Do Different Customer Experiences Really Need Different Data Platforms?

Persona work is a familiar part of customer strategy, and it often ends where it starts: a set of named, illustrated profiles on a slide, used mainly to align a marketing team on tone of voice. That's useful, but it leaves an important question unanswered: can the underlying systems actually serve these different people well, or does each persona quietly imply its own bespoke build? If you're running a persona workshop of your own soon, that's the question worth adding to the agenda before the slide gets signed off.

Working with a global travel-retail operator on its future loyalty and digital experience, we used persona work to answer that harder question directly, rather than stopping at the slide.

The proposition: stress-testing one architecture against four emotional registers

We built out four distinct traveller personas, each with a different relationship to travel retail. A budget-conscious Gen-Z traveller who scrolls past generic in-store promotions but responds to gamified, personalised content: an AI "pack-like-a-pro"experience, a treasure hunt earning points toward a reward. A frequent business traveller who wants friction removed entirely: a pre-filled cart synced to their calendar, a fast-lane QR code, a locker pickup, expenses filed automatically by the time they lands. A family travelling with young children, for whom a "Family Holiday Kit" reserved at booking and a play-corner pickup turn an historically stressful airport experience into something closer to a treasure hunt. And a luxury collector, for whom the experience is built around scarcity and ritual: a private livestream preview of a rare item, a personal shopper, a silk-boxed thank-you note after the trip.

These four personas sit about as far apart emotionally as retail personas get: one optimised around cost-conscious gamification, one around zero-friction efficiency, one around family logistics, one around exclusivity and ceremony. The test we were actually running was whether a single underlying capability stack, one identity graph tracking who the customer is across channels, one prediction engine inferring what they're likely to want, one dynamic offer service deciding what to show them, one fulfilment layer capable of a locker, a gate delivery, or a home shipment, could flex across all four, or whether serving the luxury collector properly would require building something entirely separate from what serves the budget-conscious traveller.

The answer, worked through in detail, was that it could. The same components, reused rather than duplicated, could plausibly deliver a treasure hunt for one traveller and a velvet-quiet private salon appointment for another, because what changed between personas was how the shared layer was configured and triggered, not the existence of four parallel systems underneath.

What this means for any retailer serving genuinely different customer segments

This matters most for retailers whose brand deliberately spans very different kinds of customers under one roof: value-conscious and premium, casual and serious, child and adult. The tempting assumption in that situation is that serving each segment well requires its own dedicated system, because the experiences look so different on the surface. That assumption is expensive, and often wrong.

The more useful question to ask before committing to separate builds is the one this exercise was actually designed to answer: does the underlying identity, prediction and fulfilment layer need to differ per segment, or does it just need to be configured differently? If it's the latter, and here, it was, the investment case shifts substantially. One well-built orchestration layer serving four segments well is a fundamentally different proposition, in cost and in maintainability, from four bespoke systems each serving one segment.

Conclusion

A persona exercise that stops at a slide tells you how to talk to different customers. One that stress-tests a shared architecture tells you whether you can actually afford to serve them all well.

Key takeaways

• Persona work is usually treated as a marketing alignment exercise; it's also a legitimate way to stress-test whether a shared architecture can serve genuinely different customers.
• We tested four emotionally distinct traveller personas — a budget-conscious Gen-Z shopper, a frequent business traveller, a family, and a luxury collector — against one underlying capability stack.
• The same identity, prediction, offer and fulfilment layer could plausibly flex from gamified eco-quests to white-glove concierge service, reused rather than duplicated per segment.
• For any retailer spanning very different customer segments under one brand, the key question is whether the underlying layer needs to differ by segment, or just needs to be configured differently.
• Getting the shared layer right changes the investment case from four bespoke systems to one platform, configured four ways.

If your brand spans genuinely different customer segments and you're weighing whether that means separate systems, we're happy to work through whether a shared architecture could do more than it looks like on paper.