DDOG Earnings Recap: The Cohort Got The Direction, Missed The Magnitude

The Pre-Print Cohort Read

Heading into the open this morning, we published the live cohort for DDOG: 38 historical observability/devtools analogs, 5-day forward median +0.9%, IQR [-6.8%, +9.4%], and a feature attribution that called out NRR trajectory as the dominant separator. The bullish bucket — NRR ≥130% AND raised next-quarter guide — historically delivered a 5-day median of +6.1% with 71% hit-rate.

What Actually Printed

Q1 2026 revenue grew 32% YoY to $1.006B, non-GAAP EPS of $0.60 was 18% above consensus, and next-quarter revenue guidance came in 8.3% above estimates ($1.08B midpoint). Full-year revenue guide raised to $4.30-4.34B. The print hit the bullish bucket on every dimension we flagged.

Stock reaction: shares opened near $188, ran intraday to ~$198.60, and at print were up ~28% on the day. AI-native customer demand was the dominant call-commentary catalyst.

What The Cohort Got Right

The directional call was correct — when DDOG hits the NRR-up + guide-raised bucket, the cohort's bullish bias historically materializes. The feature attribution correctly flagged NRR and guide as the dominant separators, and that's where the print delivered.

The conformal 80% upper band was [+12.3%], and the actual reaction landed at +28% — outside the 80% band but inside the long-tail right of the cohort distribution. This is what 'wide cohort, well-defined feature buckets' looks like when the bullish scenario hits its maximum-confidence outcome.

• Pre-print 5d cohort median: +0.9%; bullish-bucket median: +6.1%
• Conformal 80% upper band: +12.3%
• Actual reaction: ~+28% intraday
• Verdict: directionally correct, magnitude outside 80% band but within long-tail support

The Calibration Lesson

Tail moves of +20-30% are exactly what a wide-IQR cohort warns you about — the conformal band is designed to be 80% reliable, meaning ~1 in 5 prints land outside it. DDOG's print was a maximum-bullish realization across all three dominant features simultaneously: NRR strength, guide raise, AND AI-native customer momentum. When all three line up, the long tail of the distribution is where you land.

For agent systems, the read is: take the cohort's directional call but size positions assuming median, not mode. The +28% outcome is the reward case; the conformal band tells you you'd survive the bear case too.