Will GitLab's dollar-based net retention rate fall below 115% in any two consecutive quarters by Q2 FY2027?

The Q4 data confirms the ~1pp/quarter decline pace (122→121→119→118). From 118%, reaching below 115% for TWO consecutive quarters by Q2 FY2027 requires the remaining two data points (Q1 and Q2 FY2027) to both be below 115%. This means Q1 FY2027 must be ≤114% — a minimum 4pp drop in one quarter from 118%. At the current 1pp/quarter pace, Q1 would be ~117% and Q2 ~116%. Even a 50% acceleration to 1.5pp/quarter yields Q1 ~116.5% and Q2 ~115%, neither below 115%. The CFO's explicit guidance of 'trend down slightly before stabilizing' is the strongest contra-signal: management is guiding for deceleration of the decline, not acceleration. Enterprise DBNRR is confirmed 'very healthy,' and the pressure is concentrated in mid-market/SMB (~20% of ARR). Gross retention at a 4-year high confirms this is an expansion shortfall, not customer loss — expansion shortfalls are more controllable and less likely to produce sudden drops.

Taking the most analytically rigorous approach to the math. For YES resolution, both Q1 AND Q2 FY2027 must be below 115%. Starting from 118% in Q4 FY2026, Q1 FY2027 must show a minimum 3.01pp decline in a single quarter — triple the current pace. No identified catalyst exists for this acceleration: (1) the price increase anniversary headwind has already been absorbed through FY2026, meaning it is NOT a new FY2027 drag; (2) gross retention is at 4-year highs, ruling out a churn-driven step-down; (3) enterprise customers, representing the majority of ARR, show healthy expansion. The only scenario for YES is a severe macroeconomic shock or sudden AI-driven seat reduction — but the committee assessed the AI headcount timeline at 2-4 years, well beyond the September 2026 resolution. The calibration data from resolved markets reinforces that we may be overestimating downside risks (revenue beat underconfidence suggests systematic conservatism). Reducing to 10%.

Assigning modestly higher probability to account for tail risks the base case may underweight. Scenarios that could produce acceleration: (1) a major enterprise customer churning or significantly downsizing in Q1 FY2027 (low probability but high impact given concentrated expansion dynamics); (2) federal government spending freeze affecting the ~12% public sector ARR more severely than guided; (3) competitive loss to GitHub or a security platform competitor causing a cluster of seat reductions. Each of these is individually unlikely but their combined probability is non-trivial. The CFO's 'stabilizing' language is guidance, not a guarantee — CFOs have been wrong before, especially new CFOs (Jessica Ross took the role recently). However, even accounting for these tails, the two-consecutive-quarter requirement is extraordinarily stringent. A single bad quarter followed by a normal quarter would not resolve YES.

The math is now nearly dispositive. From 118%, the two-consecutive-quarter requirement demands both Q1 and Q2 FY2027 below 115%. The most favorable conceivable scenario for YES: Q1 drops to 114% (a 4pp decline) and Q2 drops to 113%. This would require a quadrupling of the decline rate in Q1 followed by continued acceleration. There is no historical precedent for this in GitLab's DBNRR history, and multiple structural supports argue against it: enterprise segment health, 4-year-high gross retention, Duo Agent providing an eventual (if delayed) expansion vector. The segmentation data is particularly important — mid-market/SMB pressure (~20% of ARR) cannot drive a 4pp decline on its own even if those segments showed catastrophic retention. I assign the lowest probability in the ensemble at 8%, reflecting a near-zero scenario that would require a macro shock or unprecedented competitive displacement.

Weighting the persistent decline trend and structural headwinds somewhat more than the pure math suggests. The decline trajectory has been remarkably consistent: 1pp/quarter for four consecutive quarters. While this pace is insufficient for YES resolution, the consistency itself is notable — the decline has not decelerated despite management's stated efforts to reverse it. What if the stabilization management expects simply does not materialize? If Q1 FY2027 shows continued 1pp decline (117%) and Q2 shows an acceleration to 1.5pp (115.5%), we still do not breach. Even modest acceleration does not get there. However, I assign 14% rather than single digits because: (1) the 'cohort sizes getting smaller' dynamic is structural and worsening; (2) Duo Agent's inability to contribute to expansion in FY2027 means there is no new growth lever; (3) the calibration data suggests we should be less confident in any direction, not just upside.

The resolution criteria are extremely stringent relative to current data. Two consecutive quarters below 115% from a starting point of 118% within only two remaining data points means BOTH must breach — there is no room for a 'bad quarter followed by recovery' scenario. The CFO's segmentation disclosure is actually the most dispositive piece of new information: by confirming that enterprise DBNRR is 'very healthy' and the weakness is concentrated in ~20% of ARR (mid-market/SMB), management has essentially revealed the structural floor. Even if mid-market/SMB DBNRR dropped to 100% (catastrophic), its weight in overall DBNRR is limited. The enterprise majority would need to simultaneously weaken for the aggregate to breach 115%. No evidence supports enterprise weakening — $100K+ customers grew 23% YoY in FY2026.

Assigning the highest probability in the ensemble to reflect tail risk. The persistent 1pp/quarter decline has shown no signs of bottoming despite management's efforts. If anything, the transition from 119→118 occurred while management was actively trying to reverse it. The 'stabilizing' guidance from a new CFO lacks credibility — she has been in role for less than a year and may be overly optimistic. Multiple compounding headwinds remain: no pricing tailwind in FY2027, smaller new cohorts, and Duo Agent contributing nothing to expansion. If these headwinds compound rather than offset, a 1.5-2pp/quarter decline is plausible (Q1: 116%, Q2: 114%). This scenario requires Q2 below 115 AND Q1 below 115 — the Q1 step is the binding constraint.

The two-consecutive-quarter requirement remains the binding constraint. Even in the bearish case where decline accelerates to 1.5pp/quarter, the trajectory would be Q1: 116.5%, Q2: 115%. This still does not produce two consecutive quarters below 115%. You would need approximately 2pp/quarter sustained decline — roughly double the current pace. The only catalyst for this would be a macro shock affecting enterprise tech spending broadly, not a GitLab-specific dynamic. The gross retention at 4-year highs is the most important contra-signal: companies in accelerating DBNRR decline typically show retention deterioration first. GitLab's retention is improving, meaning the expansion shortfall is the sole driver — and expansion dynamics are more variable and manageable than churn dynamics.

Splitting the difference between the structural case against YES and the uncertainty inherent in two unreported data points. The math strongly favors NO: from 118%, reaching below 115% for two consecutive quarters requires an unprecedented acceleration. But two earnings reports remain between now and resolution, and each could surprise. The committee's cross-lens consensus on DBNRR as the single most material metric adds weight to monitoring — if Q1 shows any acceleration beyond 1pp, the conditional probability of YES rises significantly. However, the base rate of SaaS companies showing sudden multi-point DBNRR drops without corresponding churn increases is very low. 15% reflects genuine uncertainty about two unknown data points while respecting the mathematical implausibility.