Data definitions: documenting what each metric means

Ask three people in the same company what “active customer” means and you may get three answers. One counts anyone who logged in this month. Another counts anyone with a paid invoice. The third counts anyone who hasn’t cancelled. All three numbers are real, and all three will turn up in different reports under the same name.

This is the quiet problem behind many reporting arguments. The figures are not wrong. They are measuring different things. A data definitions glossary fixes it, and it is one of the least glamorous and most useful documents a finance or operations team can write.

What a glossary is for

A glossary sets down, in plain words, what each metric in your reports means and how it is produced. It is not a data dictionary for engineers, listing every column in every table. It is written for the people who read and prepare reports: managers, analysts, the board, a new hire in their first week.

It does three jobs:

• Settles disagreements before meetings. “Which revenue is this?” has an answer on the page.
• Keeps reports consistent over time. A metric means the same thing in March as it did in January.
• Makes handovers possible. Someone covering the weekly report can produce the same figures as the person who usually does it.

What each entry needs

Keep the format the same for every metric. Six fields cover most needs:

Field	Example
Name	Active customers
Definition	Accounts with at least one paid invoice in the calendar month
Calculation	Count of distinct account IDs with an invoice status of “paid”, dated in the month
Source	Billing system, monthly export
Period	Calendar month, as at the last day
Owner	Finance lead

The example is illustrative; your definitions will be your own. Two fields deserve particular care.

The definition should be readable by someone who has never seen the source system. “Accounts with at least one paid invoice in the month” works. “COUNT DISTINCT acct_id WHERE inv_status = 3” belongs in the calculation field, if at all.

The owner is a role, not a name. It is the person who decides what the metric means and approves any change to it. Without an owner, definitions drift by default.

Include the exclusions

Most confusion comes from what a metric leaves out. Write the exclusions explicitly:

• Revenue: excludes GST and credit notes issued in the period
• Headcount: excludes contractors and interns
• Orders: excludes test orders and orders cancelled within 24 hours

A line like this often resolves a long-running disagreement in a sentence, because it turns out the two sides were each right about a different number.

Name the near neighbours

Some metrics are routinely confused with others. Put them side by side in the glossary, with the difference stated:

• Bookings are signed contract value in the period. Revenue is the amount recognised in the period.
• Sign-ups are new accounts created. New customers are accounts with their first paid invoice.

Readers who look up one will see the other, and the distinction sticks.

Start with the figures people quote

Don’t try to define everything at once. Start with the metrics that appear in the most-read reports and the ones that cause the most questions. For many teams that is ten to fifteen entries. A short glossary that is accurate is far more useful than a long one nobody trusts.

A practical way to find them: take the last three editions of the main weekly or monthly report and list every figure that appears in a headline or a KPI table. That list is your first draft. How to automate recurring reports uses the same audit as the starting point for setting up a report that updates itself.

Put definitions where readers are

A glossary in a shared folder nobody opens does little. Bring the definitions to the reports themselves:

• a short definitions section at the end of each recurring report
• a one-line footnote under any KPI likely to be misread
• a link to the full glossary from every report that uses it

In a board pack, a glossary page at the back is one of the simplest ways to stop arguments about mismatched numbers; board pack figures that match on the day covers the other causes.

Change definitions deliberately

Businesses change, and definitions must change with them. A new pricing model may change what counts as an active customer. When that happens:

1. The owner agrees the new definition.
2. The glossary entry is updated, with the date the change takes effect.
3. The old definition stays in the entry, marked as superseded, so past reports can still be read correctly.
4. The next report that uses the metric carries a note: “Active customers now counted on paid invoices only, from April.”

The note matters most. A metric that changes meaning silently does more damage than one that was never defined, because readers compare this month with last and draw the wrong conclusion.

Review it on a schedule

Set a light review, perhaps once a quarter, in which each owner confirms their entries are still right. It takes minutes when nothing has changed, and it catches the slow drift that happens when a source system is updated or a team starts calculating a figure slightly differently.

Clear definitions make linked figures trustworthy. Live Documents can keep each number tied to its source; the glossary makes sure everyone agrees what that number is.