Playbook: Calibrating a B2B Attribution Window From Your Own Data

Playbook: Calibrating a B2B Attribution Window From Your Own Data

Most partner programs inherit a 30-day cookie window from B2C tooling, then wonder why partner-sourced pipeline looks thin. The window should be derived, not borrowed. A defensible procedure:

— Step 1. Pull the lag distribution. Export first-touch-to-opportunity timestamps for the last 8 full quarters. You need closed-won and closed-lost, or you bias toward fast deals.
— Step 2. Find the cumulative-conversion knee. Plot the share of deals that touched first by day N. Per multiple SaaS benchmark sets, mid-market cycles cluster the knee around day 45-70; sub-$5k self-serve often resolves under 14.
— Step 3. Set the window where ~85% of attributable touches sit. Chasing the last 15% lengthens the window and inflates over-crediting.
— Step 4. Split by motion. A single window across PLG and enterprise blends two distributions and serves neither.
— Step 5. Document the assisted vs. last-touch rule so partner payouts and finance reconcile.

Trade-off worth naming: a longer window credits more partners and improves recruitment optics, but raises double-counting against paid and SDR touches. A shorter window is cleaner for incrementality but demoralizes top-of-funnel partners.

Note the correlation/causation trap: a partner appearing in long-lag deals may be a marker of complex deals, not a driver of them.

Implications: publish the window, its derivation, and its review cadence (annually, or after any motion change) so partners can model their own economics rather than guess at yours.