Walking into a new data strategy role is always a process of orientation. You inherit decisions you did not make, systems you did not choose, and a programme whose history lives in people's heads as much as in any documentation. For a data strategist joining an organisation that operates a loyalty programme, this orientation matters enormously, because loyalty infrastructure is one of the most consequential and most frequently misunderstood data assets in any business.

A loyalty programme is not just a CRM or a marketing channel. At its best, it is a longitudinal, consented, individual-level record of customer behaviour that no other data source can replicate. At its worst, it is an expensive points engine that accumulates liability, produces metrics no one trusts, and sits in a silo that has never been properly connected to the rest of the organisation's data estate.

The questions in this guide are designed to help a new data strategist move quickly from orientation to a clear-eyed assessment of where their organisation's loyalty infrastructure actually stands, what it is capable of, and where the most important gaps lie. They are not an exhaustive technical audit. They are the questions that distinguish a programme built for strategic value from one that has been running on inertia.

Before the Questions: One Prior Distinction

There is an important difference between a data audit and a data strategy assessment that is worth establishing before any evaluation begins. An audit checks what data exists and whether it is clean. A strategy assessment asks whether the organisation is structured to extract business value from it. Both are necessary, but the order matters.

For loyalty specifically, many organisations have conducted data audits and concluded they have a data quality problem. In practice, the issue is often upstream: no one has defined what decisions the loyalty data should be informing, so no one is accountable for the quality, coverage, and timeliness of the data that feeds those decisions. The questions below are strategy-first, not audit-first. Quality problems will surface naturally as you work through them.

Section 1: What Is the Programme Actually Trying to Do?

The single most revealing question in any loyalty infrastructure assessment is also the most basic. A surprising number of programmes cannot answer it clearly.

Q1  What specific customer behaviour is this programme designed to change?

If the answer is 'reward loyalty' or 'improve retention', probe further. Durable programmes are built around a specific, measurable behaviour: increase purchase frequency among mid-tier members, shift transaction mix toward higher-margin categories, reduce churn in the 90-day post-acquisition window. Generic objectives produce generic measurement, which produces programmes that cannot prove their value to a CFO.

Q2  Who defined those objectives, and when were they last reviewed?

Loyalty programme objectives set at launch often outlast the market conditions that made them relevant. A programme designed to drive in-store footfall in 2019 may be structurally unsuited to an organisation whose customers now transact predominantly through digital channels. Stale objectives are one of the most common reasons loyalty data fails to generate actionable insight.

Q3  Is the programme's success measured by outputs or by outcomes?

Output metrics (enrolments, points issued, redemption volume) are easy to produce and easy to game. Outcome metrics (incremental revenue per member versus a matched non-member cohort, retention rate uplift attributable to programme participation, customer lifetime value for members versus non-members) are harder to build but are the only measures that credibly connect the programme to business value. Find out which set of metrics the board and CFO actually see.

Section 2: What Data Is Being Collected, and What Is Being Missed?

Loyalty programmes collect a great deal of data. The more important question is whether they are collecting the right data, in the right structure, at the right moment.

Q4  Is member identity resolved consistently across channels?

The most common gap in loyalty data architecture is the inability to connect a member's online behaviour with their offline behaviour. A member who browses online, researches via app, and transacts in-store is three different people in most data systems. If identity resolution is not working, personalisation cannot work, and behavioural insight is at best partial. Ask specifically how the programme handles members who interact across multiple channels and devices.

Q5  What behavioural signals is the programme collecting beyond transactions?

Transaction data tells you what a member bought. Behavioural data tells you what they considered, what they abandoned, how they responded to an offer, how long it took them to redeem after a reward was issued, and what preceded a lapse in engagement. Programmes that capture only transactions have a fundamentally limited view of member motivation. Ask what non-transactional signals are being captured and how they are stored.

Q6  Are there meaningful gaps in the member profile, and is anyone accountable for closing them?

Progressive profiling, the practice of enriching member data over time through preference centres, surveys, and behavioural inference, is one of the most effective ways to build a richer first-party data asset. Ask whether there is a structured approach to profile enrichment, who owns it, and whether the consent framework supports it.

Section 3: How Is the Programme Measuring What Matters?

Measurement is where the gap between loyalty programmes that generate strategic value and those that merely generate activity tends to be most visible.

Q7  Is the programme measuring incrementality, or just attribution?

This is the most commercially significant measurement question in loyalty analytics. Attribution gives credit to the programme for all revenue generated by members. Incrementality isolates how much of that revenue would have occurred without the programme. The difference can be substantial. A programme that attributes all member spend to itself will consistently overstate its ROI. A programme that measures incrementality through matched cohort analysis or A/B testing produces numbers that finance teams can trust. Ask whether there is a control group methodology in place.

Q8  How is breakage being tracked, and at what granularity?

Breakage, the proportion of issued points that are never redeemed, is often treated as a financial win by accounting teams because unredeemed liability does not need to be paid out. In reality, high breakage among engaged member segments is an early warning signal of programme disengagement. Industry benchmarks suggest a healthy redemption rate sits between 15 and 25 percent. Ask how breakage is tracked: as a blended aggregate, or segmented by member tier and tenure. An aggregate that looks healthy can conceal near-zero redemption among your highest-value members.

Q9  What are the leading indicators in the measurement framework?

Lagging indicators (customer lifetime value, revenue per member, churn rate) confirm what has already happened. Leading indicators (first-redemption conversion rate, active member rate trends, reward attainability rate, engagement frequency in early member cohorts) signal what is about to happen. Well-run programmes use both. Ask which set dominates the current reporting suite and whether there is a mechanism for acting on leading indicators before churn or disengagement becomes visible in the lagging data.

Section 4: How Does the Loyalty Data Connect to the Rest of the Organisation?

A loyalty programme that operates as a standalone system, disconnected from CRM, commercial planning, customer service, and product development, is a missed opportunity at every level.

The biggest gap most enterprises have is an inability to link online behaviour with offline behaviour because they keep data acquisition programmes separate from first-party data programmes. There must be a connecting layer that assigns every interaction to a unified customer profile.

Q10  Which systems does the loyalty platform exchange data with, and how frequently?

Ask for a map of data flows: what goes into the loyalty platform, what comes out, and on what cadence. Specifically probe for CRM integration (is loyalty data available to customer service teams in real time?), commercial planning integration (do buying and category teams see member behaviour?), and data warehouse connectivity (can loyalty data be analysed alongside other enterprise data?). Manual exports and overnight batch processes are red flags in a programme expected to support real-time personalisation.

Q11  Is the loyalty data accessible for analysis, and by whom?

A loyalty programme that produces data accessible only through the platform vendor's own reporting tools, or that requires the vendor's professional services team to extract insights, is a dependency risk. Ask whether member-level data is available in the organisation's own data environment, whether analysts can query it directly, and what the data portability terms in the vendor contract look like.

Q12  Does the programme produce insight that flows back into commercial decisions?

This is the test of whether the loyalty data asset is actually being used. Ask for a recent example of a commercial or product decision that was made differently because of loyalty programme insight. If the answer is vague or hard to produce, the programme is generating data that is not reaching decision-makers in a usable form.

Section 5: Is the Infrastructure Ready for What Comes Next?

The final set of questions is forward-looking. They assess whether the programme's data architecture is positioned to support the capabilities that most organisations are now planning: AI-driven personalisation, agentic loyalty mechanics, and real-time member engagement.

Q13  Is the data architecture structured for real-time decisioning, or for batch reporting?

A programme built on nightly batch processes cannot support real-time offer delivery, in-journey engagement, or dynamic reward mechanics. Ask whether the platform supports real-time event streaming, and whether the member data model is structured in a way that allows real-time queries without performance degradation at scale.

Q14  How is consent managed, and does it support the data uses you are planning?

Consent architecture is one of the most frequently deferred governance questions in loyalty programmes. Ask what consent members have given, how it is recorded, whether it covers the uses the programme is currently making of the data, and whether it would cover planned uses such as AI-driven personalisation or data sharing with commercial partners. Gaps here are not just a compliance risk. They are a constraint on what the programme can do.

Q15  If you were to build an AI layer on top of this data asset today, what would block you?

This question surfaces the practical gap between ambition and readiness. Common blockers include inconsistent member identity resolution, incomplete behavioural data capture, inaccessible or poorly documented data models, batch rather than real-time data pipelines, and consent frameworks that do not cover AI-based decisioning. The answers to this question become the input to a realistic infrastructure roadmap.

What to Do With the Answers

These fifteen questions will not all produce clean answers immediately. Some will surface genuine uncertainty within the organisation about how the programme works or what it is measuring. That uncertainty is itself a finding, and often the most important one.

The pattern to look for is the gap between what the programme claims to do and what the data infrastructure can actually support. A programme that claims to personalise at scale but cannot resolve member identity across channels. A programme that claims to drive incremental revenue but has no control group methodology. A programme that claims to be a strategic data asset but whose data lives exclusively in a vendor's SaaS environment with no portability guarantee.

Closing those gaps is the work of a data strategy for loyalty. Identifying them clearly, with specifics and evidence, is the essential first step. The questions above are designed to get you there faster than most infrastructure reviews that focus on technology before establishing what the technology is supposed to be doing.