The quarter looks great, but something feels off
The quarter’s headline numbers can look clean: revenue up, billings solid, cash balance steady. But the unease shows up in the edges—support tickets climbing, sales cycles stretching, a bigger share of deals coming in late with “one-time” concessions. The retention line in the deck might still be packaged as “healthy,” yet the churn note gets longer each quarter, and the CFO starts emphasizing adjusted metrics a little more than last time.
That’s usually the moment to slow down. Before trusting the acceleration, it’s worth asking whether the business is getting stronger—or just running faster to stand still, replacing customers more expensively while the P&L still benefits from last year’s better cohorts.
First check: is growth coming from old customers?
Start by isolating what happened to last year’s customers before celebrating this year’s bookings. Take the beginning-of-period ARR base and reconcile it: renewals kept, dollars expanded, dollars contracted, and dollars lost. If revenue is accelerating while gross retention is slipping, the simplest explanation is often that expansion is masking leakage—price uplifts, seat true-ups, or usage overages landing on a smaller set of power users. That can look fantastic in a quarter with a heavy renewal calendar, and it can still be fragile if it’s concentrated in a few accounts or tied to one-time contract resets.
Then check the mix: how much of net new ARR came from expansion versus new logos. If expansion is doing the work, ask what it cost in timing and terms—multi-year prepay discounts, stepped pricing, or relaxed downsell clauses. Durable growth usually shows up as broad-based expansion with stable renewal outcomes, not a handful of “save” deals carrying the quarter.
When churn rises, cohorts tell the real story
Once gross retention starts drifting down, the quarterly reconciliation stops being enough. Pull the logo and revenue cohorts by start month (or quarter) and line them up: month-3, month-6, month-12 gross retention, plus expansion on the survivors. The pattern that matters is whether newer cohorts decay faster than the older ones. If the 2026 cohorts are churning earlier than the 2025 cohorts, current growth may be borrowing from a short-lived customer mix, and the income statement won’t show the full replacement burden until renewals roll through.
Watch for the practical friction points: an onboarding change that shortened time-to-go-live but increased early cancellations, a new sales segment with smaller ACVs and higher support load, or a packaging shift that pulled in “trial-like” customers. If the churn is cohort-specific, it’s also usually fixable; if it’s broad, it’s usually structural.
Separate pricing wins from discount-driven volume
The next trap is treating “pricing power” as self-evident just because ARPA moved up. In a churny quarter, ARPA can rise for the wrong reasons: the low end falls out, the survivors get repriced, and reported growth is really mix shift. So pin down what actually changed. Look for a consistent list price move applied across renewals, stable win rates, and no sudden jump in contract length. If sales is leaning on multi-year commits to pull revenue forward, that’s timing help, not pricing strength, and it usually comes with a real cost in flexibility.
Then separate unit volume from unit value. If new bookings are up but average discount is widening, “growth” may be coming from lower-quality demand that will show up later as downgrades and short renewals. A clean check is gross margin and support burden per dollar of new ARR: heavy discounting often brings noisier customers, more implementation hours, and weaker payback. If the quarter needed both higher discounts and more concessions to close, the price narrative is doing too much work.
Replacement growth: does CAC quietly spike alongside churn?
After you’ve separated true expansion from mix shift, the next question is whether the company is paying more just to keep ARR flat underneath the headline growth. Rising churn often forces “replacement growth”: more new bookings are needed to offset what’s walking out the door, and sales effort quietly shifts from compounding to refilling. The friction shows up as higher quota burden, heavier SDR coverage, and a pipeline that looks busy but less productive per rep.
So treat CAC like a moving target, not a single ratio. Compare last year’s CAC payback to the last two quarters, using a consistent definition of sales and marketing spend and a realistic gross margin assumption. If gross retention is down and CAC payback is stretching at the same time, the business may be buying its growth. A clean signal is new-logo ARR per sales head flattening while marketing spend rises—especially if it’s happening in the same segments where the newer cohorts are churning fastest.
Find the churn driver that actually changes outcomes
At this point the question stops being “why did churn tick up?” and becomes “which reason, if removed, would actually move gross retention back up?” Renewal notes are noisy, so force a small taxonomy: non-payment and forced downgrades (involuntary), budget cuts, failed implementation, missing features, security/legal blocks, competitive displacement, and “no longer using.” Then quantify two things for each bucket: ARR at risk and reversibility within the next two renewal cycles. A big bucket that’s structurally irreversible is a different risk than a small bucket that’s fixable but under-resourced.
The fastest tell is time-to-churn. If losses cluster in months 1–3, it’s onboarding, ICP, or over-discounted “trial” deals. If they cluster at month 12+, it’s value realization, product gaps, or competition. Tie it back to observable friction—implementation hours per $ of ARR, support load, and the share of renewals requiring exec escalation—and the churn story stops being narrative and starts becoming forecastable.
Stress-test forward revenue without rosy net retention
The forecast gets fragile when it assumes net retention will “normalize” while the churn evidence is still worsening. So strip the model back to gross retention first. Take the starting ARR base and run a simple forward roll with three cases: current gross retention, 2 points worse, and 2 points better—then layer in expansion only on the surviving base at the recent, not peak, rate. Keep timing honest: expansion usually lags renewals, and big reprices don’t repeat every quarter.
Now force new bookings to do the heavy lifting. If the company needs materially higher new-logo ARR just to offset churn, check whether the go-to-market engine can realistically produce it without exploding CAC payback. A “beat and raise” quarter can coexist with a forward year that is mathematically tight once the base stops compounding.
Decision rules: what would make you buy anyway?
After the stress test, the decision usually comes down to a few hard gates. I’ll buy anyway if gross retention has a credible floor (even if net retention is noisy), and the bad cohorts are already shrinking as a share of ARR. I also want proof the churn driver is being attacked with a timetable: fewer escalated renewals, lower implementation hours per dollar, or a clear ICP cut that actually shows up in recent logo cohorts.
Then I need unit economics that don’t rely on hope: CAC payback stable or improving on the last two quarters, and new-logo ARR per rep not deteriorating under higher discounting. If those hold, I can underwrite slower growth without underwriting a surprise margin hole.
