200 ₴ is what the person pays. 3 180 ₴ is what goes into her lunches. 1 020 ₴ is what is left for the kitchen. 1 882 ₴ is what was paid to advertising so that she heard about the service at all: marketing of 128 000 ₴ divided by 68 new subscribers.

Why payback of 1.85 months looks safe

Divide 1 882 ₴ by the contribution of 1 020 ₴ — a new subscriber returns the cost of her acquisition in under two months. After that, if she lives the average 9.09 months, the kitchen receives another 7.24 months of clean contribution.

And this is where everybody stops. The figure is pretty and genuinely correct — it simply answers a different question from the one the payment calendar puts to Taras.

LTV five times bigger than CAC — and there is no money

Let us put two numbers side by side. LTV / CAC = 4.93: every hryvnia put into acquisition comes back as almost five hryvnia of contribution. Money gets given for a ratio like that. The margin of that very same business is 1.75%.

There is no contradiction here, even though it looks exactly like one. Both numbers are honest, they are simply about different stretches of time. And do not confuse two quantities that only resemble each other on paper: 9.09 is months, the average lifetime of a subscriber at churn of 11%, while 9 272 ₴ is hryvnia, her total contribution over those months. The contribution arrives in portions of 1 020 ₴ for nine months. The CAC of 1 882 ₴ leaves the account in a single payment today.

Unit economics measures a business over the distance, while the bank account lives in the current week. As long as the service stands still, the shift goes unnoticed. The moment it acquires more, it pays today and receives in instalments until the end of next year.

“I showed the investor a ratio of four and ninety-three and believed in it myself. And then I realised that my nine hryvnia out of every ten will arrive in instalments until the end of next year, while the invoice from the agency comes on Friday.”

The gap between this month's advertising and the newcomers' contribution

The simplest way to see the shift is to compare two payments from the same month. The kitchen's marketing is 128 000 ₴, spent in full. The contribution given by those acquired in that same month is 68 subscribers at 1 020 ₴, that is 69 360 ₴.

The gap is 58 640 ₴. It will close in 1.85 months, and that is the good news. The bad news is that next month the kitchen will again spend 128 000 ₴ and again receive 69 360 ₴ — and so on every time, as long as the service runs.

Sixty-six for replacement: the cost item that is not in the report

Taras's churn is 11% a month. Out of six hundred subscribers that is 66 people who cancel within a month. Acquisition brings 68 new ones. Subtract one from the other: the net growth of the base for the month is 2 subscribers.

Now convert those 66 into money. To bring the base back to six hundred, the kitchen has to buy 66 people at 1 882 ₴ — that is 124 212 ₴. Out of the marketing budget of 128 000 ₴, 97% went on replacement. For growth there remained 128 000 − 124 212 = 3 788 ₴, and that is exactly two subscribers at 1 882 ₴. That is where the net growth of 2 comes from: not from the quality of the advertising, but from the fact that 66 of the 68 people bought step into the place of those who left.

Put 124 212 ₴ next to the profit of 44 000 ₴: churn costs the kitchen 2.82 times more than it earns. This is Taras's most expensive item — more expensive than the chef with his four cooks, and than the rent.

• It has no line of its own. Churn replacement sits inside marketing, and marketing in the report looks like spending on growth.
• It has no date. A person does not file a cancellation notice — she simply does not renew, and nobody records that event.
• It has nobody to blame. Acquisition is the marketer's zone, lunch quality is the chef's zone, and the cancellation lives between them and therefore belongs to no one.

Why churn does not look like a cost

Because it does not leave the account as a separate payment. Rent is visible in the statement, groceries in the invoices, and churn only in the fact that marketing adds nothing to the base although it is spent in full.

While churn is measured in per cent, it seems like a technical metric: eleven, that is not much. The moment it becomes a line in hryvnia, 11% turns out to be the biggest invoice of the month.

“For two years I asked the marketer for more subscriptions and took offence at the result. And he was giving me sixty-eight of them. It is just that sixty-six of them I immediately spent on not getting smaller.”

What cutting churn from 11% to 8% gives

Taras did not take the target 8% out of thin air: that is what subscribers who lasted more than four months were showing. At churn of 8% the base loses 48 people instead of 66, and replacing those 48 takes 90 336 ₴ instead of 124 212 ₴.

Metric	Churn of 11%	Churn of 8%
Subscribers lost per month, people	66	48
Needed for replacement at the CAC rate	124 212 ₴	90 336 ₴
Subscriber lifetime, months	9.09	12.5
LTV — contribution over the lifetime	9 272 ₴	12 750 ₴
LTV / CAC	4.93	6.77
Saving per month	—	33 876 ₴

Two lines here look alike and mean different things. The lifetime stretches from 9.09 to 12.5 — those are months. LTV grows from 9 272 ₴ to 12 750 ₴ — those are hryvnia over those months. The cost of acquisition does not change: the same 1 882 ₴.

Why 33 876 ₴ is worth more than thirty new subscribers

A saving of 33 876 ₴ every month is 77% of the profit and 406 512 ₴ a year. It demands no new advertising, and no second city either: the price stays 4 200 ₴, the variables 3 180 ₴, the fixed costs 568 000 ₴. One thing changes — how many people survive to the next billing.

To add the same 33 876 ₴ as contribution you need 33 extra subscribers at 1 020 ₴ — and they have to be paid for with 33 × 1 882 ₴ = 62 106 ₴ of advertising, which is not in the budget. Lower churn asks for nothing: with the same 68 acquired, the net growth becomes 20 people instead of 2.

“Three per cent of churn. All my life I thought of that as a rounding error. It turned out to be thirty-three thousand a month — three quarters of what I earn, and without a single new hryvnia into advertising.”

The surcharge for far districts: 39 480 ₴

Next come three items that are not in the report at all. The first lives in the courier's route. Out of six hundred subscribers, 94 have an address outside the dense delivery zone, and reaching each of them costs 420 ₴ more a month: 94 × 420 ₴ = 39 480 ₴, that is 89.7% of the profit.

Where this surcharge hides

In the average. Variable costs of 3 180 ₴ are an average across all six hundred, and the surcharge for far addresses already sits inside. A subscriber from the centre brings the kitchen more than the average 1 020 ₴, a subscriber from a far address 420 ₴ less. In the report both are the same line for 4 200 ₴.

So the kitchen sells one service at one price, but performs it at two different costs: the cheaper part of the subscribers pays for the more expensive one. As long as the delivery zone does not stand next to the address in the books, this skew has neither a sum nor anyone to blame.

Uneaten lunches: 114 480 ₴ cooked for the bin

The second item lives in the kitchen itself. A subscription is 20 lunches, and on average 17 get eaten. Three lunches per person are left unclaimed, and part of them the kitchen manages to cook.

Let us count. Variable costs of 3 180 ₴ divided by 20 lunches — the cost of one lunch is 159 ₴. That is the price of one portion, not of the subscription. Of the uneaten ones, 40% get cooked and thrown away: 3 × 0.4 × 159 ₴ = 190.8 ₴ per subscriber. Across six hundred people that is 114 480 ₴ a month, 2.6 times more than the profit.

Why this money is already spent

Check it from the other side: 1 800 uneaten lunches a month, of which 720 are cooked, 720 × 159 ₴ — the same 114 480 ₴. In the report the sum sits inside the 1 908 000 ₴ of variable costs and looks like normal cost of sales — it is just that nobody fed stands behind it.

This is the quietest loss in delivery: no refund, no complaint, no empty seat. The subscriber is satisfied, the payment went through, the lunch went into the bin.

“I haggled over every hryvnia with the vegetable supplier. And then I worked out that every month I cook seven hundred and twenty lunches nobody will thank me for, because nobody ever saw them.”

Pausing a subscription: 16 830 ₴ of contribution not received

The third item grew out of a good intention. A subscriber travels, falls ill or asks for a break — the manager sets a pause and keeps her place and terms. There are 44 such people every month, and the pause lasts on average a week and a half out of four, that is 0.375 of a month.

It is easy to overstate the loss here, and Taras overstated it at first. Gross revenue not received is 44 × 4 200 ₴ × 0.375 = 69 300 ₴. But the lunches for those days are not cooked, so variable costs of 44 × 3 180 ₴ × 0.375 = 52 470 ₴ do not arise either. The real loss is the difference: 69 300 − 52 470 = 16 830 ₴, or 38.3% of the profit. The gross figure overstates it by a factor of 4.1, and that is exactly why pauses have to be counted in contribution and not in revenue.

The place meanwhile is taken, the manager spends time on rescheduling, and the fixed 568 000 ₴ do not shrink by a single hryvnia: the kitchen rent cannot be paused.

How much Taras actually earned

Let us add up the month. The report shows a profit of 44 000 ₴. Two of the three items are contribution the kitchen does not receive: the surcharge for far districts of 39 480 ₴ and the subscription pause of 16 830 ₴. Together 56 310 ₴ — that is 128% of the profit, which is more than the kitchen earns in a whole month. If both items disappeared, Taras would have 100 310 ₴ — 2.28 times more than now.

The third item, the 114 480 ₴ of lunches cooked and uneaten, must not be added into that sum: it is money of a different nature — already spent inside the variable costs, not contribution that is missing. And alongside all three stands the most expensive item of the month: 124 212 ₴ to replace those who left.

Not one of these sums grew out of bad lunches or weak advertising. The lunches are good, the advertising delivers its 68 subscriptions every month. All four rest on numbers the books never had: on the time a person spends with the service, on the address the courier drives to every day, and on the number of lunches she really eats.

What changed when churn became a line

Taras did three things. Every subscriber got an arrival date, a cancellation date and a count of lunches eaten. Churn began to be counted in hryvnia and was put next to the advertising budget. And the surcharge for far addresses, the pause and the uneaten lunches got lines of their own and people responsible for them. Here is what two quarters showed.

• Churn stopped being a percentage. Next to the budget of 128 000 ₴ there appeared the line “replacing those who left — 124 212 ₴”, and it became visible that 3 788 ₴ is what is left for growth. Managers started calling those who had not ordered for three days running: those were exactly the people who left first.
• Far delivery got its price. The 94 subscribers with an address outside the dense zone became a separate group in the report, and for new addresses like that a surcharge appeared. The 39 480 ₴ stopped being part of the average 3 180 ₴ and turned into a decision that can be taken.
• The pause got rules. The place is held for two weeks, after which the subscription either continues or closes. The 16 830 ₴ did not disappear, but they stopped being open-ended, and the manager finally saw how many people he was holding a place for free of charge.

A new kitchen, a higher price and a bigger budget took no part in this. One thing was needed: a subscriber with a date in the line — and an answer to the question of how many of them already have that date filled in.

The takeaway for entrepreneurs. Where the client pays every month, a healthy LTV to CAC ratio guarantees nothing: it measures years, while wages are paid on Friday. The main number of a subscription is not the count of active ones and not the cost of acquisition, but the share of those who survive to the next billing. As long as churn stands in per cent in an advertising report rather than in hryvnia next to the rent, growing the base only means that you are buying the same people a second time.

Advice to finish with

• Convert churn into hryvnia. Multiply the number of those who left by CAC: 66 × 1 882 ₴ = 124 212 ₴ a month. Put that sum in the report next to the rent — it is more expensive.
• Compare replacement with the whole advertising budget. In the example 124 212 ₴ is 97% of 128 000 ₴. The remaining 3 788 ₴ is the only money really working on growth.
• Keep CAC payback next to the lifetime. Payback of 1.85 months against a life of 9.09 months — there is a margin. With a lifetime of two months the very same CAC is enough to zero out the business.
• Do not take decisions on the LTV to CAC ratio. 4.93 is a fine number for a presentation and a helpless one for the payment calendar: 9 272 ₴ arrives in instalments, 1 882 ₴ is written off at once.
• Count the cost of one unit of service, not of the subscription. 3 180 ₴ over 20 lunches is 159 ₴ per lunch, and it is this figure that shows that three uneaten lunches cost 190.8 ₴ per subscriber.
• Count the cost of delivery by zones. 94 far addresses at 420 ₴ of surcharge is 39 480 ₴ a month, 89.7% of the profit. In average variable costs of 3 180 ₴ that difference is invisible.
• Measure pauses in contribution, not in revenue. 44 people for a week and a half give 69 300 ₴ of gross revenue, but 52 470 ₴ of variables do not arise either: the real loss is 16 830 ₴, that is 38.3% of the profit.

The money does not disappear — you simply do not see it

A subscription service looks like a business about taste and logistics: cook well, deliver on time. In reality taste and logistics are the entry condition, no more. Money here is decided by one thing: how many billings a person gets through before she leaves.

These 11% will not become eight on their own. But the moment a subscriber becomes a line with an arrival date and a cancellation date, everything is visible: that the kitchen's most expensive item is 124 212 ₴ to replace people already bought, that far addresses and pauses take away 56 310 ₴ of contribution against a profit of 44 000 ₴, and that three per cent of churn is worth 406 512 ₴ a year.

How to see this in your own subscription service

Everything described above rests on a single table: every subscriber is a line showing the join date, the leaving date, the monthly contribution and what their acquisition cost. While churn lives as a feeling that some left and some joined, neither the lifetime nor the fact that most of the marketing goes on replacement rather than growth is visible. When the same data is collected automatically, churn stops being a reporting metric and becomes a cost line — which is exactly why small and medium business owners use Finmap: to see the financial picture instead of assembling it at the end of the month.

Read also

• Fitness club: why the money in your account is a debt, not a profit
• Course completion as a financial metric for an online school

Topic foundation: Management Accounting Explained: What It Is and Why Owners Need It