Downsell

Downsell is revenue lost from customers who stay. They renew, they remain counted as retained, and they spend less than they did the year before. In the 2024 data it accounts for 32% of all revenue loss, a share that has held across the two most recent editions. It is the least-managed major loss in SaaS, for a structural reason: the two metrics every board watches are built to net it away.

The Unreported Loss

Because the customer is retained, downsell never registers as churn, yet it drains close to a third of all revenue loss. The headline retention metrics are blind to it by construction, not by oversight. A company can hold its logo-retention numbers, report a healthy renewal rate, and still be bleeding revenue through accounts that quietly shrink. Nothing in the standard reporting stack forces that loss into view.

The Loss Mix

KBCM splits every lost dollar into two buckets: churn, where the customer leaves entirely, and downsell, where the customer stays and reduces spend. In 2024 that split was 68% churn, 32% downsell. Only the churn half surfaces in the retention headlines. Gross dollar retention counts a downgraded-but-retained customer as retained; net dollar retention nets the downsell against expansion and reports a single combined figure. Neither exposes downsell on its own line. A loss worth nearly a third of the total has no dedicated metric, no owner, and no dashboard, which is precisely why it goes unmanaged.

Severity by Growth Rate

The pattern is not the one intuition predicts. Downsell’s share of loss does not climb steadily as growth slows. It peaks in the middle, at 41% for companies growing 20–30%, and is lowest for the fastest growers, at 24% above 30% growth. Slow growers sit at 31%. The fastest-growing companies lose the smallest fraction of their revenue to downsell, not the largest, which argues against reading downsell as a simple symptom of a stalling business.

Source: KBCM-2025 p19: loss attribution by ARR growthVerified

One plausible mechanism: the fastest growers are still in an active land-and-expand motion, where new bookings and expansion dominate the base and CS attention is trained on net-new logos rather than defending revenue already won, which dilutes downsell’s share even if it has not disappeared. Companies in the 20–30% band have slowed enough that expansion no longer masks a weakening account, but have not yet built the dedicated downsell tracking the “What to Measure” section below argues for, consistent with the peak sitting in the middle of the distribution rather than at either end. Inference

By ARR Segment

Across segments, downsell runs 26–39% of loss. The smallest companies (<$10M, 37%) and the largest (>$50M, 39%) are hit hardest, and the mid-market ($10M–$25M, 26%) least. The shape is a U: the ends of the size distribution carry the most downsell, the middle the least, which means no segment can treat it as someone else’s problem.

Source: KBCM-2025 p19: loss attribution by ARR segmentVerified

One plausible mechanism: the smallest companies have the least dedicated CS capacity per account, so a shrinking customer draws less proactive attention before it registers as downsell, while the largest companies carry multi-year, multi-department contracts where usage can contract inside one part of the organization well before the contract itself comes up for renewal. Mid-market companies sit in the sweet spot between the two: large enough to run dedicated account coverage, small enough that a shrinking account is still visible to the team that owns it. Inference

The Causal Claim

KBCM argues the direction can run both ways. Downsell is not only a symptom of weak growth; left uncontrolled, it drains the revenue that would fund the product and expansion that drive growth, which slows growth further. Cohort analysis adds a second point that matters for how the loss is treated: customers who downsell re-expand at materially higher rates than customers who churn outright, though their re-expansion cycles run longer. Downsell is recoverable in a way that churn is not, but only if it is seen in time. Inference

Causal direction is KBCM’s editorial interpretation of the loss-mix pattern; re-expansion effects from SuccessCOACHING cohort analysis (2023–2024)Inference

What to Measure

The remedy begins with measurement, not intervention. A loss this large cannot be managed while it is hidden inside a net number. Reporting downsell as its own line, by cohort, alongside churn and expansion rather than blended into NDR, is what converts a third of revenue loss from an invisible drain into something a team can own and act on. A retained customer spending less is a warning; the standard metrics are configured to suppress it.