Consistent analytics depends on more than having the right data. It also depends on having a shared understanding of what that data represents. For one large-scale fast-food organisation, product categorisation had evolved independently across teams to meet different reporting and analytical needs. While these approaches worked within individual teams, they created manual effort, duplicated reporting and inconsistent definitions of seemingly identical product categories.
We worked with the organisation to design a centralised product categorisation framework within its data warehouse and BI layer. The objective was not simply to consistently categorise today’s products, but to establish a shared business language that could remain relevant as the organisation’s product range, underlying systems and use of data continue to evolve.
The organisation’s product data originated in operational systems designed primarily to support day-to-day business processes. While effective for that purpose, the way products were represented did not always align with how teams wanted to analyse performance. To bridge the gap, teams created and maintained their own product groupings. This introduced an important challenge: two reports could use the same category name but include different underlying products, resulting in different metrics for what appeared to be the same measure.
The solution wasn’t another report or another set of mappings. It was to establish common definitions that could be applied consistently across analytics. Working closely with stakeholders, we explored how products were grouped and discussed across the business, challenging differences in terminology, category boundaries and analytical requirements. From this, we designed a standardised framework for the data warehouse and BI layer, where categorisation logic could be centrally managed and reused.
Applying consistent categories to the current product range was only part of the challenge. For the framework to remain effective, each field also needed a clear definition, with agreed principles governing what its values represent. Without these guardrails, new products could gradually be categorised differently and the organisation would eventually find itself facing the same inconsistencies again.
Promotional products provided a useful test of this thinking. They are typically among the most frequently changing parts of the product range and naturally carry attributes relating to a particular promotion or campaign. The challenge was to distinguish these changing characteristics from the underlying attributes of the product itself. We worked with stakeholders to separate these concepts, ensuring core product categorisation remained focused on enduring product characteristics rather than allowing promotional terminology to influence fields intended for another purpose. Promotional attributes can then be represented separately without changing the fundamental definition of the product. This distinction creates a framework that is easier to apply consistently as new products are introduced.
The timing of this work was also important. The organisation is progressing through a migration of its underlying systems, meaning the structures supplying its analytical environment will change. Clearly defining product concepts independently of those systems provides a level of continuity through that transition.
Rather than allowing analytical definitions to be dictated by each new source structure, new data can be mapped back to established business concepts. This helps preserve consistency within the BI layer even as the technology underneath it changes. It also reduces the need to repeatedly rediscover historical business logic every time a source system or data structure changes.
There is another reason why explicit business definitions are becoming increasingly valuable: AI. As organisations look to AI to help people access, understand and analyse their data, providing access to the data itself is only part of the equation. Context matters too. What a field means, what each value represents, how different concepts relate to one another and which definitions the business has agreed to use all provide important context for interpreting data correctly. In legacy analytical environments, that context can be spread across reports, code, documentation and the knowledge of individual teams. Years of accumulated logic and exceptions can make seemingly simple business concepts surprisingly difficult to interpret.
By making product definitions explicit now, the organisation is building a stronger semantic foundation for future ways of using its data. Rather than asking future teams or AI-enabled tools to infer meaning from complex historical logic, agreed business context can be carried forward and reused.
At first glance, product categorisation can look like a relatively narrow data-modelling problem. Its impact can be much broader. Centralising the framework creates a path towards less manual effort for business teams, fewer competing reports based on different category definitions and a shared language for discussing product performance. It also establishes something more durable: a clear definition of important business concepts that is independent of any individual report, team or source system. Products will change. Systems will be replaced. New ways of interacting with data will emerge. The value of a shared business language is that it can provide consistency through all of them.
Better analytics does not always start with more data or another dashboard. Sometimes it starts by making sure the organisation agrees on what its data actually means.
If your organisation’s data tells different stories depending on who is analysing it, it may be time to establish a shared data language.
Contact the One51 team to talk through how we can help create trusted data models, consistent definitions and a stronger foundation for analytics.