Consumer & Omnichannel · Growth problem
Customer data is fragmented across store and digital systems
A customer who shops in-store with a loyalty card, buys online with a different email, and calls the contact centre using a third identifier looks, in most systems, like three separate people. This isn't a minor data hygiene issue: it directly understates repeat purchase and lifetime value, biases segmentation towards whichever channel has the cleanest data, and makes retention activity less effective than it could be.
Symptoms
What this usually looks like
- The same customer appears multiple times across POS, ecommerce and loyalty systems
- Reported repeat purchase rate looks lower than store staff's anecdotal experience suggests
- Segmentation and targeting rely heavily on whichever channel has the most complete data
- Loyalty scheme sign-up doesn't reliably link to online account activity
- Marketing suppression and consent records aren't consistent across channels
- Customer service can't see a full purchase history when a customer contacts any channel
Diagnostic questions
What we would test first
- Sample a set of customer records across POS, ecommerce and loyalty systems and manually check for likely duplicate identities
- Measure loyalty/customer ID capture rate at point of sale across a sample of stores and shifts
- Compare channel-reported repeat purchase rate against a manually resolved sample for the same customers
- Review how consent and suppression preferences are currently synchronised, or not, across systems
- Interview CRM and store operations teams on known data quality gaps and workarounds already in use
- Assess technical feasibility and cost of a basic identity resolution layer given current systems
Root causes
Why it happens
- 01
There is no shared, persistent customer identifier
Loyalty numbers, email addresses, phone numbers and POS transaction records are collected independently by different systems with no reliable matching key, so identity resolution has to be inferred probabilistically or not attempted at all.
- 02
Store data capture is inconsistent
Loyalty card usage, email capture and phone number collection at point of sale depend heavily on staff behaviour and till workflow, producing gaps that are much larger and more variable than equivalent online data capture.
- 03
Systems were bought and integrated at different times for different purposes
POS, ecommerce platform, loyalty programme and CRM are frequently from different vendors, implemented years apart, with no data architecture that anticipated needing to resolve identity across all of them.
- 04
Consent and suppression records aren't centralised
A customer who opts out via one channel may still be contacted via another if consent isn't managed centrally, creating both a customer experience problem and a compliance risk that grows as channels multiply.
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 |
|---|---|---|
| Identity resolution / match rate across systems | The percentage of transactions or customer records that can be confidently matched to a single customer profile.Low match rates mean every downstream metric — retention, LTV, segmentation — is working from an incomplete and likely biased sample. | Low match rates mean every downstream metric — retention, LTV, segmentation — is working from an incomplete and likely biased sample. |
| Blended repeat purchase rate | Repeat purchase measured across all channels for a resolved customer, not per channel account.Compare against channel-reported repeat rates to quantify how much is currently hidden by fragmentation. | Compare against channel-reported repeat rates to quantify how much is currently hidden by fragmentation. |
| Loyalty capture rate at point of sale | How consistently store transactions are linked to a loyalty or customer record.Varies significantly by store and shift; worth reviewing at that granularity rather than as a network average. | Varies significantly by store and shift; worth reviewing at that granularity rather than as a network average. |
| Total customer value by resolved identity | A more accurate lifetime value figure once cross-channel activity is combined.Should be materially different from any single channel's LTV figure; if it isn't, resolution is probably still incomplete. | Should be materially different from any single channel's LTV figure; if it isn't, resolution is probably still incomplete. |
| Consent/suppression consistency rate | Whether an opt-out in one channel is honoured across all others.A compliance and trust issue as much as a data quality one — worth checking directly, not assuming. | A compliance and trust issue as much as a data quality one — worth checking directly, not assuming. |
Measurement traps
What can mislead you
- Looks fineOur CRM database has grown steadily, so our customer data is improving
- Database growth can simply mean more duplicate and partial records are being created, particularly if acquisition activity outpaces any identity resolution or deduplication effort.
- Looks fineLoyalty sign-up rate is high, so we understand our customers well
- A high sign-up rate doesn't guarantee the loyalty ID is being captured at every transaction or linked to online activity, so the depth of understanding may not match the breadth of membership.
- Looks fineSegment sizes from our CRM look statistically solid
- If CRM data over-represents customers who engage primarily online or opt into marketing, segments built from it will be systematically unrepresentative of the full customer base, including higher-value but less digitally visible customers.
Outcome
What better looks like
Not a promised number. A clearer basis for the next investment decision.
- Repeat purchase and lifetime value are reported on a resolved-identity basis, with the gap to channel-reported figures made explicit
- Loyalty and account capture at point of sale is tracked and actively managed as an operational KPI
- Consent and suppression preferences apply consistently regardless of which channel a customer used to set them
- Segmentation and targeting draw on a merged view of the customer, not whichever system is easiest to query
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
Loyalty and retention don't hold together across channels
Loyalty and CRM activity often only reaches a partial, digitally-biased view of the customer base, weakening retention.
Consumer & Omnichannel
Digital's influence on store sales isn't being measured
Search, store locator and calls are shaping footfall and phone enquiries, but digital only gets credited for sales it closes directly.
B2B & Lead Generation
We can't tell which sources actually drive pipeline
Marketing and sales disagree about which channels actually generate revenue, and the reporting can't settle it.
Questions about this problem
- Do we need a customer data platform to fix this?
- Not necessarily as a first step. The diagnostic assesses whether existing systems can be better linked before recommending new infrastructure, since process and capture fixes often deliver meaningful improvement on their own.
- How do you measure identity resolution without perfect data?
- Through sampling and probabilistic matching on available identifiers, which is enough to estimate the scale of the problem and prioritise fixes, even before a permanent solution is built.
- Is this a GDPR or compliance review as well?
- It touches on consent consistency, which has compliance implications, but it isn't a substitute for a formal compliance audit; we'd flag risks for your data protection function to address.