Annual Operating Plan

When to use

Use whenever the annual plan needs to reconcile what leadership wants (a top-down growth or margin target, often set for investor or board reasons) against what a driver-based bottom-up model actually produces. Especially use when the current "plan" is really just the top-down number with a supporting narrative built backward to justify it, rather than a genuine reconciliation.

What it does

Compares top-down target and bottom-up build line by line via a bundled calculator, computes the gap and gap percentage for every line item, and flags any line with more than a 10% gap — forcing an explicit resolution (revise the target, find real drivers to close the gap, or accept it with a named owner) instead of letting the two numbers get quietly blended into a plan nobody actually built.

Method

1. Build the bottom-up plan first, independently. Use driver-based-budget-builder's discipline: real volume, price, and cost drivers by department, not a top-down number divided by twelve.
2. State the top-down target explicitly — the number leadership or the board has in mind, and why (a growth rate expected by investors, a margin target tied to a prior commitment).
3. Run the bundled calculator (scripts/aop_reconciliation.py) to compare every line item, revenue, margin, opex, and any other planned metric, top-down versus bottom-up, with the gap and gap percentage shown.
4. Treat every flagged gap (over 10%) as requiring one of three resolutions, not a rounding error to smooth over: revise the top-down target to something the bottom-up model can actually support, find specific additional drivers (a new channel, a price change, a cost action) that would close the gap for real, or explicitly accept the gap with a named owner and a plan to close it during the year.
5. Never let the published plan silently split the difference between top-down and bottom-up without documenting which resolution was chosen and why — a plan that's an unexplained average of two numbers is not a real plan, it's a compromise nobody can defend later.
6. Revisit the reconciliation at each planning checkpoint (quarterly re-forecast), not just once at the start of the year — the gap-closing actions from step 4 need to actually show up in the numbers, or the same gap reappears next cycle.