Consumer & Omnichannel · Growth problem
Loyalty and retention don't hold together across channels
Loyalty programmes are usually designed with good intentions and measured by membership growth, but retention depends on something harder: reliably recognising a customer wherever they shop, and using purchase history from every channel to decide what to say to them and when. Where identification is patchy, retention activity ends up serving the customers who are easiest to see, not necessarily the ones most worth keeping.
Symptoms
What this usually looks like
- Loyalty member identification rate at point of sale is inconsistent by store or channel
- Email and SMS activity is based only on online purchase history, missing in-store behaviour
- High-value in-store customers receive generic or infrequent communication
- Category-level purchase frequency data is incomplete for customers who shop across channels
- Retention campaigns show reasonable open rates but limited effect on repeat purchase
- Cross-channel repeat purchase and LTV can't be reported with confidence
Diagnostic questions
What we would test first
- Measure member identification rate at point of sale across a representative sample of stores and shifts
- Compare loyalty tier assignment against a manually resolved cross-channel purchase history for a sample of customers
- Design a holdout test for an upcoming retention campaign to measure effect on repeat purchase, not just engagement
- Review CRM segmentation logic for how heavily it weights digital vs. in-store behaviour
- Audit category frequency data completeness for customers known to shop across channels
- Interview store staff on current loyalty sign-up and identification prompts at the till
Root causes
Why it happens
- 01
Member identification at point of sale is inconsistent
Whether a store transaction gets linked to a loyalty account often depends on staff prompting, till workflow friction, or customer willingness to provide details in the moment, producing wide and largely unmanaged variation in identification rates.
- 02
CRM decisioning is built primarily on digital behaviour
Email opens, site visits and online purchases are the easiest signals to capture, so CRM segmentation and triggers are often skewed towards digitally active customers even when a large share of value sits with primarily in-store shoppers.
- 03
Incentive structures don't reflect cross-channel value
Loyalty rewards and tiering are sometimes calculated on incomplete purchase history, understating true customer value and misallocating incentive spend towards customers who look valuable only because their record is more complete.
- 04
Retention is measured by activity, not by resulting behaviour change
Campaign performance is often reported on opens, clicks and short-term redemption rather than on whether it changed underlying repeat purchase frequency or category spread, making it hard to tell what's actually working.
Evidence
The numbers we would look at
These are the metrics that make the constraint visible, and the cuts that stop them being reassuring by accident.
| Metric | What it tells you | How we read it |
|---|---|---|
| Member identification rate at point of sale | The foundation metric for whether loyalty and CRM data can be trusted at all.Report by store and shift, not just as a network average, since the variation is where the fixable problem usually lies. | Report by store and shift, not just as a network average, since the variation is where the fixable problem usually lies. |
| Cross-channel repeat purchase rate by cohort | Whether customers who engage with more than one channel repeat purchase at a different rate to single-channel customers.Needs resolved identity; without it, this metric will be biased towards easily-identified customers. | Needs resolved identity; without it, this metric will be biased towards easily-identified customers. |
| Category frequency and breadth by resolved customer | How many categories a customer buys into and how often, once channel data is combined.A key input to genuine value-based segmentation, often unavailable without identity resolution. | A key input to genuine value-based segmentation, often unavailable without identity resolution. |
| Retention campaign effect on repeat purchase (not just engagement) | Whether communications actually change buying behaviour, not just get opened.Requires a holdout or control group; open and click rates alone don't establish causation. | Requires a holdout or control group; open and click rates alone don't establish causation. |
| Loyalty tier accuracy against true cross-channel value | Whether reward tiers reflect actual customer value or just the visible, digitally-captured portion of it.Worth spot-checking a sample of high in-store spenders against their assigned tier. | Worth spot-checking a sample of high in-store spenders against their assigned tier. |
Measurement traps
What can mislead you
- Looks fineLoyalty membership numbers keep growing, so the programme is succeeding
- Membership growth measures sign-up, not engagement or retention; a growing base with a falling or flat active-member rate points to a programme that's acquiring but not holding attention.
- Looks fineEmail engagement metrics for loyalty campaigns are strong
- Engagement metrics reflect the digitally-visible portion of the loyalty base and say nothing about the members who shop mainly in-store and may not even be reliably reached by email.
- Looks fineOur top loyalty tier customers are our most valuable customers
- If tiering is based on incompletely captured purchase history, some of the most valuable customers, particularly frequent in-store shoppers with poor identification, may be sitting in lower tiers than they deserve.
Outcome
What better looks like
Not a promised number. A clearer basis for the next investment decision.
- Member identification rate is tracked and actively managed as an operational KPI, not left to chance
- CRM segmentation and loyalty tiering are built on resolved cross-channel purchase history where feasible
- Retention campaign success is measured against repeat purchase behaviour, using holdouts where practical
- Store teams understand their role in loyalty identification and are supported with simple, low-friction processes
Where a Growth Diagnostic would start
A three to four week senior review across demand, discovery, acquisition, conversion, retention, measurement and capability — sequenced so this problem is either confirmed as the constraint or ruled out early. Read alongside the consumer & omnichannel model page for how we frame the wider system.
Related problems
Consumer & Omnichannel
Customer data is fragmented across store and digital systems
POS, loyalty, ecommerce accounts and CRM rarely resolve to the same customer, so retention and value metrics understate reality.
Consumer & Omnichannel
Growth has plateaued once single-channel gains are exhausted
Each channel has been optimised individually, but total growth has flattened because nobody owns the space between them.
Ecommerce & Retail
Low repeat purchase and retention
New customers convert once but don't return often enough. A review of repeat purchase, cohorts and CRM's real contribution.
Questions about this problem
- Do we need to redesign the loyalty programme itself?
- Not necessarily. Many of the issues here are about identification and data use rather than the programme's reward structure, which is often the last thing that needs changing.
- How do you test retention campaign effect without a holdout, if our systems don't easily support it?
- Where a formal holdout is hard to build, a phased rollout or matched comparison group can usually approximate one well enough to draw a useful conclusion.
- Is store staff training likely to be part of the recommendations?
- Often yes, in some form, since identification rate is frequently more a matter of prompt design and staff workflow than of technology.