Retention: GDR, NDR, and Loss Composition, The Certified Data

GDR, NDR, and the Expansion Premium

Retention Dispersion

Source: 2025 Survey, p. 9

The 75th-percentile NDR, the industry’s best quartile boundary, fell from 113% to 106% between 2022 and 2024. In 2022, a company at 106% NDR was the median; by 2024, that same performance put it at the top-quartile line. The bottom quartile sat at 95%, outright net contraction, for two consecutive years. Meanwhile the 25th-percentile GDR improved from 78% to 82%, which is consistent with a selection effect: the worst gross retainers of 2022 may simply no longer be in the respondent pool. Inference

2024 Gross-Loss Composition

Source: 2025 Survey, p. 19

Churn dominates downsell in every cut, ranging from 59% of losses at best (the 20–30% growth cohort) to 76% at worst (the fastest growers, whose losses skew toward customers acquired quickly and lost entirely). Downsell is a large-company and moderate-growth phenomenon; full churn is everyone’s problem. See Downsell: The Unreported Loss for the full treatment.

Reading the Retention Gap

The two retention series moved differently, and the difference is what this section is about.Companies lose roughly 14 cents of every ARR dollar annually, and the three actual years show no improving trend. Net retention, meanwhile, fell from 106% to 101%, which means the entire deterioration happened in the expansion layer. The derived gap row makes this explicit: the expansion premium compressed from 20 points to 15 across the actuals, and measured from its 2021 peak of 23 points (2022 Survey), the cushion has thinned by roughly one-third. Inference

This section underwrites The NDR Crisis primarily, with supporting evidence for The Churn Tax and The Downsell Problem.