Guides
Loyalty data and measurement
Loyalty data and measurement turns member activity into the financial and behavioural evidence needed to run a programme. The core is simple: track earn, burn, breakage, tier movement, and inactivity against a clear baseline. That shows whether rewards change behaviour or just cut margin. Robust measurement links the points ledger to the income statement and isolates incremental lift. This guide gives the working path from raw records to programme decisions.
What this covers
This page covers the measurement of loyalty programme data: how member activity becomes financial and behavioural evidence, and how that evidence drives operating decisions. It includes points issued, points redeemed, breakage, liability, tier movement, inactivity, and the cost of each incremental behaviour. It does not cover general customer data platforms, privacy compliance, or the design of reward catalogues. Those areas sit outside this cluster. The cluster's 31 definitions and 8 decisions provide the shared vocabulary and the contested questions for loyalty measurement.
The boundary is deliberately narrow. Upstream data collection, such as point of sale integration, matters here only when it changes the measurement record. Downstream campaign analytics become part of loyalty measurement only when they isolate the programme's incremental effect on member behaviour. In short, this guide is about the ledger and the scoreboard, not the creative or technical infrastructure around them.
How the pieces fit together
A loyalty data system starts when a member joins. The join record carries an identifier, the enrolment channel, and the initial tier. Every subsequent earn, burn, adjustment, expiry, and tier change is logged against that identifier. The log must be stable enough to serve as the programme's financial record, because points are a liability until they are redeemed or expired.
The financial spine runs from issued points to redeemed points. When a member earns a point, the programme records a cost and a deferred revenue obligation. When the point is redeemed, the obligation is settled and the cost of the reward is recognised. Points that never redeem become breakage, which reduces the liability. Breakage is not a target to maximise blindly; it is an estimate based on historical redemption behaviour and programme design. The liability model must reconcile to the accounting record every month, or the measurement is not yet trustworthy.
Behavioural measurement sits on top of that financial spine. It compares members against non-members, or before and after enrolment, to track purchase frequency, basket size, cross-category buying, and tier progression. The aim is not to prove that members are valuable in absolute terms. The aim is to isolate the value the programme creates, which is the difference between observed behaviour and what would have happened without the programme.
Reporting turns these records into a regular operating rhythm. A measurement dashboard should show the liability balance, redemption rate, breakage rate, active member rate, tier mix, and cost per incremental behaviour. These metrics must reconcile to the book of record. If they do not, the data model is wrong before any business decision is made. The register's 57 curriculum modules and 87 vendor profiles support the tooling and model choices, but the measurement discipline comes before any software selection.
Where programmes get this wrong
Many programmes treat enrolment as the headline metric. The register shows 217 programmes profiled, with 214 having known tier status and only 112 of those running tiers. Enrolment alone hides the fact that a programme can be large but inert. The first wrong turn is to report member count without defining an active member.
Breakage is either ignored or set as a hard target. If a programme records every issued point as a liability forever, it overstates cost and underinvests. If it assumes breakage too high, it understates liability and risks a cash call when redemption spikes. Both errors come from failing to model historical redemption curves by cohort and age of points.
Measurement often stops at redemption. A high redemption rate can mean engaged members or a reward catalogue that is too easy, burning margin without changing behaviour. The missing test is incremental lift: do members exposed to the programme spend more than a control group would have spent anyway. Without that test, the programme is a discount with extra steps.
Tier data is sometimes treated as a vanity metric. The register shows 61 of 61 airline programmes with known status run tiers, but only 25 of 46 retail programmes with known status do. Some sectors force tiers where they do not fit the purchase rhythm, creating complexity without changing behaviour. Measurement should ask whether tiers shift share of wallet or merely reward the already loyal.
How to work through it
Start with a data audit. List every system that creates or changes a member record: point of sale, ecommerce, call centre, and partner channels. For each system, record whether it sends a unique member identifier, a timestamp, a transaction value, and an event type. Most audits find missing identifiers and duplicate records before any metric is calculated.
Define the event taxonomy next. Use the register's 31 definitions and 8 decisions to settle what earn, burn, adjustment, expiry, transfer, and tier change each mean. Without that shared language, two teams will report different numbers for the same programme.
Construct the outstanding points balance. Begin with the points created in a period, take away those used and those expired, and apply the expected non-redemption share derived from past cohorts. Check the result against the accounting system each cycle. Stop and repair the data if the two do not agree.
Set the measurement plan. Choose the metrics that matter for this programme's business model: for a retailer, share of wallet and purchase frequency; for an airline, incremental revenue per available seat mile; for a coalition, cross-partner purchase. Write down the control group or baseline before launching any change.
Produce a regular report that separates financial and behavioural lines. The financial view shows liability, breakage, cost per point issued, and cost per point redeemed. The behavioural view shows active members, tier mix, redemption rate, and incremental lift. The report should be short enough to read in one sitting and tied to a decision each period.
Revisit the programme economics quarterly. Use what you measured to adjust earn rates, redemption thresholds, expiry rules, and tier benefits. The goal is not more data, but better decisions. Adjust what you measure only when a decision depends on it.