Ask your team one question this week:
Which of our subscribers is this business actually built on?
Not how many. Which. Which ones came in through which channels, on which offers. Which ones are growing on their own, and which ones are quietly bleeding out inside a healthy-looking average.
I can tell you what happens next, because it happens the same way almost everywhere. First a pause. Then someone pulls up a dashboard. The dashboard has subscriber counts, revenue, growth, churn, NRR. It does not have the answer. So an analyst gets assigned to "pull something together," and weeks go by.
That pause is the most expensive silence in the subscription economy right now. To understand why, don't start with the dashboards. Start with the money. The money moved first.
Follow the money
A company is worth its revenue times a multiplier. For most of the last decade, one thing set the multiplier: how fast you grew. Money was nearly free, so a dollar arriving five years from now was worth almost as much as a dollar today. Investors paid for speed and asked nothing else. In SaaS, the dial was ARR growth. In media, it was subscriber adds. Volume dials, both.
Then money got expensive. Starting in early 2022, rates went from near zero to over five percent, and every future dollar took a haircut. The question investors ask changed with it. Not "how fast are you growing?" but "will this revenue still be here in five years?"
You can watch the two eras collide on a single date. On April 19, 2022, Netflix reported it had lost 200,000 subscribers, its first decline in more than a decade. The stock fell 35% the next day and $54 billion in market value was gone. That wasn't the new math arriving from nowhere, and it wasn't the old math working as designed. It was the collision: the most-watched volume dial in the subscription economy went negative in the first year the market had stopped forgiving. From that day, nobody trusted the old dial to carry a valuation on its own.
The paper trail is public. Meritech, which runs the regression on software multiples every week, marks the break plainly: after the early-2022 selloff, growth and profitability converged as the things that set the multiple. Growth alone stopped being the formula. KeyBanc and Sapphire's annual survey puts prices on it: public SaaS companies below 20 on the Rule of 40 trade at a median 2.6x EV/NTM revenue. Above 40, the median is 7.7x. Software Equity Group, a bank that actually sells these companies, publishes the retention ladder: below 100% net revenue retention, about 3.1x. Above 120%, about 9.3x. And SaaS Capital's valuation model runs on exactly three inputs: public multiples, ARR growth, and NRR.
The variables never changed. The weights did, when money stopped being free.
The executives followed the money. Netflix told shareholders in its Q1 2024 letter that its primary financial metrics are now revenue and operating margin, and starting with Q1 2025 it stopped reporting paid memberships every quarter. The number that cost them $54 billion got deleted from the earnings report. Bob Iger said it straight out on Disney's Q3 FY2023 earnings call: subscriber growth "had been the key measure of success for many," before he reset the streaming business around profitability. And Match Group put it in writing to shareholders. Its Q3 2023 letter said the company was "essentially resetting the Payer base at a lower number but paying a higher rate." Read that again. A public company chose fewer subscribers, on purpose, put it in an SEC filing, and called it progress.
So the money moved in 2022. The executives moved in 2023 and 2024. The reporting hasn't moved at all. Inside most subscription companies, the numbers a board reviews still answer the question the industry inherited from the growth era: how many, and how much. The market now prices a different question: which.
Why your numbers can't disagree with you
Every metric on the standard stack is a blend, and a blend averages away the exact information the decision needs.
NRR is the clearest case. A 105% NRR can describe a healthy business with broad expansion and modest churn. The same 105% can describe a business where two strong cohorts are covering for decay everywhere else. Same number. Opposite companies. Opposite correct strategies. A number that reads the same when things are fine and when they're not isn't information. It can't disagree with you. And it can't be managed — NRR is a readout, not a lever. The number is made by what's inside it.
The same flaw runs through the whole stack. Blended CAC hides the channel that's efficiently buying churn. Average LTV hides that "lifetime" means eleven years in one cohort and eleven months in another. Topline churn can't tell you whether you're losing subscribers you could afford to lose or the ones you can't. These aren't rounding errors. The subscribers inside one average can be two different businesses — one compounding, one dying — and the blend reports their midpoint as if anybody actually paid you the midpoint.
None of this is a missing-data problem. The raw material exists — that's exactly how a diligence team can rebuild your cohort curves without ever asking you. But it's all organized to answer how many and how much. Nobody has the which view handy, because nobody asked for it. That's a question problem, not a data problem, and every decision on the table — pricing, packaging, channel mix, retention spend, and what the company is worth — turns on which.
What the board is actually approving
When a board approves a growth plan, it's approving a theory about which customers the company should buy more of. If the reporting can't tell a compounding subscriber from a decaying one, the board approves that theory blind. Then the operating team executes it blind, chasing headline numbers that reward volume no matter what it's made of.
The failure mode is specific and I'd bet you've seen it. The business hits its numbers for six or eight quarters while the quality of the base rots underneath. By the time blended NRR finally turns, the problem is years old and expensive to fix. The metrics were green the whole way down.
And the bill now comes due at the worst possible moment, because buyers stopped trusting the blends before you did. Diligence teams don't read your reporting. They take your raw billing data and rebuild the cohort curves themselves. Every consumer subscription S-1 contains cohort disclosures the company never produced until bankers demanded them. The "which subscribers" question gets answered at pricing time no matter what. The only open variable is who answers it first. You, with time to act on what it shows. Or your buyer, with the information advantage pointed straight at your price.
Why nothing on your stack answers this
The obvious comeback: isn't this what the analytics budget is for?
Look at what the tools actually do. Billing platforms, product analytics, revenue metrics, customer success, churn prediction. Every one of them measures how many and how much with more precision than the generation before it. Better dashboards. Same question.
Precision on the wrong question is not progress. A churn score tells you who's leaving. It doesn't tell you whether they were load-bearing. A revenue metrics tool computes NRR flawlessly. It can't tell you the number is misleading. The which question lives between the behavioral data and the commercial data, at the cohort level, over time. The stack, as sold, is not organized to sit there.
That's not a knock on the vendors. They're answering the question the market asked for thirty years. It's a vacancy. The most consequential question in the subscription economy has no owner.
What to do on Monday
Run the audit once, by brute force, on the data you already have.
Ask which ten percent of your base you'd rebuild the company around, and whether your acquisition spend is buying more of those subscribers or fewer. Ask what your NRR is without your top two expansion cohorts. Ask which channel-and-offer combinations produce the subscribers you still have three years later, and which behaviors mark those subscribers in their first ninety days.
Expect it to hurt. In most companies this means analysts hand-stitching identity across a billing system, a product analytics tool, and a CRM that were never built to agree on who a subscriber is. That pain is not a staffing problem. It's the finding. When the most consequential question in your business takes heroics to answer once, and can't be answered continuously at all, you've learned something true: your measurement layer was built to instrument a race you're no longer running.
The subscription businesses that get asked to justify their value over the next few years — to boards, to markets, to buyers — won't be separated by how fast they grew. The market already re-priced speed. It pays for durability now, and every company knows it, so every company claims it. The claim costs nothing. They'll be separated by the proof: whether they can show which customers they're built on. The question is already being asked at the top of the market. The answer, so far, has no name.