The Board Pack Is Three Weeks Old the Day It Lands

A mid-market manufacturer can tell you, to the hour, how many enquiries came in last week, which rep is top of the board, and how the pipeline moved since Monday. Ask the same business what it actually earned last month — after cost of goods, after the payroll and distribution it carried, after the commission and receivables still sitting unpaid — and the precision disappears. You get an estimate, or a promise to check with finance. The pipeline is instrumented to the minute. The money is not.

For years the accepted answer — and I held it too — was that a capable ERP and a decent accounting system were enough. Add a stack of Excel templates on top, export the trial balance monthly, and the management pack falls out the other side. Buying anything more looked like solving a process problem with software. I was half right. The tooling was rarely the binding constraint. What I underestimated was how much of the month a small finance team loses to moving data by hand — pulling the ledger out of QuickBooks, the deals out of the CRM, volumes and cost out of the operational systems, and reconciling all three into one version before anyone can even ask a question of it.

By the time that work is done and the board pack is printed, it is describing a business that no longer exists. The quarter has moved on. Margin has already eroded on the last three weeks of shipments; cash has already been lent to a slow-paying customer; a capital project has already drifted past its budget. The pack is a museum piece — accurate about a moment that has passed, silent on the one you are standing in. That is not a reporting problem. It is a data foundation problem wearing a board-pack costume.

The practical shift came with Microsoft Fabric. For an owner-operator the part that matters is unglamorous and decisive — one governed place where the general ledger, the CRM and the operational data land together, in a form analytics can read directly.

Azure Data Factory pulls the source data on a schedule. For most mid-market industrial businesses that means QuickBooks or Xero for the ledger, a CRM such as Microsoft Dynamics 365 or HubSpot for the pipeline, and the operational systems that hold volume, cost and delivery. Larger estates swap the accounting source for SAP S/4HANA or SAP ByDesign — the pattern does not change. It all lands in OneLake, the single store underneath Fabric, and the reconciliation that used to happen every month by hand happens once, in the pipeline.

A semantic model then defines the financial logic in one place — gross margin, contribution, EBITDA, the cash conversion cycle, weighted pipeline, customer net margin — so every view draws on the same definitions rather than each analyst’s private version in a private spreadsheet. Direct Lake lets Power BI query that data without the overnight refresh cycle that used to make “live” a generous word, and Power Automate chases the receivables and flags the exceptions that need a human. The management pack stops being a monthly artefact rebuilt by hand and becomes a screen that recomputes as the underlying figures change.

None of this removes the controller’s judgement. It removes the three weeks of plumbing that used to sit in front of the judgement.

Abstract architecture is easy to nod along to and hard to picture. So here is a working example — a financial command centre built on the same Microsoft stack, with illustrative figures, styled the way it would sit inside your own Power BI tenant.

Move through the eleven screens. The Executive Summary carries the headline: revenue of USD 21.2M against a USD 21.1M budget and USD 18.9M last year, EBITDA of USD 1.55M against a USD 1.42M plan, and cash generation trending beside it. Then drill: the P&L with variance to budget and prior year, OpEx by category, capital projects with IRR and payback, the balance sheet with liquidity and covenant status, ROIC broken down the DuPont way, the cash flow statement with the working-capital cycle, the forecast against budget, and — on the same screen, not in a separate tool — marketing spend and cost per lead, weighted pipeline, and customer economics with the accounts most at risk. Click through all eleven screens — or open the interactive version to change the period and segment and watch every figure, variance line and trend recompute. This is the shape of a management pack when the data underneath it is live rather than rebuilt each month.

The reason this belongs on one screen is that the questions the three seats ask are not separable. The CEO wants to know whether the business is compounding. The CFO wants to know whether it can fund itself. The MD wants to know which product, customer and project to push next week. In most businesses those three answers live in three different spreadsheets maintained by three different people, and they disagree — because they were built off different extracts on different days.

Profitability that moves weekly, not quarterly. In an industrial business, gross margin shifts with input cost, mix and discounting — week to week, not quarter to quarter. Seeing EBITDA a month late means the erosion is already booked before anyone can price or negotiate against it. On a live pack, the variance to budget is visible while it is still actionable.

Working capital — DSO, DIO and DPO, and the cash conversion cycle they add up to. Cash trapped in slow-moving stock or overdue receivables is the most expensive money on a mid-market balance sheet, and it is usually invisible until the overdraft tightens. Computed live against operational reality, the cycle shows where cash is trapped while you can still release it.

Cash and covenant headroom. A forecast of the cash position, the runway it implies, and the distance to any banking covenant — read from the same governed numbers as everything else, not a separate treasury sheet that is a fortnight behind.

Return on capital, the DuPont way. ROIC and ROE decomposed into margin, asset turnover and financial leverage tells the owner which lever actually moves returns — and, as often as not, the answer is that a one-point margin gain does more than any amount of chasing revenue. Capital projects sit beside it with their IRR, payback and spend-to-date, so the investment conversation runs on returns rather than enthusiasm.

The commercial engine — on the same screen as the money. Marketing spend and cost per lead, weighted pipeline against target, and customer economics that step gross revenue down to net margin per account and flag the ones ageing or at risk. This is where a command centre earns its name: the demand-generation spend, the pipeline it sources, and the margin it eventually books stop living in three disconnected tools and start being read as one chain.

This is where most vendors go quiet, so I will not. A command centre is only as honest as the systems beneath it, and there are three places it breaks that you should hear before you commission one.

The chart of accounts and the CRM stages have to mean something. If QuickBooks has drifted into forty overlapping expense categories, or the CRM lets a rep park a dead deal in “negotiation” for a year, the command centre will render that mess faithfully and at speed. The unglamorous mapping and hygiene work at the start is not optional — it is most of the value. We diagnose that before we build, every time.

A live view is not a substitute for the close. Direct Lake gives you management numbers that recompute through the month; it does not perform your period-end accruals, your revenue recognition judgements or your statutory close. The command centre is the steering wheel between closes. The controller still signs the accounts, and should.

Multi-entity and multi-currency consolidation is where the modelling gets real. Two entities on one accounting system is straightforward. Five entities across QuickBooks, SAP and a local package, in three currencies with intercompany eliminations, is a genuine data-engineering exercise — doable, but not a weekend job, and anyone who tells you otherwise is selling. And the forecast is only ever as good as the CRM discipline feeding it; garbage pipeline in, confident-looking forecast out.

The shift is not that you get prettier charts. It is that the speed of the business stops depending on who happens to be looking. Today, the margin problem gets caught when a diligent analyst notices it in next month’s pack. On a live command centre, the variance surfaces the week it happens, at the decision point, whether or not anyone was watching. Speed becomes structural rather than dependent on vigilance.

It also changes what the monthly meeting is for. When the numbers are trusted, current and shared, the meeting stops being a reconciliation exercise — an hour spent arguing about whose spreadsheet is right — and becomes a decision meeting. Which customers to keep, which developer or supplier to lean on, which capital project to accelerate or kill. Most operations leaders leave the first proper look at their own data with three decisions they did not know they were carrying.

You are not funding a learning exercise. The point of unifying the ledger, the CRM and the operational data is not the platform. It is that the person who owns the P&L can finally see it move — and act while acting still changes the outcome.

We build in four phases, and we start small on purpose. Discover: a short diagnostic on your actual estate — what the ledger is, what the CRM is, where the operational data lives, and which three numbers are hurting most. Prototype: one governed view built on your real data, not a template, so you can judge it against reality rather than a demo. Deploy: the command centre wired into Fabric with the refresh cadence and access your team needs. Expand: the next entity, the next forecast, the next automation.

The promise is first value in six weeks — a working view of your own numbers, not a fifty-slide roadmap. It runs on the Microsoft stack you likely already pay for: Microsoft Fabric, Power BI, Power Platform, reading from QuickBooks, Dynamics 365 or SAP, and the systems underneath. Fixed scope or a monthly retainer, and Amit stays accountable for the delivery rather than handing it to a junior.

If the board pack in front of you is describing a quarter that has already moved on, that is the thing worth fixing first. Book 30 minutes with a MyData consultant and bring your worst report. No slides. No pitch deck. No obligation to proceed.