Teams reconcile the same numbers repeatedly
Reports use different definitions, cut-off times or treatments of returns. Meetings become debates about which figure is correct instead of what action to take.
06 / Data Architecture
Know which data to trust, and who can fix it.
Clarify the ownership, meaning and movement of business data across your systems. Build a practical foundation for operations, reporting and decisions that depend on consistent information.
Independent advice. A clear scope. Decisions your team can act on.
OrderCustomerInventory
OriginAuthoritative sourceWhere is this fact created and corrected?
ContractDefinition, quality & freshnessWhat can downstream teams rely on?
UseOperations, analytics & AIThe same meaning across different tools.
Trace the number back to the decision.Shared meaning does not require one giant database.
Illustrative decision framework
Retail & DTC focusIndependent of platform vendorsAdvisory shaped around your decision
Which record should this decision trust, and why?
More data access does not resolve conflicting definitions. Reliable decisions require a clear account of where information originates and how it changes.
Ownership stops at the system boundary
Each application has an administrator, but nobody owns the business definition or the correction process when data crosses into another system.
New initiatives expose old inconsistencies
A customer experience, planning or AI initiative depends on information that is incomplete, stale or difficult to trace back to its source.
Get close to the work.
Make the decision clearer.
We work with the people who own the outcome and the teams who will deliver it. Your constraints shape the recommendation.
- 01
Start with a decision that needs better data
Choose a reporting or operational problem. Agree the entities, definitions and systems that influence it before expanding the scope.
- 02
Trace the information
Review representative records and existing mappings with business and technical owners. Identify where definitions, identifiers or timing diverge.
- 03
Assign ownership and correction paths
Agree authoritative sources at the appropriate level, quality expectations and resolution responsibilities. Turn the findings into a staged improvement plan.
A recommendation you can put to work.
Establish the definitions and responsibilities needed for the decisions in scope, with a realistic path to better data.
A domain ownership map
Define business owners and authoritative sources for key entities and attributes. Distinguish responsibility for meaning from responsibility for technical storage.
A critical data flow map
Trace creation, transformation and consumption across systems. Identify duplication, timing assumptions and points where meaning or identifiers change.
A quality and reconciliation framework
Specify relevant rules for completeness, freshness and consistency. Define who investigates exceptions, how corrections propagate and how unresolved issues are reported.
A prioritised improvement plan
Sequence work around business impact and dependencies. Recommend targeted changes to ownership, contracts, models or tooling with explicit acceptance criteria.
Put it to the test
“Available inventory” means three different things
Illustrative scenario: the warehouse reports physical stock, commerce displays sellable stock and planning uses a forecast adjusted for incoming deliveries. The figures differ for legitimate reasons, but their labels and timing obscure those differences.
The questions worth answering
- Which definition does each business decision actually require?
- Where are reservations, damaged stock and delayed updates handled?
- Who owns a correction when a record is wrong at its source?
The architecture makes the definitions and flows explicit, preserving useful distinctions while reducing avoidable reconciliation work.
Related case study
Read the case study Data readiness shows up in the workflow
Explore a related retail AI case study where the quality of operational evidence shapes what a workflow can safely do.
A few practical questions.
Scope, ownership and the next step should be clear before an engagement begins.
Does this mean putting all data in one database?
No. Different systems can remain authoritative for different entities, attributes or stages of a process. The goal is clear ownership and dependable exchange.
Will you recommend a new data platform?
Only if the requirements and evidence support it. Ownership, definitions and data contracts may need attention before additional tooling would be useful.
Can this support an analytics or AI initiative?
Yes. The scope can focus on the information a specific initiative requires. It identifies prerequisites and responsibilities; data engineering and remediation remain with your delivery teams.
Let's talkBring the decision
Bring the decision
you need to make.
Tell us what is changing, what is at stake and where you need an independent perspective. We will discuss the right starting point and agree the scope before work begins.
Still framing the problem?
The free Retail Architecture Risk Score can help surface where systems, ownership or delivery capacity need a closer look.
Start with the Risk ScoreThe work rarely stops at one boundary.
Where we work
Independent advice for your regional team.
Shared reporting and AI depend on reliable operational data. We trace the definitions, ownership and handoffs behind the records your teams need to trust.
Remote collaboration, with workshop and meeting arrangements agreed during scoping.