Business intelligence2 min read

One definition per metric: why your dashboards disagree

When two reports show different revenue for the same month, the problem is rarely the maths. It is that the business never agreed what the number means.

Written by The metadatum team

Analytics charts on a laptop screen

Every growing organisation reaches the meeting where two people bring two different numbers for the same thing. Revenue, active customers, on-time delivery. Both reports are "right", because each calculates the figure its own way. The meeting turns into a debate about the data instead of a decision.

Where the differences come from

  • Different filters. One report excludes refunds, another doesn't. One counts trial users, another only paying ones.
  • Different timing. Orders counted when placed in one place and when invoiced in another.
  • Different sources. Finance reads from the ledger, sales from the CRM, and the two never quite reconcile.
  • Logic copied by hand. The same calculation rebuilt in five dashboards, each drifting a little as people fix things locally.

A semantic layer: define it once

A semantic layer is a single place where each metric is defined, in code, with its filters, timing and source. Every dashboard, spreadsheet and AI assistant reads the metric from there instead of recalculating it.

Without oneWith one
Each report computes revenue its own wayRevenue is defined once and reused
Fixes are made in one dashboard at a timeA fix in the definition reaches every report
Nobody knows which number is officialThe definition is the official number

Getting started

  1. Pick the ten metrics leadership argues about most. Start there, not with everything.
  2. Agree the definition in writing, with the people who own each number, before writing any code.
  3. Encode it once in the semantic layer, with tests that check known results.
  4. Point the most-used dashboards at it and retire the copies.

The goal isn't a perfect catalogue. It's that the next time a number is questioned, the answer is a definition everyone already agreed, not another spreadsheet.

  • AI

    Why AI pilots stall at the data, not the model

    A pilot that dazzles in a demo often falters in production. The reason is rarely the model. It is the data underneath it, and that is fixable.

  • Data engineering

    Data contracts: stop upstream changes breaking everything downstream

    A data contract is a promise between the team that produces data and the teams that depend on it. Here is how to put them in place without slowing anyone down.