Education and subscription services: deferred revenue and cohort economics

An education company - or any subscription business - carries a large and fast-growing liability called deferred revenue, and a nervous reader sees a swelling liability and worries. They have it exactly backwards. That liability is one of the healthiest things about the business: it is fees already collected in cash, for courses the company will deliver over the coming months. The company has the money; all it "owes" is the service. So a growing deferred-revenue balance is future revenue already banked. In this sector the liability that looks like a debt is a leading indicator of revenue, and the recognised-revenue line a reader instinctively trusts is the laggard. illustrative

The mechanics run in reverse of an ordinary sale. Collect the fee upfront - a year's tuition, an annual subscription - then recognise it as revenue steadily over the period the service is delivered. Cash comes in first; revenue earns out over time; the not-yet-earned portion sits as deferred revenue. So a burst of new enrolments shows up first in the cash collected (billings) and the deferred-revenue balance, and only feeds into recognised revenue over the following months. Read recognised revenue alone and you are looking at the past. Read billings and deferred revenue and you are looking at the future - and read the cohorts, whether each intake stays and pays back what it cost to acquire, for whether that future is durable.

Deferred revenue is a healthy liability. When a student enrols, the company collects the fee upfront and recognises it as revenue steadily over the period it delivers the course. The portion collected but not yet earned sits as deferred revenue - a liability, because the company owes the service it has been paid for. But it is a liability only in accounting form: the cash is already in the bank, and the obligation is to teach, not to pay anyone. This is the reverse of a manufacturer, which delivers the goods and then waits to be paid. A large and growing deferred-revenue balance therefore means the company has collected a lot of fees for services it will deliver - future revenue already banked. A shrinking balance signals sales slowing. Reading it as a worrying debt inverts its meaning entirely. In an Indian filing under Ind AS 115 it usually appears as a contract liability (sometimes labelled "income received in advance" or "unearned revenue"), tucked inside other current liabilities on the balance sheet, with the movement disclosed in the revenue note - so it is often a line you have to go looking for rather than one the headlines hand you.

Billings and deferred revenue lead; recognised revenue lags. Because the cash comes first and the revenue is earned out, the recognised-revenue line reflects sales made over the preceding period, trickling out. The leading indicators are billings - the cash actually collected this period, which captures the latest sales in full - and the deferred-revenue balance, the stock of future revenue banked. A business whose billings and deferred revenue grow faster than its recognised revenue is accelerating: new sales are running ahead of the P&L. One whose billings and deferred revenue are flat is running down its pipeline even while recognised revenue still rises from past sales. The recognised-revenue line is the comforting number that tells you the past; billings and deferred revenue are the signal that tells you the future.

Cohort economics decide durability. The forward pipeline is only valuable if the customers stay, so the real economics live in the cohorts. For each intake, cohort economics asks three things: what did it cost to acquire them, how many stay each period (retention), and what lifetime value do they generate? A business whose cohorts retain well - students who stay for years, subscribers who renew - earns many times its customer-acquisition cost over the customer's life, so growth is durable and profitable. A business whose cohorts churn is a leaky bucket that must keep spending just to stand still, and its growth is expensive and fragile. Retention and the ratio of lifetime value to acquisition cost sit beneath the aggregate enrolment and revenue numbers, and they are what tell you whether the growth is real. Note that these are operating disclosures, not audited statement lines - companies choose whether and how to report them, which is itself a signal: a business proud of its retention publishes the curve, and one that stays quiet about cohorts is usually quiet for a reason.

The distinctive thing about education and subscription is when the cash arrives relative to the revenue, and whether a growing liability is a good sign or a bad one. Set it beside businesses that collect on other schedules and the inversion is plain.

The reader must invert two ordinary instincts. First, do not read the deferred-revenue liability as a debt: for a manufacturer a swelling liability would be a worry, but here it is cash-backed future revenue and its growth is a positive. Second, do not read recognised revenue as the current state of the business: for a manufacturer it measures today's trade, but here it lags the billings and deferred revenue that lead it. And beneath both, the real economics are in the cohorts - whether customers stay and pay back their acquisition cost - which the enrolment and revenue lines never show. The scary liability is a good sign, the trusted revenue line is the laggard, and reading them literally gets the business backwards.

Read the composite education company across five years. Recognised revenue grew steadily from ₹1,200 crore to ₹3,000 crore - a healthy line, but a lagging one. The leading indicators tell a stronger story: billings, the cash actually collected, ran ahead of recognised revenue throughout - ₹3,600 crore against ₹3,000 crore in the final year - and the deferred-revenue balance grew from ₹600 crore to ₹2,000 crore. That growing liability is not a debt to worry about; it is ₹2,000 crore of fees already collected for courses still to be delivered, so it is future revenue banked. The gap between billings and recognised revenue, and the rising deferred balance, mean new sales are running ahead of the P&L - the business is accelerating, and recognised revenue will keep rising as this deferred revenue earns out. illustrative

Now read the durability through the cohorts. Active students grew from 300,000 to 770,000, and - the number that matters most - cohort retention rose from 82% to 88%. High and rising retention means students stay and keep paying, so each cohort generates far more lifetime value than it cost to acquire, and the growth is durable. Revenue per student held steady at around ₹39,000, so the growth is genuine expansion of the base, not price rises masking churn. Had retention been falling - students leaving after the first term - the same billings growth would have described a leaky bucket, spending to acquire students who churn out, its future revenue evaporating as fast as cohorts drop away. The rising retention is what confirms the growing billings and deferred revenue represent sticky customers rather than a revolving door.

A reader who saw only recognised revenue would have found a solid but ordinary growth story. A reader who read the billings, the deferred revenue and the cohort retention saw an accelerating business with sticky, profitable customers and future revenue already in the bank. And a reader who saw the growing deferred-revenue liability and worried about it would have misread the single healthiest thing on the balance sheet.

How it comes up. When an education or subscription company reports revenue growth, a sharp analyst looks past the lagging line to the leading indicators and the cohorts: "Recognised revenue grew, but how did billings and deferred revenue grow, and what's your cohort retention and the lifetime-value-to-acquisition-cost ratio - is retention holding across your newer, discounted cohorts?"

A good answer.

"Right to look past recognised revenue. Billings grew 24% against 20% recognised, and deferred revenue is up 33%, so new sales are running ahead of the P&L. On cohorts, first-term retention is 88% and rising, and lifetime value to acquisition cost is about 4.5x. Importantly, retention in our newer cohorts is in line with the mature ones - the discounting we did to acquire them hasn't hurt stickiness."

It gives billings and deferred-revenue growth ahead of recognised revenue, the retention and the lifetime-value ratio, and confirms the newer cohorts retain. You can judge both the forward pipeline and the durability.

An evasive answer.

"We're thrilled with our strong revenue growth and expanding learner base. We're a mission-driven platform transforming education at scale, and our north-star metrics are trending well. Momentum is strong across all our programmes."

It cites "strong revenue growth" (the lagging number) and "expanding learner base" without billings, deferred revenue, or cohort retention, and offers "north-star metrics trending well" in place of the actual retention and lifetime-value figures. A business acquiring students who churn out, or whose newer cohorts retain poorly, would answer exactly this way.

The follow-up nobody asks. "What is your retention curve by cohort, and how does lifetime-value-to-acquisition-cost look for your most recent, discounted cohorts specifically?" If "strong growth, expanding learner base" is allowed to stand, an investor credits enrolment growth without knowing whether the students stay. The silence is the tell - either retention is falling in the newer cohorts, or the customers are being acquired for more than they will ever pay back.

Reading the liability as a debt. Deferred revenue appears as a liability, and a swelling liability instinctively looks like rising indebtedness. But it is fees already collected in cash for services not yet delivered, so a growing balance is future revenue banked - the strongest evidence of forward sales, mistaken for a red flag. What it cannot tell you, though, is the margin at which that revenue will be delivered: the cash is banked, but teaching the course has a cost, and if delivery is poor or expensive the deferred revenue earns out thin and the customer does not renew. Future revenue banked, yes - but its value depends on the margin and the satisfaction it produces, neither of which the balance itself is.

Reading recognised revenue as the state of the business. Because cash is collected upfront and earned out over time, the recognised-revenue line lags - a business can show rising recognised revenue while its new sales are already slowing, or ordinary recognised revenue while its billings and deferred revenue accelerate. Reading the P&L line alone is reading the past; the billings and the deferred balance lead it and reveal whether the future is building or emptying.

Reading enrolment growth as success without the cohorts. A business can grow its headline customer count by acquiring people who churn out quickly - a leaky bucket, expensive and value-destroying if they leave before paying back acquisition cost. It is precisely the fastest-growing edtech and consumer-subscription businesses, buying customers who do not stick, that have destroyed the most capital. And even the reported cohort metrics deserve suspicion: what counts as a retained customer, over what period, and how lifetime value is estimated all involve assumptions a company under pressure can make generously, and a flattering blended retention figure can hide poor retention in the newer or discounted cohorts. The right things to read are retention and lifetime value - but read the curve by cohort, not the headline, because the acquisition costs and retention that look healthy today can erode as competition intensifies and alternatives proliferate.