One version of the numbers: how to stop arguing about the figures

Every department walks into the management meeting with its own spreadsheet, and the figures do not match. Much of the meeting goes into reconciling them, and the decision waits until someone brings back “the right number.” The causes are few and well known, and most of them are solved by management decisions.

Why it happens

When two departments report different figures for the same metric, a bad sum is rarely the reason. It is usually one of these four causes:

• Different definitions. For sales, an “active customer” may be anyone who bought this quarter; for customer service, anyone with a contract in force. Each figure is correct under its own definition, which is why they will never match.
• Scattered information. Part of it lives in spreadsheets on someone’s laptop, part in Drive, part in the ERP and part in WhatsApp and email conversations that never reach a system. Each report uses whatever is at hand.
• Different cut-off dates. Sales closes the month by order date and finance by invoice date. One report was downloaded on Monday and the other on Thursday.
• Nobody owns the figure. When a number does not add up, there is no one to ask. Each department defends its version and no one has the authority to settle it.

None of these causes is fixed by a tool alone. A dashboard that mixes different definitions shows the disagreement faster, but it does not settle it.

One definition and one owner per figure

The first step is a short document, approved by management, covering the metrics that actually drive decisions. For each one, write down:

• What counts and what does not. The rule in one or two sentences, with the borderline cases already decided.
• Where it comes from. The system or file that serves as the source, and only that one.
• Which cut-off applies. The date that governs (order, invoice or payment) and the day the period closes.
• Who answers for it. One person, by name and role.

The owner does not have to calculate the figure. The owner answers for the definition and explains the changes. When two departments disagree, management decides on the owner’s proposal, and the decision is written into the definition with its date. If a definition changes, the change is dated too, so comparing one quarter with the last does not need footnotes.

One source, no migration

Once the definitions are written, the figures need to come from one place. That does not mean replacing the ERP or banning spreadsheets: the operation keeps its tools, and what changes is where the meeting’s numbers come from.

In practice, a connection reads each system (the ERP, the shared sheets in Drive, the CRM and the service channels) and applies the agreed definitions. The result is a single set of figures, refreshed as often as each metric needs. Nobody copies data from one file to another, so nobody copies it wrong.

Conversations are data too: the requests and complaints that arrive over WhatsApp or email can be counted when they go through a channel that records them.

One dashboard per role, with the source of every number

The same source feeds different views. What changes from one role to the next is the level of detail; the figure stays the same:

• Operations. The day in detail: open cases, pending work by department and what falls due this week.
• Management. The month’s trend against target, and the deviations that call for a decision.
• Board. A few metrics, with the same definition month after month, so periods compare without extra explanations.

Every number on the dashboard shows its origin: which system it comes from, which cut-off and which definition apply. When someone questions a figure in the meeting, the answer is on the same screen and the conversation returns to the decision.

Reports nobody assembles by hand

In a regulated company, the report for the board or the supervisor is often put together by hand: someone downloads data from several systems and adjusts it in a spreadsheet. Every time the exercise is repeated, small differences appear that then have to be explained.

When reports come from the same source and the same definitions as the dashboards, what the board sees matches what management sees. Every reported figure can be rebuilt: which data it used, which cut-off and who approved the definition. If the supervisor asks how a number was obtained, the answer already exists.

The rule of thumb: if a figure is going to leave the company, it must be traceable to its origin without relying on the memory of whoever assembled it.

How we do it at Xplouse

We start with the definitions and the owners, together with management. Then we connect what already exists (spreadsheets, Drive, the ERP and conversations), with no migrations, and build dashboards per role with the source of every figure in view. From diagnostic to production in 30 days. The data layer shows how it works.