Have Any Questions?
Get in Touch
Have Any Questions?
Get in Touch
Data Platform Architecture & Design

The right architecture decision, made once. Before the build begins.

Bad infrastructure and data architecture compounds. The decisions your team makes informally in the first two weeks of a build are the ones you pay to reverse in month six. We make those decisions deliberately, completely, and on record, before the first sprint begins.

🏛️
Architecture
Decision Record
2 weeks • Fixed price

Not a slide deck. An architecture document your team builds from.

The output is a complete, implementation-level architecture decision record. Every design decision is made, documented and defensible before the first sprint begins.

What the document covers
  • Current state review — technical debt scoring, identified anti-patterns
  • Target state architecture on Fabric or Databricks — medallion layers, security, governance
  • Data modeling framework, ingestion strategy, transformation framework
  • Platform recommendation if undecided between Fabric and Databricks
  • Phased implementation roadmap with milestones and resource plan
The document is yours. Your team builds from it — with TrueNorth or independently.
How the Engagement Works

Two stages. One clear path. No surprises at either end.

Stage 1 is where you get certainty. Stage 2 is where you get the platform. You don’t commit to Stage 2 until Stage 1 gives you everything you need to make that decision confidently.

 
Greenfield

“We’re building a data platform from scratch and we need to get the foundation right first time.”

No platform exists yet. You have a mandate to build a modern cloud-native data platform and your engineering team needs a complete architecture to build from, not a high-level diagram they have to interpret into design decisions every sprint.

  • Platform selection, full reference architecture from Bronze through Gold, data modeling standards, ingestion patterns, governance and security model. Everything the build team needs to start without gaps.
Platform Replacement

“Our current platform is so poorly architected that we need to start over, not extend it.”

A platform exists but has grown without design, ad-hoc schema additions, no medallion structure, brittle pipelines, no governance. The team knows it needs to start over. The engagement designs the replacement properly so the same problems are not built into the new platform.

  • Clean break target architecture, technical debt inventory of the existing platform to scope the replacement, migration sequencing from old to new.
Architecture Redesign

“The platform works but adding new use cases keeps breaking what we already have.”

The platform is in use but has scaling and reliability problems rooted in poor architecture decisions. The data team spends most of its time firefighting rather than building. The engagement identifies what is structurally wrong and designs the fix, without requiring everything to stop.

  • Current state technical debt assessment, targeted redesign of the layers causing the problems, phased remediation plan that maintains business continuity throughout.
Platform Chosen, Not Designed

“We have Fabric / Databricks licensed and provisioned. We don’t know where to start.”

The platform has been chosen, often by a corporate licensing decision or executive mandate, but nobody has designed the architecture for your specific workloads. The platform is ready. The engineering team is not sure what to build on it or in what order.

  • Workload-specific architecture design, medallion layer configuration for your data profile, ingestion and transformation framework, semantic model structure. Accelerates time-to-first-value significantly.
What the Engagement Produces

Seven sections. Every architecture decision your build team needs made.

The output is not a slide deck. It is a complete architecture decision record, the format engineering teams work from directly. Every section is produced in 2 weeks and covers a category of decisions that, if left unmade, will be made inconsistently during the build by whoever is working that sprint.

Current State Assessment
Platform inventory, technical debt scoring, identified architectural anti-patterns and their root causes. For greenfield customers: data source inventory, volume and velocity profile, use case analysis that informs every design decision that follows.
Platform Recommendation
A structured recommendation based on your specific workload profile, existing Microsoft licensing, team capability and long-term roadmap. Not a generic vendor comparison, a specific, defensible recommendation your CTO can present to the board.
Target State Architecture
Medallion layer design (Bronze/Silver/Gold zones), security and access model, data governance framework, monitoring and observability approach. The complete structural design your engineering team builds to, with enough specificity to build without coming back for design decisions.
Data Modeling Framework
Entity design principles, naming conventions, surrogate key strategy, slowly changing dimension approach, aggregation patterns. Without a modeling framework, every engineer makes different decisions. With one, the data model is consistent, navigable and extensible from day one.
Ingestion Strategy
Source system connectivity patterns, batch vs. streaming decision framework, incremental load design, error handling and retry logic approach. The ingestion layer is where most data platform performance problems originate, designing it properly upfront prevents the most common class of rework.
Transformation Framework
Which transformations happen at which layer, dbt or native Fabric/Databricks tooling decision, testing strategy, documentation standards. The transformation framework determines whether your Silver and Gold layers are trustworthy, or whether business users stop trusting the platform within 6 months of go-live.
Phased Implementation Roadmap
Phases with clear deliverables, milestones, resource requirements and success criteria. The sequence is designed to deliver early value, business users see working data in the platform within the first phase rather than waiting for a complete build. Includes dependency mapping so the team knows what must be built in what order and why.
0
Architecture document sections, every decision your build team needs made upfront
1
Year TCO model included in every Blueprint Sprint, legacy vs. cloud
Starting points covered — greenfield, replacement, redesign, platform chosen
Inside the Architecture Document

Built for engineers to build from. Not for leadership to file away.

The document is structured as an architecture decision record, the format engineering teams actually use. Every section has enough specificity that the build team can work from it without returning for design clarification every sprint.

Section 1

Current State & Technical Debt Inventory

What exists, what it costs the team to maintain, what architectural decisions created the current problems. For greenfield: data source inventory and use case analysis.
Section 2

Platform Decision Record

Microsoft Fabric vs. Databricks — recommendation with structured rationale, evaluation criteria and decision log. If platform is already chosen, this section documents why and confirms the architecture fits the choice.
Section 3

Target State Architecture

Medallion layer design, zone boundaries, security model, access control framework, governance approach, monitoring strategy. The structural blueprint for the platform.
Sections 4 & 5

Data Modeling & Ingestion Framework

Modeling standards, naming conventions, key strategies. Source connectivity patterns, batch vs. streaming decisions, incremental load design, error handling approach.
Sections 6 & 7

Transformation Framework & Implementation Roadmap

Layer-by-layer transformation decisions, tooling selection, testing strategy. Sequenced build phases with milestones, dependencies, resource requirements and early-value delivery points.

Why an architecture decision record and not a slide deck? Slide decks describe architecture. ADRs specify it. An engineering team receiving a slide deck still has to make hundreds of design decisions during the build. An engineering team receiving an ADR has those decisions already made, documented, reasoned and defensible.

The practical difference is weeks of design-during-build conversations that do not happen because the answers are already in the document. That is where the 2-week investment pays back fastest.

The document is also a communication tool. The CTO presents Section 2 (platform decision) to the board. The Head of Data uses Sections 3–5 to onboard new engineers. The project manager uses Section 7 (roadmap) to track progress. The same document serves every audience.

What this engagement does not include

  • The build itself — Architecture & Design produces the specification; the build is a separate engagement
  • Data Platform Migration — if you have a named legacy platform to migrate, the Migration Blueprint Sprint is the right entry point (or runs after this engagement)
  • BI and reporting layer — Power BI report design and semantic model build are covered by BI Modernization
  • Data engineering implementation — pipeline build, ETL/ELT development and operational management are separate engagements
What Comes Next

The architecture engagement unlocks everything else.

A properly designed platform architecture is the foundation every other TrueNorth product deploys faster and more reliably on. What comes next depends on your starting point.

 
If you have a legacy platform to migrate

Data Platform Migration Blueprint Sprint

The architecture engagement clarifies the target state. The Migration Blueprint Sprint then scopes the route from your legacy platform to it, TCO model, risk register and phased migration plan in 2 weeks.
 
Once the platform is built

TrueFinance, TrueSCM or TrueRevenue FastTrack

FastTrack products deploy faster and with better performance on a properly designed platform. The architecture engagement creates the foundation. FastTrack delivers the reporting and analytics layer on top of it.
To go deeper on engineering patterns

BI Modernization & Data Governance

Architecture & Design establishes the platform layer. BI Modernization addresses the reporting and semantic model layer. The Data Governance Assessment builds the comprehensive governance framework on top of the architecture the engagement started.
 
Get Started

Start with a platform architecture session.

60 minutes. We walk through your starting point, your platform situation and what the 2-week engagement produces for your specific case. You leave with a clear picture of what the architecture document covers and what your build team gets from it.

US Office

SA Office