Test 02
The Pressure Discount Test
Do deals closed under quarter-end pressure retain and expand worse than deals that closed cleanly?
What everyone does instead
Quarter-end push is universal and unmeasured. Deals that slip are chased, discounted and escalated until they close, and the closing is recorded as a success. Whether that cohort behaves differently afterwards is not something any standard report will tell you, because the pressure signal lives in snapshot history that gets overwritten.
What to compute
- Split closed-won deals into a pushed cohort (three or more months in Commit, two or more close-date slips) and a clean cohort (fast close, no slip).
- 24-month net revenue retention for each cohort, computed from billing rather than modelled.
- The blended NRR implied by the current mix, and the NRR implied by a smaller pushed share.
The arithmetic
Pushed 3+ months in Commit, 2+ slips 30% $11.9M 94% NRR
Clean fast close, no slip 70% $27.7M 121% NRR
100% $39.6M
Blended today 0.30 x 94 + 0.70 x 121 = 112.9%
Pushed share at 15% 0.15 x 94 + 0.85 x 121 = 117.0%
Gain from mix alone 4.1 points of NRRWorked example
In one worked example the pushed cohort (30% of bookings) retained at 94% over 24 months while the clean cohort retained at 121%. Halving the pushed share moves blended NRR from 112.9% to 117.0% with no change to customer success, product or pricing.
Illustrative figures. The relationship shown is the hypothesis the test exists to check.
What changes if this is true
At most SaaS valuations, four points of net revenue retention is worth more than the bookings a quarter end push produces. If that holds in your data, the trade you make every ninety days is a bad one, and you can stop making it without changing anything about how customers are served.
What you need and how long
Weekly forecast snapshots, four to eight quarters
- Where it lives
- CRM field history or warehouse
- Days
- 1
ACV at start, at renewal, and expansion bookings
- Where it lives
- Billing
- Days
- 1
Cohort construction and NRR calculation
- Where it lives
- Analysis
- Days
- 2
Every test runs on the same two extracts, forecast history from the CRM and contract history from billing, pulled by a RevOps analyst in about an hour. Everything is measured against your own outcomes: no benchmark, no composite, no assumed uplift.