Why a loyalty marketing reporting dashboard needs fixed definitions

What to take away

Every metric on the dashboard needs one written definition, one owner and one source system before the first chart goes live.
Consent and data-broker checks belong on the dashboard build checklist, not in a separate compliance document.
A register template with metric name, code, definition, source system, refresh frequency and owner ends most review arguments.
A change log beside the dashboard records the date, reason and approver for every definition change.

A loyalty marketing reporting dashboard fails for boring reasons. Two teams count active members differently. Referral revenue includes VAT in one report and not the other. The fix is a definitions register that changes only through a recorded decision.

Start with the register

List every metric the business asks for. For each one, write the plain-English definition, the source system, the refresh frequency and the person accountable. Keep the list in one file the dashboard links to, so nobody has to email an analyst to check what "active member" means.

Metric register entry fields

Metric name and short code
Plain-English definition
Source system
Refresh frequency
Named accountable owner

Metric name	Code	Plain-English definition	Source system	Refresh frequency	Owner
Active member	AM90	A member with a qualifying purchase in the last 90 days	Loyalty platform	Daily	Head of CRM
Referral revenue	RR_NET	Net revenue from referred orders, after VAT and refunds	Ecommerce order table	Daily	Finance analyst
Redemption rate	RED_EARN	Points redeemed divided by points earned	Loyalty platform	Weekly	Loyalty manager

Definitions must be narrow enough to test. A member with a qualifying purchase in the last 90 days is active. An "engaged customer" is not.

Give each metric a short code as well as a name, so exports and charts carry the same label. Link the register as a CSV or Google Sheet to Power BI, Looker Studio or Tableau, and keep the metric code as the column header in every export.

Where nobody will own a metric, retire it instead of leaving it on the page.

Treat the register as the first step in the cycle set out in the loyalty marketing measurement and reporting guide 2027, which connects definitions to board-level reporting.

Pick the source of truth

Most schemes pull from a loyalty platform, an ecommerce order table and a payment provider. These disagree. Choose one system as the source of truth for each metric, and state that choice in the register. Label any reconciliation system a check, not a second answer.

Reconcile the loyalty platform against the ecommerce order table daily, and both against the payment provider at month end before the finance close.

Choosing the source of truth

List loyalty platform, order table, payment provider
Pick one source per metric
Record the choice in the register
Label other systems as checks
Agree the period close date
Restate prior period once for refunds

Agree the date boundary as well, because a month that closes at midnight will disagree with finance if finance closes earlier. Refunds and chargebacks land late, so restate the previous period once rather than editing history every week.

If you buy audience data to target lapsed members, read the ICO guidance on organisations using marketing services of data brokers before the data reaches the dashboard, because consent and lawful basis decide what you can report and reuse.

Check the provider and vendor-reported numbers

A vendor dashboard is a source of numbers, not a definition. Ask for the metric code, the extract date and the refresh schedule. Reproduce one figure in your own warehouse before it appears on the board.

Company law sets duties a supplier must meet, so check the Companies Act 2006 rather than accepting a sales description. Match the contracting entity to the name on the invoice, because a brand you cannot identify is a reporting gap.

Run the two-analyst test

What Analyst A does

Draft definition: "Engaged" means opened an email in 30 days
Same extract: Runs the count on the loyalty platform
Compare totals: Total is lower than Analyst B
Rewrite: "Active member" means a qualifying purchase in the last 90 days

What Analyst B does

Draft definition: "Engaged" means opened an email or redeemed a point in 30 days
Same extract: Runs the count on the same extract
Compare totals: Total is higher than Analyst A
Rewrite: Applies the same wording

Decision

Draft definition: Wording is ambiguous
Same extract: One extract date
Compare totals: Draft fails
Rewrite: Both totals match

Run this test before launch. If the totals differ, the wording is still too loose. Record the agreed wording in the register.

Set the review rhythm

Run a monthly definitions review with marketing, finance and data present. Any change gets a date, a reason and a named approver. Publish the change log beside the dashboard.

Watch the common failures

Dashboards drift when owners leave, when a campaign invents a bespoke definition, or when a vendor supplies a number nobody can reproduce. The loyalty marketing measurement mistakes in England article lists the patterns that recur most, so read it before your next quarterly review.

Use the ONS labour market overview only as an external check on new member sign-ups and redemption rates, never as a dashboard source.

Common questions

How many metrics should a loyalty dashboard show?

Start with fewer than twelve. Each one needs an owner and a definition, and a long list usually means several metrics are unowned and quietly wrong.

Should referral and loyalty metrics share a dashboard?

Yes, if they share a member record. Separate dashboards hide the overlap where a referred member also redeems points.

How often should definitions change?

Rarely, and never mid-quarter without a change log entry. A definition change resets trend lines, so it needs a date and an approver.

Who owns the dashboard?

One named person in the marketing or data team. Committees can review it, but a single owner keeps the register current.