Beyond the Data Swamp:
In a nutshell…
- Total data centralisation often creates more confusion than clarity
- AI is emerging as the connector that makes existing systems work harder
- The real blockers aren’t technical, they’re cultural and human
- Focus on linking what matters, not everything you can
- Data integration succeeds when teams align on purpose, not just platforms
For all the talk about unified data, most retailers are still living in a world of silos. Systems don’t talk, teams don’t align, and the “single customer view” everyone dreams of too often turns into a data swamp: huge, murky, and impossible to navigate. The irony is that the problem isn’t data itself, it’s how organisations think about it.
At Tech Leaders Europe in October 2025, Hive members shared the same frustration: everyone wants joined-up insight, but few agree on what “joined up” should actually look like. Here’s what they’re learning, and how they’re starting to turn data from a headache into a competitive advantage…
The challenge: connection without control
Retailers are operating across more platforms, partners and geographies than ever. That means thousands of data inputs that are often inconsistent, incomplete or misaligned. One Hive member summed it up neatly: “We’ve got 1,500 size charts. Every brand has its own logic. Fixing it manually would take years.”
The temptation is to pour everything into one big system. The result is centralisation without structure. Disconnected data doesn’t become useful just because it sits in the same bucket. In fact, retailers say it often gets worse: harder to find, slower to query, and full of duplication.
The better question isn’t “how do we get everything into one place?”, but “how do we make sure the right things talk to each other?”
What are Hive members doing?
Selective integration beats total unification
Businesses chasing a single customer view are learning to narrow their focus. Linking a handful of core datasets, such as product, price, customer and inventory delivers more value than trying to connect everything. Smaller, better-connected systems are also easier to govern and maintain.
AI is bridging gaps that tech teams can’t
Artificial intelligence is helping retailers clean, map and interpret data without forcing massive rebuilds. One member described using AI to automate the mapping of inconsistent sizing charts, transforming what used to be months of manual work into hours. Others are training models to enrich metadata for search and personalisation, allowing them to extract insight from what they already have rather than starting from scratch.
Over-protection kills progress
Several members admitted that their predecessors went too far on data security. Entire datasets were locked down, encrypted and isolated (“for GDPR reasons”), leaving marketing and customer teams unable to extract any value. The consensus: anonymise where you must, but don’t design for fear. Design for use.
Unified data unifies people
When data becomes accessible across functions, so does decision-making. Promotions stop clashing with stock levels, fulfilment teams understand trading priorities, and everyone works from the same reality. The commercial impact is real: faster reactions, fewer mistakes and less internal friction.
Culture is the hidden barrier
Technology can’t fix teams that don’t talk. Many retailers still structure data ownership by department rather than outcome, which means the “truth” changes depending on who you ask. The businesses making the most progress are embedding cross-functional product and tech teams, and teaching engineers to solve problems through communication, context and collaboration.
Next steps
1. Start with purpose, not plumbing. Before investing in another data platform, define the business decisions it needs to enable
2. Connect what counts. Link the few datasets that drive the biggest operational and customer impact
3. Let AI do the heavy lifting. Use it to match, clean and contextualise data where manual work is impossible
4. Treat data as a people project. Create shared KPIs and language across teams. The best dashboards are useless if no one agrees on what they mean
5. Build for access, not perfection. Perfect data doesn’t exist; accessible, actionable data does
Retailers don’t need another dashboard, they need a way to see clearly. The winners in the next phase of retail won’t be those who own the most data, but those who can make the most of what they already have.
