A single source of truth for operational reporting
Why most companies lack a single authoritative number, and what it takes to create one - without a large IT project.
If you ask three people from the same company - the sales director, the CFO and the operations director - for the same number, say last month's revenue, you often get three different answers. Sometimes they differ by a few percent, sometimes substantially. And each person is confident their number is correct.
This is not a mistake by specific people. It is a symptom that the company lacks a single source of truth - and everyone uses the system or methodology closest to their department.
Why this happens
Data in a company lives in several systems: CRM, accounting system, bank, marketing platform, Excel exports. Each system counts something different. The CRM counts deals by contract signing date, the accounting system by posting date, the CFO adjusts for prepayments and returns, sales looks at the pipeline.
Each of these methodologies is correct in its own way - for its own purpose. The problem is not that anyone is wrong. The problem is that these methodologies are nowhere explicitly documented and agreed upon.
The result: every meeting begins with arguments about numbers instead of discussions about decisions.
What a single source of truth means in practice
A single source of truth is not one database or one system. It is an agreement about which number is the official one for each metric and where it comes from.
It looks like this: for the "revenue" metric, the official number comes from the accounting system, is counted by posting date, includes VAT, and is calculated in local currency at the central bank rate on the transaction date. All other numbers are analytical views, not official reporting.
That agreement alone is already a major step. But it also needs technical implementation: making that number available to the right people at the right moment without manual assembly.
Typical traps
The first trap is starting with technical implementation before locking down the methodology. Building a data warehouse before definitions are agreed means automating confusion.
The second trap is trying to cover all metrics at once. The right approach is to pick 5 to 7 key operational metrics that most often create disagreements, and start with those.
The third trap is treating this as a purely IT project. The methodology is agreed by department heads, not by developers. IT implements what people have agreed upon.
Practical steps
If you want to reach a single source of truth in operational reporting:
- Bring together the heads of key functions and agree on 5 to 7 metrics that matter to everyone.
- For each metric, document: where the data comes from, what methodology is used to calculate it, who owns the definition.
- Implement this technically - even as a single report that updates automatically from one source.
- Record the agreements in writing - in a document that new employees can consult.
This process takes weeks, not months - if done pragmatically. A single source of truth is not an IT project. It is a management agreement with a technical implementation.