Board Reporting: What to Include and a Sample Report

It is the Thursday before the meeting. Someone has a slide open on one monitor and a billing dashboard on the other, typing a revenue figure across by hand. Someone else is in Slack asking whether the pipeline number should come from the CRM or from the spreadsheet the sales lead maintains, because the two have disagreed for three months. A third person is rebuilding last quarter's chart because the original file lives on a laptop that left the company. This is what board reporting actually looks like inside most companies under a thousand people, and it is why board packs go out late, contain numbers nobody can defend, and get read for the first time in the room.

The failure is almost never analysis. The finance team knows how to calculate gross margin. The head of sales understands the funnel. What breaks is assembly: figures are re-keyed from six or seven systems in the two days before the meeting, definitions drift quietly between versions, and when a director asks where a number came from, the honest answer is that it came from a cell somebody typed.

This guide gives you a section by section structure for a board report, a plain sample outline you can copy, and one discipline that fixes more problems than any template redesign: every number carries its source.

Why board reporting breaks at the assembly step

Sit with a board pack that went badly and trace the errors backwards. You will usually find one of three things, and none of them is a modelling mistake.

Re-keying. A number existed correctly in a system and became wrong when a human moved it. Someone copied last month's cell and forgot to update it. Someone pasted a figure that was current on Tuesday into a pack that shipped on Friday. Someone transcribed 4.18 as 4.81. This is the single most common defect in board reporting and it is entirely mechanical.

Definitional drift. The metric name stayed the same and the meaning moved. Active customers included trials in Q1 and excluded them in Q2 because a different analyst built the query. Annual recurring revenue was gross in one pack and net of churn in the next. Nobody lied. The definition simply lived in somebody's head instead of in the document.

Missing provenance. The number is correct and current and still useless, because when a director asks how it is calculated the room says "I think that comes from the CRM" and the conversation stops being about the business. A figure that cannot be traced is a figure that cannot be challenged, and a board that cannot challenge figures is not doing oversight.

All three are assembly problems. They live in the gap between systems that hold the truth and the document that presents it, and that gap is currently staffed by a person with a spreadsheet and a deadline. If your reporting process still routes through manual copies of exports, the structural fixes in Alternatives to Excel for Data Analysis and Reporting are worth reading before you redesign the template, because a prettier template built on the same manual pipeline produces prettier errors.

What a board actually needs from the pack

A board is not management. Directors are not there to run the company, and a pack written as though they were produces forty pages of operational detail that answers questions nobody asked.

The board has three jobs. It has a fiduciary duty to know the company is solvent, compliant, and honestly reported. It has a strategic duty to test whether the plan is still the right plan. And it has what governance writers call a generative duty: helping the executive team think about problems that do not yet have a shape. A good pack serves all three, and it does so by answering four questions in order.

Are we on plan? Not "how are we doing", which invites narrative. On plan, against the numbers the board approved, with the variance stated plainly.

What changed since the last meeting? Boards meet quarterly or monthly. They hold a mental model of the company that is weeks stale. The pack's job is to update that model, not to re-teach the business.

What could hurt us? Risks, with movement since last time, not a static register that has said "amber" for two years.

What do you need from us? Decisions, approvals, introductions, consents. Stated as asks, with the information required to say yes attached.

If a page in your pack does not serve one of those four questions, it belongs in the appendix or in the bin. That single test usually removes a third of a typical board meeting report.

What to include in a board report, section by section

Here is the structure that survives contact with real boards. Order matters: the argument should be readable top to bottom by a director who has fifteen minutes.

1. The one page summary

One page, prose, written by the CEO in their own voice. Three to five paragraphs: the state of the business, the two or three things that actually moved, the biggest problem, and the asks. No charts. If a director reads only this page, they should be able to hold a competent conversation.

Write it last, after the numbers are final. Written first, it becomes a wish and the rest of the pack spends twenty pages arguing with it.

2. Headline metrics

Six to ten numbers, on one page, each with the current value, the prior period, the plan, and the variance. Not thirty. A headline metrics page with thirty rows is a data dump that forces the reader to do the prioritisation you were supposed to do.

Choose metrics that a director could act on: revenue and its growth rate, gross margin, net burn, cash and runway in months, a leading demand indicator such as qualified pipeline, a retention measure, and headcount. Company shape changes the list, but the count should not creep. Every metric on this page needs a named owner and a written definition.

3. Variance against plan

This is the section that separates a real board report from a status update. For each headline metric, show plan, actual, and variance, then explain the variance in a sentence. Not "revenue was below plan" but "revenue was 6% below plan because two enterprise deals slipped from June to July; both are signed as of the 14th."

Variance analysis is the analytical core of the pack, and it is where technique matters most: choosing the right comparison base, separating volume effects from price effects, and distinguishing a timing shift from a genuine miss. The methods in Financial Data Analysis: Methods, Metrics, and Tools apply directly here, particularly the discipline of decomposing a single variance into its drivers instead of reporting one aggregate number and a shrug.

4. Functional narrative

Short sections from each function: go to market, product, engineering, people, finance. Two thirds of a page each, maximum. Each one states what changed, what is on track, and what is not. The temptation is to let each function write its own mini board pack. Resist it. Cap the length and the pack stays readable.

Sales and marketing numbers in this section should come from the CRM, not from a parallel spreadsheet, and if your team is still maintaining that parallel spreadsheet, the reason is usually that the CRM is not trusted rather than that it is missing data. What CRM Stands For and What a CRM System Really Does is a useful reset on what that system is supposed to be the record of.

5. Risks and the watch list

A live register, not a compliance artifact. For each risk: a one line description, an owner, the direction of travel since the last meeting, and what would have to be true for it to become a crisis. Three to seven items. If everything is green, you are not being honest, and boards notice.

Operational and security incidents belong here in summarised form: how many, of what severity, mean time to resolve, and any that are still open. If you do not currently have a clean count to report, the intake and classification problem is worth solving at the source rather than reconstructing by hand each quarter, which is the argument made in Incident Reporting Software: What to Look For in 2026.

6. Asks and decisions

Explicit. Numbered. Each one states the decision required, the recommendation, the options considered, and the deadline. An ask buried in paragraph four of a functional update is not an ask, it is a hope.

7. Governance and consent items

Option grants, board minutes approval, policy adoptions, statutory filings, auditor matters. Routine by design, moved through as a consent agenda so the meeting does not spend twenty minutes on items that require a vote but not a discussion. Companies with reporting obligations beyond financials, sustainability disclosures in particular, will carry a standing item here; the selection criteria in ESG Reporting Software: How to Choose the Right Platform matter once that obligation becomes an audited one rather than a narrative page.

8. Appendix

Full financial statements, cohort tables, detailed pipeline, org chart, anything a director might want to interrogate. The appendix has no length limit and no obligation to be read. That is the point: it lets the main pack stay short without hiding anything.

A sample board report outline

Here is a plain sample board report you can adapt. The numbers are placeholders, included to show the shape of a line rather than to describe any real company. Page counts assume a monthly or quarterly operating board of five to seven people.

Page 1: Cover. Company name, meeting date, period covered, distribution list, and a data as-of timestamp. The timestamp is not decoration. It tells every reader which reality the pack describes.

Page 2: CEO summary. Prose. Example opening line: "We closed the month at [ARR figure], 4% under plan, entirely because of two enterprise deals that slipped into July and have since signed. Cash is [figure] with [n] months of runway at current burn. The one thing I want your help on is the VP Engineering search, which is now six weeks behind."

Page 3: Headline metrics. A table with columns for metric, actual, prior period, plan, variance, and source. Example row: "Net revenue retention, 108%, 111%, 110%, minus 2 points, [billing system, cohort as of month end]."

Pages 4 to 5: Financial performance. Revenue by segment, gross margin, operating expense by category, net burn, cash bridge. Each with variance against plan and a one sentence explanation. Operating expense lines that repeatedly surprise the board are usually the ones assembled from manual claims and card statements, which is a fixable input problem rather than a reporting one; Expense Report Software: How to Pick the Right Tool covers what to look for if that line is the one that keeps moving.

Page 6: Go to market. Pipeline coverage, win rate, sales cycle length, new logos, expansion, churn. Sourced from the CRM with the query definition stated.

Page 7: Product and engineering. What shipped, what is next, delivery against the roadmap the board last saw, and any material technical risk. Avoid velocity metrics. Boards cannot interpret them and they invite the wrong argument.

Page 8: People. Headcount actual against plan, open roles and time to fill, attrition, any leadership changes.

Page 9: Risks and watch list. The register described above.

Page 10: Asks and decisions. Numbered, with recommendations.

Page 11: Consent agenda. Approvals, minutes, grants.

Appendix. Statements, cohorts, detail.

Eleven pages plus appendix. If yours runs to forty in the main body, the extra twenty-nine are almost certainly detail that belongs behind the divider.

The rule that fixes most board packs: every number carries its source

Adopt one rule and enforce it without exception: no figure appears in the pack without a source annotation. Not a footnote at the back. Attached to the number.

A source annotation has four parts: the system of record, the definition or query, the as-of moment, and the owner. In practice it looks like a short bracket after the metric, or a dedicated column in the metrics table. "ARR $4.18m [billing system, active subscriptions net of refunds, as of 31 July 23:59 UTC, owner: Finance]."

This looks pedantic for about one meeting. Then three things happen. Directors stop asking where numbers come from and start asking what the numbers mean, which is the conversation you wanted. Definitional drift becomes visible, because a changed definition now changes visible text. And the person assembling the pack can no longer paste a number they did not verify, because the annotation forces them to name where it came from.

The table below maps each section to the question it answers, where its numbers usually live, and the failure mode that shows up when the source is not pinned down.

That last row is worth a moment. Consents, grants, and signed agreements only become urgent during fundraising or an acquisition, at which point a missing approval is expensive. Keeping those documents in a system with real retrieval rather than a shared drive is the sort of unglamorous work described in Legal Document Management Software: A Law Firm Guide, and the principle holds for any company, not just firms.

Cadence, length, and the pre-read discipline

Three operating rules do more for board reporting quality than any redesign.

Ship the pack five to seven days ahead. Not the night before. Directors with day jobs need a weekend. A pack that arrives at 11pm the night before guarantees a meeting spent reading aloud.

Never present the pack. The meeting agenda should assume the pack was read. Open with five minutes of headline confirmation, then spend the time on two or three real discussions and the asks. If your board culture cannot support that yet, say explicitly in the invitation that the first ten minutes are silent reading. It is awkward once and normal thereafter.

Fix the cadence and the format. Same sections, same order, same metric definitions, every time. Boards read comparatively. A pack that reorganises itself each quarter destroys the reader's ability to notice change, which is the entire value of periodic reporting.

One nuance on monthly versus quarterly. Many companies send a short monthly update and hold a quarterly board meeting with a fuller pack. That works well, provided the monthly uses the same metric definitions as the quarterly. When the monthly has its own casual definitions, the quarterly becomes a reconciliation exercise and you have created work rather than reduced it.

Where Skopx fits, and where it does not

Be clear about what a tool can and cannot do here, because this category attracts overclaiming.

Skopx does not produce a formatted board deck. It is not a slide generator, not a dashboard builder, not a business intelligence platform, not a data warehouse, and not a CRM. If you want a designed pack with your typography and your chart style, you will still build that in your deck tool or your reporting layer, and you should.

What Skopx addresses is the assembly problem described at the top of this article. It connects to nearly 1,000 tools a company already uses, including Gmail, Slack, Stripe, HubSpot, QuickBooks, and Google Analytics, and it answers questions in chat with citations back to the systems the figures came from. That means the person building the pack can ask for the current month's revenue net of refunds, the pipeline created in the quarter, or the open incident count, and get an answer with its source attached rather than a number pasted from a dashboard that was current on Tuesday.

Two other pieces are relevant. The insights engine watches connected systems for anomalies and surfaces the things worth explaining before a director spots them, which is how a variance section stops being a surprise. And workflows, which are automations you build by describing them in chat, can run a recurring pre-board summary: ten days before the meeting, pull the current figures with their sources, note what moved since the last pack, and post the result where the pack author works. The deck still gets built by a human. It just starts from live data with provenance instead of from last quarter's file.

Pre-board figure pull

10 days before board meeting

Recurring schedule tied to the board calendar

Pull finance figures

Revenue, margin, burn, cash from billing and accounting

Pull go to market figures

Pipeline, win rate, churn from the CRM

Pull risk and incident counts

Open incidents by severity, time to resolve

Compare against plan

Variance per metric with the prior pack as the baseline

Attach source to every number

System, definition, as-of timestamp, owner

Draft the summary

What moved, what is off plan, what needs explaining

Post to the pack author

Delivered to Slack with citations intact

On the commercial side, Skopx is Solo at $5 per month and Team at $16 per seat per month, and it uses your own AI provider key with zero markup, so model spend goes to the provider rather than through a reseller margin. Full detail is on the pricing page. For a finance or operations team whose real cost is two days of senior time per board cycle, the arithmetic is usually straightforward, but the honest framing is that this removes the fetching and the transcription, not the judgement. Someone still has to decide what the numbers mean and write the summary that says so.

The same discipline applies to cost lines that surprise boards repeatedly. Infrastructure and data platform spend is a common offender, and it is rarely a reporting failure so much as an absence of anyone watching the trend between meetings; Retail Data Platform Cost Control: Where the Money Goes walks through where that spend hides.

Common board reporting mistakes

Treating the pack as a performance. A board report that only contains good news trains the board to distrust it. The most useful packs lead with the problem.

Letting metric definitions live in queries instead of in writing. If the definition is only in SQL, it will change when the analyst changes.

Reporting activity instead of outcomes. Meetings held, features shipped, campaigns launched. Boards want the result, not the effort.

Burying the ask. If you need a decision, it goes on the asks page with a recommendation, not in a paragraph on page nineteen.

Using the appendix as a dumping ground with no index. An unindexed hundred page appendix is the same as no appendix.

Rebuilding the pack from scratch each cycle. Structure should be fixed, content should change. Rebuilding is how definitions drift.

Assembling in the last forty-eight hours. Every failure mode in this article gets worse under time pressure. Move the assembly earlier and automate the fetching, and most of them stop happening.

Frequently asked questions

What should be included in a board report?

At minimum: a one page CEO summary, six to ten headline metrics with plan and variance, a financial section, short functional updates, a live risk register, and an explicit list of asks and decisions. Everything else belongs in an appendix. A useful test is that every page in the main body should answer one of four questions: are we on plan, what changed, what could hurt us, and what do you need from us.

How long should a board report be?

Ten to twenty pages in the main body, plus an unlimited appendix. Length is not the real constraint; readability under time pressure is. If a director with fifteen minutes cannot follow your argument from the summary and the metrics page, the pack is too long regardless of its page count.

How far in advance should the board pack go out?

Five to seven days before the meeting. That gives directors a weekend to read and lets them send questions in advance, which turns the meeting into a discussion rather than a recital. Sending the night before is the most reliable way to guarantee that nobody has read it.

What is the difference between a board report and a board deck?

In practice the terms are used interchangeably, but the useful distinction is format and purpose. A board report or board pack is a written document meant to be read in advance, with prose, tables, and an appendix. A board deck is the presentation material used in the room. Most well run boards send a full pack as the pre-read and use very few slides live, because slides compress an argument into bullets and boards need the argument.

How do you present bad news in a board report?

Early, plainly, with the cause and the plan. Put it in the CEO summary rather than letting a director find it on page fourteen. State what happened, why, what you are doing about it, and what you need. Boards forgive misses far more readily than they forgive discovering a miss themselves, because the second one is evidence about the reporting, not just the result.

Should board report numbers be audited?

Monthly and quarterly operating figures are management numbers and are not audited. That is normal and expected. What matters is that they are consistently defined, traceable to a system of record, and reconciled to the audited statements at year end. Stating the source and the as-of timestamp next to each figure is what makes that reconciliation possible without a forensic exercise.