A SaaS dashboard showing MRR climbing up and to the right is one of the most dangerous things a founder can look at. Not because the number is wrong. Because it is technically correct — and still hiding a fire.
Here is the pattern. You close new accounts. Some existing customers upgrade or add seats. Top-line monthly recurring revenue ticks higher. Everyone sees progress. The board deck looks healthy. And underneath, a growing share of last quarter's cohort is quietly canceling or downgrading.
Expansion revenue papers over contraction. The total goes up, so nobody asks hard questions. Until one month the new logos slow down — because they always do — and MRR drops. The team scrambles. But the churn didn't start that month. It started six months ago. You just couldn't see it through the aggregate number.
Think of a bathtub with the faucet on full and the drain half-open. The water level rises, so you assume the plumbing is fine. Turn the faucet down slightly and the water drops fast.
Expansion revenue works the same way. A handful of power users upgrading their plans can offset dozens of smaller accounts leaving. In aggregate, the math works. At the cohort level, it is a disaster. You are replacing durable revenue with revenue that depends on a small number of accounts continuing to grow.
This is especially common in products with usage-based pricing or seat-based tiers. The accounts that expand tend to be large. The accounts that churn tend to be small. The dollar-weighted average tells a story of health. The logo-weighted view tells a story of attrition.
The founder who catches this usually catches it by accident. Maybe they pull retention data for a fundraise and realize that fewer than half of the customers who signed up eight months ago are still active. Or they look at a cohort chart and see every monthly cohort decaying at roughly the same rate — a pattern the top-line MRR chart never revealed.
That moment is disorienting. You thought you had a growth problem. You actually have a product-market fit problem, at least for a segment of your customer base.
The dashboard that makes this undeniable is not the one showing MRR over time. It is the one that breaks revenue into cohorts — showing how each month's customers behave over their lifetime. When you see the same curve bending down for cohort after cohort, there is no ambiguity. The aggregate number was the lie. The cohort curve is the truth.
Net revenue retention (NRR) is better than raw MRR growth, but it is not immune to the same distortion. If your NRR is 105%, that sounds solid — until you learn it is composed of 130% expansion from your top decile and 75% retention from everyone else. The average flatters the distribution.
The discipline that works: break NRR by segment. By plan tier. By acquisition channel. By company size. By use case. You are looking for the segments where retention is strong and the segments where it is weak. Those weak segments are not a growth opportunity. They are a cost center masquerading as revenue.
Once you see the cohort data clearly, the correction follows a predictable arc.
Month one: stop the bleeding. Identify the segments with the worst retention. Understand why they leave. Common reasons: the product does not solve their core problem, onboarding fails them, or they bought on a promise the product doesn't deliver yet. Do not try to save every account. Focus on the ones where the product actually fits.
Month two: fix onboarding and expectations. Most churn happens in the first 60 days. If customers are not reaching a meaningful outcome in that window, no re-engagement email will save them. Tighten the path from signup to value. Be honest on the marketing site about who the product is for — and who it is not for.
Month three: measure the new cohort. The real test is whether the cohort that signed up after your changes retains better than the ones before. If it does, you have made progress. If not, you have more work to do. Either way, you now have the feedback loop you were missing.
The takeaway is simple but uncomfortable: growth metrics are only trustworthy when paired with cohort-level retention. A rising MRR chart without a retention breakdown is a number without context. And a number without context is a guess dressed up as a fact.
Track the total. Celebrate the wins. But always check the cohorts. The revenue number that masks churn will eventually stop masking it. The only question is whether you catch it on your terms or the market catches it for you.
Be the first to comment.
0 comments
Loading comments...