Home › Case Studies › Bookings vs. Revenue Reconciliation Model

Bookings vs. Revenue Reconciliation Model

A 20-deal worked model that bridges sales bookings to recognized revenue line by line — including amendments, partial-month starts, and ramped contracts. Built to end the quarterly "but Sales said we booked X" debate between RevOps and Finance.

The problem

Every quarter Sales would report a bookings number. Every quarter Finance would report a revenue number. The two never matched, and the conversation about why turned into a forensic accounting exercise instead of an operating discussion. Multi-year ramps, mid-contract amendments, and partial-month starts each introduced a different bridging adjustment, and nobody had documented the rules in one place.

The approach

Pick 20 representative deals covering the full taxonomy: clean new-business, multi-year with annual ramp, mid-term upgrade, mid-term downgrade, partial-month start, dollar-for-dollar swap, and the awkward stuff. Walk each deal from contract date through monthly recognition through the end of the term, applying ASC 606 where relevant. Build the model so that bookings → ARR → MRR → recognized-revenue is one continuous calculation with all bridging items called out. Automate variance analysis between forecasted and actual recognition.

The outcome

• A single shared model that Sales and Finance both trust as the source of truth for any deal-level question.
• Faster quarter-close — the bridging walk is automated, not rebuilt from emails and slack threads.
• Visibility into variance drivers (amendments, partial-period starts, ramp inflection points) so Finance can flag forecasting risk earlier.
• A documented policy for the edge cases that previously required a 30-minute phone call to settle.

What I would do differently

Treat the model and the policy doc as a single deliverable, not two artifacts. The first version shipped with a clean spreadsheet but the rules for "when does an amendment trigger re-bridging" were buried in cell comments. Finance and Sales had different mental models for the same edge case for another quarter before we wrote the policy in plain English.