Measurement
Part of When to rebuild loyalty marketing measurement and reporting
Choosing loyalty marketing key metrics without vanity numbers
How to choose loyalty marketing key metrics that survive board scrutiny, with a candidate list, selection criteria and England-specific data checks.
What to take away
- A high-street retailer in England with 40,000 members cannot tell whether the scheme pays for itself from sign-ups alone.
- Choose a short set of loyalty marketing key metrics that link recruitment to repeat purchase, margin and retention, and agree definitions before anyone builds a dashboard.
- Score each candidate on business link, data quality, actionability and cost to collect, then drop anything failing two or more.
- Record the period, cohort and source for every figure you publish.
- Include referral metrics beside scheme metrics, because referral and loyalty normally run as one programme.
- Re-check the set each quarter, because metrics that work at launch often lose value as a scheme matures.
Which metrics belong in a loyalty scorecard?
Most reporting problems in England start when a team leads with sign-ups, then cannot explain falling revenue per member. Choose deliberately rather than inheriting whatever the platform shows first, and write down what decision each candidate metric would change before anyone builds a dashboard.
Four Jobs of a Loyalty Scorecard
- Acquiring membersjoin rate, cost per enrolled member
- Changing behaviourrepeat purchase, redemption, basket size
- Protecting margingross margin, breakage, reward cost
- Staying crediblecomplaints, opt-out rate, data accuracy
Five to seven metrics is usually enough; more than that and the board stops reading. Traps appear as the list grows, particularly double counting and unclear baselines.
A candidate list of loyalty key metrics
Start wide, then score. A loyalty platform's own data will usually support:
- Repeat purchase rate
- the share of members who buy more than once in the period.
- Purchase frequency
- the average number of transactions per active member.
- Average transaction value
- revenue divided by number of transactions.
- Active member rate
- members who transacted, not merely enrolled.
- Member penetration of sales
- member transactions as a share of total revenue.
- Redemption rate
- points or rewards redeemed as a share of those issued.
- Member retention or lapse rate
- the share of a cohort still active after a set period.
- Cost per enrolled member
- recruitment spend divided by enrolments.
- Revenue per member
- revenue divided by active members.
- Customer lifetime value
- expected margin per member over the expected relationship.
Referral metrics belong on the same shortlist, because most schemes run loyalty and referral as one programme:
- Referral rate
- the share of members who refer at least once.
- Referral conversion rate
- referred prospects who enrol as a share of referrals made.
- Referred-member share of acquisition
- new members from referrals as a share of all new members.
- Referral cost per acquisition
- referral rewards and programme cost divided by referred members acquired.
- Referred-customer value
- revenue or margin per referred member against non-referred members.
What makes a number a vanity metric
A vanity metric moves without telling you whether the scheme creates value or what to do next. Sign-ups are the classic case: they rise with marketing spend and say nothing about margin. Total points issued, app downloads, rewards-page views, follower counts and cumulative members ever enrolled behave in the same way.
A substantive metric has a named owner, a decision attached and a direction that would change behaviour if it moved. Revenue per member, active member rate, member penetration of sales and referred-member share of acquisition clear that bar; points issued and app downloads rarely do.
How do you score each candidate metric?
Score each metric from 1 to 5 on four criteria, then multiply by a weight. Business link and data quality carry the most weight, because a weak link or dirty data makes a number useless however cheap it is to produce.
Metric Scoring Criteria and Weights
Criterion
- Business link
- 40%
- Data quality
- 30%
- Actionability
- 20%
- Cost to collect
- 10%
Weight
- Business link
- No clear decision
- Data quality
- Manual, incomplete
- Actionability
- Cannot be changed
- Cost to collect
- Bespoke, high effort
Score 1
- Business link
- Informs budget or design
- Data quality
- System-generated, audited
- Actionability
- Team can influence
- Cost to collect
- Already in platform
Score 5
- Business link
- Data quality
- Actionability
- Cost to collect
Scoring candidate metrics
Weight
- Business link
- 40%
- Data quality
- 30%
- Actionability
- 20%
- Cost to collect
- 10%
Score 1
- Business link
- No clear decision attached
- Data quality
- Manual, incomplete, disputed
- Actionability
- Cannot be changed by the team
- Cost to collect
- Bespoke build, high effort
Score 5
- Business link
- Directly informs a budget or design choice
- Data quality
- System-generated, complete, audited
- Actionability
- Team can influence within a quarter
- Cost to collect
- Already in the platform, low effort
Multiply each score by its weight, add the four results and rank. Drop or rework anything scoring below 3 overall. For example, a team paying £400 a month for a loyalty platform might score redemption rate highly on cost to collect. It scores poorly on business link if nobody acts on it. Run the same test on sign-ups and the contrast is plain: sign-ups are cheap to collect, but unless a budget or design decision depends on them they fail business link, which is what makes them a vanity number.
What does England-specific context change?
Consumer protection in England shapes how you describe rewards and points. The Consumer Rights Act 2015 covers unfair terms and notices. That includes terms letting a scheme cut reward value without clear notice, so track how often terms change and how members hear about it. Read the legislation on unfair terms and notices before publishing a benefits statement.
Market context also affects which metrics are comparable. The Office for National Statistics publishes business services data that helps you size the sector and set benchmarks instead of copying figures from unrelated industries.
If your team needs to sharpen how it reads member behaviour, customer experience training from the Chartered Institute of Marketing (CIM) covers programme design and measurement, which gives you a firmer basis for the definitions you publish.
How do you keep the set honest over time?
Write a definition for every metric: numerator, denominator, period, cohort and exclusions. Store it beside the dashboard so new analysts cannot reinterpret it. Review quarterly, retire metrics that no longer drive a decision, and read common measurement mistakes in England when a figure keeps being disputed.
Publish a short evidence note with each board pack, stating the source, the date range and any known gaps. Label illustrative figures as illustrative. That habit stops a rough estimate hardening into a quoted fact.
Common questions
How many metrics should a loyalty scheme report?
Five to seven is a practical ceiling for most England-based teams. If a metric cannot change a decision within a quarter, move it to an appendix or drop it.
Should sign-ups be a headline metric?
Only alongside cost per enrolled member and a retention measure. Sign-ups alone say nothing about whether the scheme creates value.
How often should definitions be reviewed?
At least quarterly, and whenever the platform, pricing or reward structure changes. Undocumented changes to a denominator cause most disputed numbers.
Can we compare our metrics to industry benchmarks?
With care. Use published market data for context, but compare against your own history first, since scheme design and customer base vary widely.
