Table of Contents
You’re in a meeting, and a customer profile page appears in a slide deck.
People read it. And then you notice something small happens.
An awkward pause, or someone says, “That used to be true”.
Or nothing is said, but you can feel that different teams have different versions of the customer in their minds.
This is why understanding retail customers isn’t primarily a data problem. Most retailers have data. The harder question is whether the image built from that data still reflects who the customer actually is.
How does the picture go out of date?
Customer profiles are often treated as finished outputs. Research is commissioned, findings are presented, and a persona gets “signed off”. And from that point, it tends to stay because there is no ongoing process to test the profile against the current reality.
Retail has moved quickly since 2020. The Centre for Retail Research projected over 17,000 UK store closures in 2025. The Centre for Cities has reported that footfall remains more than 10% below pre‑pandemic levels even after several years of recovery.
Hybrid working changed shopping routines. The cost‑of‑living shift changed not just what customers can afford, but how spending feels. That’s a psychological shift that doesn’t show up in transactional data.
A profile built around the old routine describes a world that has moved on. Most businesses know this in theory. The problem is that the task of checking whether the customer on the slide still matches the customer in the store is often overlooked.
What staff turnover does to customer knowledge
There is a cost to staff turnover that rarely appears in a budget line. It removes the “why” behind customer understanding.
The CIPD and ONS put annual turnover in UK wholesale and retail at around 41%, meaning that across the sector, nearly half the workforce changes over each year.
What leaves with people isn’t only time and capacity. It’s the context behind the research. The insight manager who commissioned the segmentation model knew what they saw in stores, what the data excluded, and why a particular framing was chosen. That context rarely transfers in a handover document. What transfers is the output, like the deck, the persona, the approved summary.
The next team inherits the picture without the story behind it. And because it looks official (well-presented, approved, embedded in templates), they use it. They build plans from it. And the gap between the approved profile and the actual customer widens.
This is made worse when research budgets tighten.
The BRC’s 2026 CFO Survey found that 84% of retail finance leaders now rank labour and employment costs in their top three concerns. In that environment, refreshing customer understanding starts to feel like something for “when things settle down”.
Things rarely settle down. So the picture gets older still.
When a score replaces a picture
As original research recedes, something tends to fill the gap. Not new insight, unfortunately, something easier. A satisfaction score. An NPS figure. A weekly metric that drops neatly into a report.
This happens because the measure already exists, is consistent, and provides the comfort of “tracking the customer”.
But a score is not a customer. It can tell you that something changed; it can’t tell you what the customer was experiencing, thinking, or trying to achieve.
The cost‑of‑living period makes this clear.
Research from Savills and McKinsey on UK grocery habits found that own‑brand products now account for more than half of all grocery spending. Some of that shift is straightforwardly about budgets. But some of it reflects a change in how spending feels, even in households where income hasn’t materially fallen.
That kind of change is hard to spot if your main signals are transactions and a satisfaction score.
Those measures tell you what happened, but not what it meant to the customer; whether they were trading down reluctantly, changing what “good value” feels like, or simply choosing what feels safer.
You only hear that when you talk to customers about how they’re making decisions, not just what they bought.
As explored in Episode 20 on customer satisfaction measurement, the issue isn’t that these metrics are wrong. It’s that they are narrow. And when they become the dominant signal, the richer picture of who the customer is gets thinner.
Understanding retail customers requires the frontline
The most current customer intelligence in most retail organisations sits with the people who interact with customers every day.
A long‑serving store manager notices which regulars have changed their habits. A sales assistant hears which questions come up most often.
The pattern is there, but it rarely goes anywhere. It stays with the people on the shop floor, in quick chats at the end of a shift, in things a store manager mentions in passing, in the questions staff keep hearing. These useful snippets don’t make it back into the “official” view of the customer because there’s no straightforward route for them to return to the Support Centre.
Meanwhile, the approved customer profile was built elsewhere by someone else at a different time.
If it is reviewed at all, it tends to be reviewed away from the shop floor.
And so the two drift apart: what the profile says and what the frontline sees.
This is the territory explored in Episode 17 on the conditions that shape retail customer experience; the gap between how customer‑centricity is described at the centre and how it plays out in the places where customers actually are.
More often, it’s the result of systems that haven’t been built to carry customer insight in both directions.
Key takeaways
- Customer data and customer understanding are not the same thing. Data describes what customers did. Understanding clarifies what it means, and most retailers prefer to maintain the former rather than the latter.
- Customer profiles age. Without an active process for testing them against current reality, they keep being used long after the customer they describe has changed.
- High staff turnover removes institutional customer knowledge, the context and reasoning behind research, not just the people who held it. What’s left is the output without the thinking.
- Satisfaction scores are useful but narrow. When they become the primary way a business “knows” the customer, the picture thins, and the identity drift explored in Episode 24 can progress without leaving a trace in the metrics.
- Frontline teams often hold the most current customer intelligence. The question is whether there is any mechanism for that intelligence to reach the people making decisions, or whether it remains local and never reaches anyone.
Resources mentioned
- Centre for Retail Research — UK store closure projections, 2025
- Centre for Cities — UK retail footfall data, 2024
- CIPD / ONS Annual Population Survey — staff turnover in wholesale and retail, 2022–23
- British Retail Consortium CFO Survey 2026 — labour cost pressures in UK retail
- Levitt, B. & March, J.G. (1988) — Organisational Learning, Annual Review of Sociology — the competency trap concept
