Yagnesh RamalingamThe 30-Day XP&A Reset
A four-week sequence for a finance and ops pair who are tired of reconciling last month before they can plan next month.
At a glance
- Time
- Four weeks, one step a week
- Who runs it
- A finance lead and an operations lead, working as a pair
- You leave with
- One close number, with an owner and a window in which it can change
- Five metric contracts
- A list of retired shadow forecasts
- Three recurring questions on a sourced path
How it works
Read this first. Then fill it in.
Best run with a finance lead and an operations lead, working as a pair.
XP&A fails in a predictable order: contested actuals, then a forecast that becomes a political artefact, then a planning platform that distributes the disagreement faster. The reset takes them in the same order, because each step stands on the one before: settle the actuals, then the metrics and the forecast built on them, and only then how answers reach people.
Four weeks
- Week 1: freeze actuals. One close number, one owner, one window in which it can change.
- Week 2: write down the five metrics planning leans on most, each as a contract with an owner.
- Week 3: retire the shadow forecast. If a number is not in a contract, it does not enter the room.
- Week 4: put the first three recurring questions on a sourced path, answered where the question is asked.
It is operating work, and it has to happen whether or not you buy another platform. Each week ends with one thing written down: the close number and its owner, five metric contracts, the list of retired forecasts, and three sourced question paths. The plan below keeps all four in one place.
Before you run it
What happens in the 30-day XP&A reset?
Week 1 freezes last month’s actuals. Week 2 writes five metric contracts. Week 3 retires the shadow forecast. Week 4 puts three recurring questions on a sourced path.
Who should run it?
A finance lead and an operations lead, working as a pair. One owns the close, the other defends the operating definitions that feed it.
Do we need a new planning platform first?
No. It is operating work, and it has to happen whether or not you buy another platform. A platform on top of contested actuals distributes the disagreement faster.
The template
Plan the four weeks. One step a week.
Everything you type stays in this browser tab. Nothing is sent or saved, so print it or save it as a PDF before you close the page.
0 of 4 weeks done · Next: week 1, freeze actuals
Week 1
Freeze actuals
One close number, one owner, one window in which it can change.
0 of 3 written
Week 2
Write five metric contracts
The five metrics planning leans on most, each written as a contract with a named owner.
0 of 5 contracts
A contract counts once it passes the first four checks of the semantic layer checklist: owner, one source, a written definition and a change log.
Week 3
Retire the shadow forecast
If a number is not in a contract, it does not enter the room.
None listed yet
Week 4
Put three questions on a sourced path
Three recurring questions, answered from one source, where the question is asked.
0 of 3 live
A week counts as done only when its rows are written and ticked. A tick on an empty row counts for nothing.
Keep going. More free templates.
- Interactive template · 90 minutes
The Decision Latency Audit
A 90-minute working session your team can run without me: pick five recurring questions, time the path from ask to trusted answer, and see where the latency actually lives.
Open the template - Interactive checklist · 12 checks
The Governed Semantic Layer Checklist
The contract a metric needs before two dashboards are allowed to show it. Twelve checks. If a metric fails any of the first four, it is not ready to be on a slide.
Open the template
Work with me
Found a number nobody owns? Bring it to a call.
A free 30-minute call with Yagnesh. You'll find out if the problem is the data, the definition or who owns it.