Grace Williams·
Wrote our month-end ops close as a runbook with owners and cutoff times — first month ever we finished before the 3rd
Converts a braindump of month-end tasks into a sequenced runbook with owners, cutoff times, dependencies, and a parallel-work plan.
Month-End Close Runbook Builder
You are an operations planner converting a chaotic month-end close into a sequenced runbook. Our close currently "happens to us" — tasks surface from memory, dependencies collide, and it drags into the second week.
Braindump of everything that happens at month-end (unordered, from the team):
{{tasks_braindump}}
Team members and their roles: {{team}}
Hard external deadlines (payroll, filings, board reporting): {{deadlines}}
How long the close currently takes: {{current_duration}}
Build the runbook:
1. **Normalize the braindump.** Merge duplicates, split compound tasks ("reconcile and report" is two), and flag anything that isn't actually month-end work (weekly tasks hiding in the pile — they leave the runbook).
2. **Map dependencies honestly.** For each task: what must be COMPLETE before it starts, not what usually happens before it. Distinguish true dependencies from habit-order — habit-order is where the parallel time hides.
3. **Assign each task:** owner (role, with named backup), day (D-2 to D+3, where D is the last business day), cutoff time where a hard deadline demands one, and expected duration.
4. **Find the critical path** — the dependency chain that determines total close length. Mark it. Every optimization conversation starts here, so also state which single task, if shortened, shrinks the whole close.
5. **Build the checklist artifact:** the final runbook as a table (day | task | owner | backup | depends on | done-signal), where each done-signal is observable — "posted in #close channel" not "finished".
6. **Pre-close preparation list:** everything movable to D-5 or earlier that we currently do during the close.
End with the two process changes that would remove the most calendar days, with their cost.
Ergebnisse
## Normalized task list — 31 braindump items → 22 runbook tasks
Merged: three versions of "chase missing receipts" into one task. Split: "reconcile banks and update cash report" (two tasks, different owners). **Evicted from the runbook:** subscription audit (quarterly, not monthly), CRM cleanup (weekly task hiding here), "update the wiki" (real, but not close-blocking — moved to D+5 follow-up).
## Critical path (marked ▶)
▶ Receipt collection cutoff → ▶ credit card reconciliation → ▶ expense categorization → ▶ P&L draft → ▶ management review → ▶ board pack numbers. Total: 4.5 working days — this chain IS your close length.
**The single highest-leverage task: receipt collection.** It currently ends D+2 because chasing starts D+0. Move the chase to D-5 with an automated reminder, cutoff D-1 at 17:00, and the whole path shifts left by two days. Nothing else on the list comes close.
## Runbook
| Day | Task | Owner | Backup | Depends on | Done-signal |
|---|---|---|---|---|---|
| D-5 | Receipt chase reminder goes out | Ops Coord | Office Mgr | — | Reminder posted in #team |
| D-2 | Vendor invoice cutoff | Ops Coord | Finance Asst | — | AP inbox zero, posted in #close |
| D-1 17:00 | ▶ Receipt cutoff (hard) | All staff | — | reminder | Missing-receipt list posted |
| D-1 | Payroll final check | Finance Asst | Ops Coord | — | Payroll provider confirmation email |
| D+0 | ▶ Bank reconciliation | Finance Asst | Ops Mgr | — | Recon sheet ✓, posted in #close |
| D+0 | ▶ Credit card reconciliation | Ops Coord | Finance Asst | receipt cutoff | All lines matched or flagged |
| D+1 am | ▶ Expense categorization | Finance Asst | Ops Coord | both recons | Uncategorized = 0 in accounting tool |
| D+1 | Revenue recognition entries | Ops Mgr | Finance Asst | bank recon | Entries posted, screenshot in #close |
| D+1 pm | ▶ P&L draft | Ops Mgr | — | categorization + revenue | Draft link in #close |
| D+2 am | ▶ Management review (30-min slot, pre-booked) | Ops Mgr + MD | — | P&L draft | Review meeting notes posted |
| D+2 | ▶ Board pack numbers final | Ops Mgr | — | review | Numbers locked in board doc |
| D+2 | Accruals + prepayments | Finance Asst | Ops Mgr | categorization | Schedule updated |
*(non-critical-path tasks D+0–D+3 omitted here for length — full table has all 22)*
## Pre-close preparation (move to D-5 or earlier)
- Recurring journal templates prepared and dated
- FX rates table updated
- Prior-month open items reviewed so they don't ambush D+1
- Review meeting calendar slot booked for D+2 (currently scheduled ad hoc, costing 1–2 days of waiting)
## Two changes that remove the most days
1. **Receipt cutoff at D-1 with a D-5 automated chase** — removes ~2 days. Cost: one automation in your expense tool + a one-time policy announcement that late receipts roll to next month (needs MD backing the first time someone tests it).
2. **Pre-booked recurring review slot D+2** — removes 1–2 days of scheduling drift. Cost: one recurring calendar invite and the discipline not to move it.
Projected close: **D+2 close, D+3 buffer** — versus your current "sometime in week two".
Modell: Claude Sonnet 4.6
25 Likes17 SavesScore: 18
1 Kommentar
Chloe Adams·
Habit-order is where the parallel time hides — that line justifies the whole exercise. Most close delays are exactly this.