Month-end reporting has two halves. The close is reconciling the bank, posting accruals and prepayments, running depreciation, checking the control accounts, and locking the period; it's accounting work, it's done in the ledger, and it should stay there. The pack is what's produced after the close: profit and loss against budget, balance sheet, cash summary and forecast, aged debtors and creditors, the operational KPIs, and the commentary. In most finance teams the close takes three to five working days and the pack takes another two to three, and it's the pack where the time goes and where automation pays back, because once the ledger is locked, everything on the pack except the commentary can produce itself.
What's on the month-end checklist, in order?
This is the list I work from with clients. Day numbers are working days after month end and are typical for a small or mid-sized UK business, not targets.
| Step | Owner | Typical day | Can a model take it over? |
|---|---|---|---|
| Bank reconciliations complete | Bookkeeper | 1 | No. Ledger work. |
| Sales and purchase ledgers posted and matched | Bookkeeper | 1 to 2 | No |
| Accruals, prepayments, deferred revenue posted | Accountant | 2 to 3 | No, but a model can flag what's missing versus last month |
| Depreciation and fixed asset movements | Accountant | 3 | No |
| Payroll journal posted | Accountant | 3 | No |
| Stock valuation and adjustments | Accountant / ops | 3 to 4 | No, but a stock report can highlight the adjustments needed |
| Intercompany reconciled (if multi-entity) | Accountant | 4 | Partly; a model can show the mismatches |
| Control accounts reviewed (VAT, PAYE, debtors, creditors) | Accountant | 4 | Partly; the check can be on a page |
| Period locked | FD / controller | 5 | No |
| P&L vs budget vs last year, month and YTD | Finance | 5 to 6 | Yes |
| Balance sheet with movements | Finance | 5 to 6 | Yes |
| Cash summary and 13-week forecast | Finance | 6 | Yes for actuals; forecast inputs need a person |
| Aged debtors and creditors with trend | Finance / credit control | 6 | Yes, if month-end snapshots are stored |
| Operational KPIs (utilisation, pipeline, orders, headcount) | Finance + ops | 6 to 7 | Yes, if the source systems are connected |
| Variance commentary | Finance + budget holders | 7 | No. A person writes it. |
| Pack assembled and distributed | Finance | 7 to 8 | Yes |
Source: my working checklist, built from the month-end processes I've seen in UK businesses; your days will differ.
Two things stand out. The close is all "no". The pack is nearly all "yes". And the pack is where the second week goes.
What should the pack contain?
The minimum that a board or an owner will read:
- One page of headlines. Revenue, gross margin, EBITDA or net profit, cash, overdue debt, one or two operational numbers. Each against budget and last year. Arrows.
- P&L vs budget vs last year, month and year to date, by the reporting lines managers own, with variances marked favourable or adverse consistently. The variance rules are in Budget vs Actual.
- Balance sheet with movements since last month and the ratios the board watches.
- Cash. Actual movement for the month and the 13-week forecast ahead. The forecast is covered at The 13-Week Cash Flow Forecast.
- Debtors and creditors. Ageing as at month end, the trend over twelve month-ends, the top ten with owners. Shape at The Aged Debtors Report.
- Operational KPIs. The three to six numbers that explain the P&L: utilisation for a services firm, orders in hand for a manufacturer, occupancy for a venue, pipeline coverage for a sales-led business.
- Commentary. What happened, why, what's being done. Written by a person, against the variances that pass the materiality rule.
I've written the pack up in more detail at What Goes in a Monthly Management Accounts Pack; this post is about the process.
What should you automate first?
In this order, because each step pays for the next:
1. The mapping from nominal codes to reporting lines. One table: nominal code, reporting line, group, sort order. It's a hundred rows, it takes an afternoon, and it's the reason the P&L can build itself. Every later step depends on it.
2. The P&L and balance sheet from the ledger. Once the mapping exists, a model reads the posted journals from the ledger (Xero, Sage, QuickBooks, Business Central all expose them; routes at How to Connect Your Accounting Software to Power BI), applies the mapping, and the two statements are measures. Budget goes in as a table at the same grain. This removes the export-and-paste.
3. Month-end snapshots of ageing and balances. A scheduled copy of the aged debtors, aged creditors and key balances at each period end into a table. Cheap to do, and after six months it's the trend chart the board asks for.
4. The reconciliation page. The model's P&L total next to the ledger's, difference shown. Zero or it doesn't go out. This is what makes the automated pack trustworthy and it's the step that's most often skipped.
5. Operational KPIs from the other systems. CRM, job system, timesheets, stock. Each one is a second source and a join on a shared key. Do these after the finance side is solid, one at a time.
6. Distribution. The pack is a Power BI report in the client's workspace; the PDF export or the scheduled email is a setting. Row-level security means each department head sees their own lines without a separate pack.
What stays manual: the close itself, the forecast assumptions, and the commentary. Automating those is either impossible or a bad idea.
How much time does it save, and what does it cost?
I'm not going to print a percentage; I don't have a survey and I won't invent one. What I can say from the packs I've built: the second week of the month (the pack) collapses to the commentary day once the ledger is closed, because the numbers are already on the page. The close doesn't change.
Costs: Power BI Pro is £10.80 per user per month on Microsoft's UK page as of September 2026, for everyone who opens the pack. The build, if the ledger is the only source (steps one to four above), is a single-source job and it's what the First Report is for: one system, one report, the headline number reconciled to the ledger to the penny, £1,950 fixed, ten working days, details at the First Report. Steps five and six with a second system are a Reporting Sprint.
Common questions
What is month-end reporting?
The process of closing the accounting period (reconciling, posting adjustments, locking) and then producing the management pack: P&L against budget, balance sheet, cash, debtors and creditors, KPIs and commentary. The close is ledger work; the pack is reporting.
How long should month-end take?
For a small or mid-sized UK business, the close is typically three to five working days and the pack another two to three, so the pack lands in the second week. Once the pack produces itself from a locked ledger, the second week collapses to the commentary.
What's in a month-end reporting pack?
Headlines, P&L vs budget vs last year (month and YTD), balance sheet with movements, cash actual and forecast, aged debtors and creditors with trend, operational KPIs, and commentary. The detail is at What Goes in a Monthly Management Accounts Pack.
Which month-end steps can be automated?
Everything after the period lock except the commentary and the forecast assumptions: the statements, the variances, the ageing and its trend, the KPIs, and the distribution. The close itself (reconciliations, accruals, depreciation) stays in the ledger and with the accountant.
What should we automate first?
The mapping table from nominal codes to reporting lines. It's the foundation for the P&L and balance sheet producing themselves, and it takes an afternoon. Then the statements, then month-end snapshots, then the reconciliation page.
Does automating the pack mean the numbers can't be trusted?
The opposite, if the reconciliation page is there. A pack that shows its own tie-out to the ledger every month is more trustworthy than one assembled by hand, because the check is visible. A pack without that page shouldn't go to a board, automated or not.
Do we need Power BI for this, or will Excel do?
Excel with Power Query can do steps one to four for a single entity, and for some businesses that's the right answer; see Power BI vs Excel: When the Spreadsheet Is the Right Answer. Power BI wins when the pack is shared with people who should only see their lines, when there's a second system, or when the trend history needs storing.