Management Reporting Playbook

Executive summary

A management reporting playbook is a written system for how a company turns closed books into a monthly decision. The core claim of this one is simple: the variance narrative is the product. The table of actuals against budget is raw material. The product is the short written account of what changed, why it changed, whether it will persist, and what someone is going to do about it.

Most management reports fail in the same way. They arrive late, they contain forty tables, and the reader has to work out the story alone. The CEO reads the EBITDA line, the board member reads the cash line, and the lender reads the covenant line. Nobody reads the other thirty-seven tables, and nobody can tell from the report which variances the finance team believes matter.

The fix is not a better dashboard. It is a discipline with four parts:

1. A fixed monthly cycle with named owners and dated deadlines, so the pack arrives when the decisions are still open.
2. A pack structure that leads with the narrative and puts tables behind it as evidence.
3. A variance method that decomposes every material variance into price, volume, cost rate, mix or timing before anyone writes a sentence.
4. A feedback loop, so last month's action items are the first page of this month's review.

This playbook gives you each part, with templates you can copy, an illustrative EBITDA bridge with the arithmetic shown, prompts for using AI on the drafting, and a 30/60/90-day rollout. It is written for CFOs, controllers, FP&A leads and the CEOs who consume their work, in companies large enough to have a monthly close and small enough that one finance team produces everything.

When to use this

Use this playbook when:

• The monthly pack takes more than a week after close and still gets asked "so what happened?" in every review.
• Different audiences (CEO, board, lender, investors) each get a differently assembled report, and the numbers do not always agree.
• Variance commentary is written from memory at the last minute, or consists of phrases such as "timing" and "higher than expected."
• A new owner, sponsor or lender is about to impose a reporting calendar and you want to build it on purpose rather than on request.
• You are scaling from a founder-reviewed spreadsheet to a repeatable reporting process.

Do not start here if the close itself is unreliable. If account reconciliations are routinely incomplete or restated after the pack goes out, fix the close first. See monthly close acceleration: a 10-day close plan. A good narrative on bad numbers is worse than no narrative.

Diagnostic questions

Answer these honestly before changing anything. Each one points to a specific part of the playbook.

1. How many working days after month-end does the management pack reach the CEO? Is that date written down and met?
2. Who owns each page of the pack by name, not by team?
3. Does every material variance have a written explanation that names a cause, not a label?
4. Can the person who wrote the explanation show the decomposition (price, volume, cost rate, mix, timing) behind it?
5. Is there a written rule for what counts as material, or does it vary by who is reading?
6. Do the board pack, the lender package and the CEO pack tie to the same set of closed numbers?
7. What happened to the action items from last month's review? Where are they recorded?
8. How many metrics are in the pack, and could the CEO name the five that matter most?
9. When actuals miss, does the pack say whether the miss is one-off, run-rate, or reversing?
10. Is the forecast updated in the same cycle as the actuals, so the pack shows what the full year now looks like?
11. Which tables have not changed a decision in six months?
12. Who reads the narrative before it goes out, and do they have authority to send it back?

If you answered "no" or "I don't know" to questions 3, 4 or 7, you are producing reports, not managing performance. Start with the variance method below.

Framework: the narrative stack

The playbook rests on a stack with four layers. Each layer feeds the one above it, and a reader should be able to stop at any layer and still have the right level of detail for their role.

Two rules follow from the stack.

The narrative comes first. Tables are evidence for the narrative, not substitutes for it. If a table does not support a sentence in layers 1 or 2, question why it is in the pack.

Every layer ties. The number in the headline is the number in the scorecard is the number in the statements. One closed dataset feeds everything. When the board pack and the lender package are built separately from the same ledger, they drift.

The five variance causes

A variance narrative that names a cause must choose from a short list. Using the same list every month forces the analysis and makes patterns visible over time.

"Other" is not a cause. If more than a small fraction of the variance is labelled other, the decomposition is not finished. A sixth label, "one-off," is a persistence judgement rather than a cause, and it belongs in the persistence field of the template below.

The persistence test

Every variance carries a second label: will it persist? The three answers are one-off (will not recur), run-rate (will repeat at roughly this level), and reversing (timing that will come back). The persistence label is what turns a variance into a forecast input. A favorable variance that is reversing should not raise the full-year outlook. An unfavorable variance that is run-rate should lower it.

Step-by-step: the monthly reporting cycle

The cycle below runs from the last day of the month to the review meeting. Day counts are working days after month-end (WD+n). Adjust them to your close calendar, but write them down and hold them.

1. Hard close complete (WD+0 to WD+x). The controller finishes the close checklist and locks the period. Owner: Controller. Output: closed trial balance and a list of any known open items with estimated amounts.
2. Data freeze and tie-out (WD+x). FP&A pulls actuals from the locked period into the reporting model and ties totals to the trial balance. Owner: FP&A lead. Output: a tie-out sheet with zero unexplained differences.
3. Variance decomposition (WD+x+1). FP&A runs the bridges: revenue into volume, price and mix; gross margin into volume, price, mix and cost rate; operating expenses by line and by owner. Owner: FP&A lead. Output: a variance register listing every variance above the materiality rule with its size and a first-pass cause.
4. Cause confirmation (WD+x+1 to +2). FP&A sends each flagged variance to the budget owner with the decomposition attached and asks for cause, persistence, and action. Owner: budget owners (department and plant heads). Output: confirmed cause and persistence for each flagged item, in writing, within one working day.
5. Narrative draft (WD+x+2). FP&A writes the headline and the variance paragraphs from the confirmed register. Owner: FP&A lead. Output: draft narrative, no longer than two pages.
6. CFO review (WD+x+3). The CFO challenges the narrative against the numbers: does each sentence trace to a figure, does the headline match the largest driver, are the actions specific. Owner: CFO. Output: approved narrative, or a returned draft with named gaps.
7. Forecast refresh (WD+x+3). The full-year outlook is updated for the persistence labels: run-rate items flow forward, one-offs do not, reversing items shift periods. Owner: FP&A lead. Output: updated outlook with a bridge from the last version. For the method, see rolling forecast vs annual budget.
8. Pack assembly and release (WD+x+4). The pack is assembled in the fixed order, checked for ties, and released to the leadership team before the review. Owner: FP&A lead. Output: final pack, version stamped.
9. Review meeting (WD+x+5). Thirty to sixty minutes. Agenda: prior action items first, then headline, then variances, then decisions. Owner: CEO chairs; CFO presents. Output: decisions and actions recorded with owner and date.
10. Action log update (same day). Actions go into the log that opens next month's review. Owner: FP&A lead. Output: updated action log.

The calendar is the discipline. A pack released two days after the review is a history lesson. If the schedule slips, cut tables before you cut the cause-confirmation step, because the confirmation is what makes the narrative worth reading.

Materiality rule

Write the rule down. A common structure is a dollar floor and a percentage floor that must both be met, plus a short list of lines that are always explained regardless of size (for example, revenue, gross margin, headcount, and any line with a covenant or contractual consequence). Set the dollar floor by asking what amount would change a decision, and the percentage floor by asking what deviation is noise for that line, using the line's own trailing volatility rather than a number from elsewhere. Revisit the rule each quarter. A rule that flags everything is as useless as no rule.

Pack structure

The pack has a fixed order. Fixed order lets readers go straight to what they own and lets you spot a missing page.

For the board-specific version of this structure, with page-by-page guidance, see board reporting pack: structure and template. For the writing method behind sections 1 and 2, see variance analysis narrative: writing the why.

Choosing what goes on the scorecard

Pick metrics by decision, not by availability. For each candidate, ask what the CEO would do differently if it moved. If the answer is "nothing," it belongs in the appendix. Companies that try to track everything end up reviewing nothing. Design the set deliberately; the method is in KPI design: choosing 12 metrics that run the company, and the executive-facing layout is in executive dashboard design. Existing coverage of what executives usually track is in financial KPIs for executives.

Worked example: a price, volume and cost bridge

Illustrative. The numbers below are made up to show the method. They are not benchmarks and do not describe any real company.

A manufacturer reports for one month. Plan assumed 40,000 units at $100 each, with a variable cost of $60 per unit and $1,100,000 of operating expenses. Actuals: 38,000 units at $110 each, a variable cost of $66 per unit, and $1,150,000 of operating expenses.

Step 1: the plan-to-actual table

Read alone, this looks like a good month: revenue up 4.5%, EBITDA up 4.4%. The narrative has to say why, and the table cannot.

Step 2: decompose revenue

• Volume effect: (38,000 minus 40,000) times the planned price of $100 = ($200,000).
• Price effect: ($110 minus $100) times actual units of 38,000 = $380,000.
• Check: ($200,000) + $380,000 = $180,000, which equals the revenue variance.

Step 3: decompose gross profit

• Volume effect on gross profit: (2,000) units times the planned margin of $40 per unit = ($80,000).
• Price effect: $380,000 (price drops straight to gross profit).
• Cost rate effect: ($66 minus $60) times 38,000 units = ($228,000), unfavorable.
• Check: ($80,000) + $380,000 + ($228,000) = $72,000, which equals the gross profit variance.

Step 4: operating expenses and the EBITDA bridge

Operating expenses are $50,000 over plan. Suppose the budget owners confirm that $30,000 is recruiting fees to fill two open roles and $20,000 is software that was budgeted in the next quarter.

Check: 500,000 minus 80,000 plus 380,000 minus 228,000 minus 50,000 = 522,000.

Step 5: write the narrative

Here is a headline and one variance paragraph, built from the register.

Headline. EBITDA was $522,000, $22,000 (4.4%) above plan, but the result masks three larger movements: price added $380,000, unit cost took back $228,000, and 2,000 fewer units cost $80,000. The month was carried by price. Decision needed: whether to hold the new price while unit cost is unresolved.

Unit cost: ($228,000), run-rate unless the cause changes. Variable cost per unit was $66 against $60 planned, a $6 gap on 38,000 units. Plant management attributes it to input cost changes and not to scrap or labor efficiency (to be confirmed with the cost accounting detail). If the input cost holds, the gap repeats each month at the current volume, so the full-year outlook has been lowered accordingly. Owner: Head of Operations; action: confirm the cause of the $6 gap by the next review.

Notice what the narrative does. It names the largest driver, not the largest line. It reports size, cause, persistence, action and owner. It flags what is still unconfirmed instead of asserting a cause nobody has checked. And it makes a statement about the forecast. A reader who stops after the headline knows the story is price against cost, not growth.

What the example teaches

• The favorable EBITDA variance of $22,000 is the least informative number in the pack. The components, which total $22,000 only after offsetting each other, are the information.
• The volume loss of 2,000 units is the one component with no explanation yet. The next review needs a cause: lost customers, supply, or timing.
• If price is a deliberate increase, the question for the commercial team is whether the volume loss is related. See price realization rate drift for how realized price is tracked against list.

Templates

All tables below are designed to be copied into a spreadsheet or document.

Variance register

Variance paragraph structure

1. Line and size, with direction: "Gross margin was $72,000 above plan."
2. Cause, named from the list, with the decomposition: "Price added $380,000, cost rate cost $228,000, volume cost $80,000."
3. Evidence for the cause, or a statement that it is unconfirmed.
4. Persistence: one-off, run-rate or reversing.
5. Effect on the full-year outlook.
6. Action and named owner, with a date.

Review meeting agenda (60 minutes)

1. Prior action items: status of each, by owner (10 min).
2. Headline and decision needed (5 min).
3. Variances above the rule, largest first (20 min).
4. Scorecard exceptions (10 min).
5. Forecast and cash outlook (10 min).
6. New actions, owners, dates (5 min).

Reporting calendar

Metrics and KPIs for the reporting process itself

The reporting process is itself a process, so measure it. These measures judge the quality of the reporting system, not the performance of the business.

Set targets by method. Use contractual deadlines first (a lender or investor reporting date is fixed), then your own best trailing performance, then named public peers where their filing calendar is visible. Do not take a target from an article and apply it to your close.

Common mistakes

1. Writing the narrative before the decomposition. If you write "revenue was favorable on strong demand" before you know the price and volume split, you may be wrong about the cause. Decompose first.
2. Using "timing" as a cause without naming the reversal. Timing is only an explanation if you can say in which period the amount reverses. Otherwise it is a label for ignorance.
3. Explaining the net, not the components. A $22,000 favorable variance made of $380,000 of price and $358,000 of offsets is not a $22,000 story.
4. Letting finance guess causes that operations own. The analyst can compute the cost rate gap. Only the plant head can say why. Send the decomposition to the budget owner and make the response part of the calendar.
5. No materiality rule. Without one, the narrative either explains everything (and nobody reads it) or whatever the author happens to remember.
6. Separate numbers for separate audiences. When the board pack, the lender package and the CEO pack are built independently, they disagree by small amounts and every disagreement consumes a meeting. Build from one closed dataset. See lender reporting package for the lender view and LP and investor portfolio reporting for the investor view.
7. Dropping the action log. A review that does not begin with last month's actions teaches the organization that actions are optional.
8. Too many metrics. A forty-metric scorecard converts every review into a scavenger hunt. Cap the scorecard and move the rest to the appendix.
9. Restating quietly. If a prior figure changes, say so on the page, with the reason. Quiet restatement destroys trust faster than a bad month.
10. Ignoring the forecast. A report that explains the past without updating the outlook leaves the reader to do the hardest part.

AI prompt assistance

AI is useful for drafting and for challenge. It is not a source of facts about your business and it will produce confident, wrong causes if you let it. In each case you supply the data and the confirmed causes, and you check the output against the ledger before anything is released. Do not paste confidential data into a tool your company has not approved for that purpose.

1. Draft variance paragraphs from a confirmed register.

Below is a variance register with plan, actual, variance, the decomposition into price, volume, cost rate, mix and timing, the cause confirmed by the budget owner, persistence, and the action. Write one paragraph per row using this structure: size and direction; cause with the decomposition figures; persistence; effect on the full-year outlook; action with owner and date. Use only the information in the register. If a cause is marked unconfirmed, say so in the paragraph. Do not add any figure or explanation that is not in the register.

Check: tie every dollar figure in the output to the register. Delete any sentence that states a cause the budget owner did not confirm.

2. Challenge the narrative.

Here is the draft headline and variance narrative for this month. Act as a skeptical board member. List the five questions you would ask, the sentences that assert a cause without evidence, and any place where the headline and the largest variance disagree. Do not rewrite the narrative.

Check: answer each question yourself from the underlying data. Add answers to the narrative where a reader would want them.

3. Build the bridge arithmetic for review.

I will paste plan and actual units, price, unit cost and operating expenses. Compute the volume, price and cost rate effects on gross profit and the bridge from plan EBITDA to actual EBITDA. Show every formula and confirm that the components sum to the total variance. Flag any case where the components do not reconcile.

Check: recompute one component by hand and confirm the bridge sums to the reported variance. If you have mix, specify the mix method in the prompt, because there is more than one defensible way to calculate it.

4. Convert the pack to an executive summary.

Here is this month's approved narrative. Write a 120-word summary for a CEO who will read only this: result versus plan, the two largest drivers, whether they persist, and the decision needed. Use only facts in the narrative.

Check: confirm every number against the approved narrative before sending.

Implementation guide

Days 1-30. Take last month's pack and rewrite its first two pages in the new structure. This is the cheapest test of the method. Then write the calendar, agree the materiality rule with the CEO, and name the budget owner for each P&L line. Do not rebuild the model yet.

Days 31-60. Run the full cycle once for real. The first run will expose slow steps, usually cause confirmation and tie-out. Fix those before adding pages. Define each scorecard metric in writing (numerator, denominator, source, owner) so the number means the same thing every month.

Days 61-90. Run the cycle a second time, on the calendar, and compare. Add the forecast refresh and the reconciliation across audience packs. Retire tables nobody used. At day 90, review the process metrics against your own first-run baseline.

Governance: the CEO owns the standard (what the pack must answer). The CFO owns the output and signs the narrative. The FP&A lead owns the cycle. Budget owners own their causes and actions. The Controller owns the underlying close.

Downloadable assets

The variance register, the variance paragraph structure, the review agenda and the reporting calendar above are designed to be copied into a spreadsheet or document as they are. No calculator for this topic exists on the site today. The EBITDA bridge in the worked example uses the formulas shown, which you can reproduce in a spreadsheet. For a related tool that models margin levers, see the EBITDA improvement calculator.

Related resources

• Board reporting hub: the full set of pages on board, lender and variance reporting.
• Board reporting pack: structure and template: page-by-page structure for the board version.
• Variance analysis narrative: writing the why: the method for sections 1 and 2.
• Lender reporting package and LP and investor portfolio reporting: the external audiences.
• KPI design: choosing 12 metrics that run the company and executive dashboard design: choosing and presenting the scorecard.
• Monthly close acceleration: a 10-day close plan: the upstream step that sets your release date.
• Operating review system: how the monthly review fits a wider cadence.
• Financial KPIs for executives, rolling forecast vs annual budget and EBITDA bridge and continuous tracking: supporting method.

Frequently asked questions

What is a management reporting package?

A management reporting package is the monthly set of documents that explains financial and operating performance to company leadership. It normally holds a headline, a variance narrative, a KPI scorecard, financial statements, cash and forecast pages, and an action log. The test of a good one is whether a reader can tell what changed and what to do about it without further questions.

What should a monthly management report include?

A monthly management report should include a short headline, written explanations of material variances, a capped KPI scorecard, the P&L with an EBITDA bridge, cash and liquidity, working capital, an updated full-year outlook, and an action log. Statements and schedules belong in an appendix. The narrative leads, and tables serve as evidence for it.

How do you write a variance analysis?

Decompose each material variance into volume, price, cost rate, mix and timing, confirm the cause with the budget owner, then write one paragraph that states the size, the cause, whether it persists, the effect on the forecast, and the action with an owner. Check that the components add up to the total variance before writing.

How long should the monthly reporting cycle take?

The cycle should end with a review meeting within a few working days of the hard close, and the dates should be written down and held. The right number depends on your close, so use your own trailing best and any lender or investor deadline as the target. If you must cut something to keep the date, cut tables, not the cause-confirmation step.

Who should own the variance narrative?

The CFO owns and signs the narrative, the FP&A lead writes it, and the budget owners confirm the cause of each variance in their area. This split matters because finance can compute a gap but usually cannot explain why operations produced it. Without named confirmation from the owner, the narrative becomes a guess that looks like a fact.

How this works in a System of Performance

A reporting cycle only improves if each month's causes, actions and outcomes stay linked to the next review. In most companies those live in separate spreadsheets, emails and slide decks, so the review starts from memory. A Performis routine can take on a recurring job like this, with the scope and fit agreed with you: it works from the records your team supplies, prepares the work for your named reviewer, and keeps each approval on record. To see where your own reporting loop breaks, bring it to us as a launch partner.