DDOG Trades With Wide Earnings Tails
Mid-cap-to-large-cap SaaS observability earnings produce some of the widest one-day reaction distributions in tech. DDOG's 12-print history (back to 2019) has a median earnings-day |move| of 8.7%, with a max of -23% (Q4 2022) and a max of +18% (Q3 2021). The cohort retrieval pulls these 12 prints plus cross-ticker analogs from MDB, NET, SNOW, and CFLT with similar customer-count and large-customer-mix profiles.
Across the combined 38-print observability/devtools cohort, the 5-day forward median is +0.9%, the 10-day median is +1.6%, and the IQR is [-6.8%, +9.4%]. That width is the cohort telling you not to bet on direction without conditioning on the right feature.
Net-Revenue-Retention Is The Lever
The feature_importance ranking from the cohort is unambiguous: net-revenue-retention (NRR) trajectory dominates everything else, including headline revenue beat magnitude. The cohort splits cleanly:
- NRR ≥ 130% AND raised next-quarter guide (n=14): 5-day median +6.1%, hit-rate 71%
- NRR 115-129% (in-line) (n=15): 5-day median +0.8%, hit-rate 53%
- NRR < 115% OR cut next-quarter guide (n=9): 5-day median -7.4%, hit-rate 22%
- Large customer ($100k+) net-add count: secondary feature, amplifies whichever NRR bucket the print lands in
The Pre-Print Setup
DDOG has drifted -3.4% over the prior 10 sessions into the May 7 print, on declining RVOL — a defensive setup. Options-implied move sits at ~9.2%, slightly above the 12-print median. The pre-drift-negative subset of the cohort has a 5-day median of +1.7% (slight upward bias from oversold conditions), but the conformal 80% band remains wide at [-9.1%, +12.3%] for the 5-day forward.
The base-rate read: this is a print where the headline alone won't tell you the direction. The cohort needs the NRR number to retrieve a tight forward distribution.
What To Listen For
On the call, the post-print cohort match weights commentary on three things: NRR trajectory (was it 'flat with last quarter' or 'modestly higher'?), AI/observability product attach rate among existing customers, and any color on the macro spend environment among large enterprise accounts. These three together explain >70% of the within-cohort 5-day return variance in the historical analog set.
Agent systems building positions off DDOG prints should not retrieve the cohort until after the NRR number is disclosed on the call — otherwise the IQR is too wide to be actionable.