Organisations sometimes skip the think big part in favour of containing scope, moving fast and demonstrating value. It looks like progress. It often is — briefly. Then the organisation shifts, priorities change and we find ourselves painted into a corner.
Conversely we have all heard the stories about over engineered solutions – architecturally brilliant on paper – building for something that may never come – eventually too complex to maintain – a bottleneck for the business.
Both are liabilities, and both preventable.
The good news is – we dont have to choose between speed and longevity. We can move fast without compromising the future — designing architectures that meet today’s goals while evolving deliberately with the organisation. We just have to ask the right questions.
The decisions that matter most aren’t technical
Organisations that extract durable value from their data platforms are those that understood their own context before committing to architectural patterns. Medallion architectures, lakehouse models, data mesh, warehouses — these are interesting topics that we tend to dive straight into. But without context, they are distractions.
-
Lakehouse or warehouse? It depends
-
Kimball or Datavault or 3NF? It depends
-
Centralise or federate? Mesh, fabric? It depends
-
ETL or ELT? It depends
Their value and applicability is determined by how well they fit the organisation applying them and where it wants to go.
Before domain design, before platform selection — understand the non-technical factors that can make or break for your initiatives.
-
What are your strategic objectives — the ones you say out loud, and the ones you don’t?
-
How is your organisation structured, and how does that shape the way data flows and decisions get made?
-
What are your team’s real capabilities, and how do teams actually work together across domains?
-
How mature is your data governance, and what change is realistically possible?
-
How will this be funded, and what behaviours and incentives does that create?
-
What principles are truly non-negotiable — and what are you genuinely willing to trade off? What are we optimising for?
-
What other initiatives are in flight that are dependent on, or will impact this initiative?
Ask harder questions earlier, be honest about the answers — and treat what you learn as architecture inputs just as important as the technical ones.
Without structure, these conversations rarely happen — or they happen too late. This is what Level 0 of the Intuitas Data Intelligence Blueprint is all about.
– Benson