Interfaces multiply, incidents are hard to assign and one datum has several definitions.
CIO / CTO guide
Scope architecture, data and integrations before multiplying interfaces
Turn business needs into data contracts, criticality, ownership and an architecture roadmap.
For: CIOs, CTOs and architecture, applications, data and integration leadersWhat to clarify before discussing a solution
Which events and master data should be authoritative so systems remain decoupled without losing operational coherence?
The answer does not lie in an isolated feature: it is built by connecting rules, actors, data and evidence.A readable target architecture prioritised by critical flows, with contracts, observability and degraded modes.
Four decisions to make before configuration
These decisions make the scope testable and prevent structural choices from being discovered during the project.
Map flows by business criticality
The decision must specify the rule, its owner, the required data and the evidence used to accept it.
Assign each master datum to an owner
The decision must specify the rule, its owner, the required data and the evidence used to accept it.
Define expected latency, recovery and idempotency
The decision must specify the rule, its owner, the required data and the evidence used to accept it.
Test degraded modes before deployment
The decision must specify the rule, its owner, the required data and the evidence used to accept it.
A readable end-to-end flow
Scope is not a list of modules. It is a chain of events, ownership and decisions.
1Business capabilities
2Authoritative master data
3Events and contracts
4Security and continuity
5Observability
In the first scoping
- One representative flow and its useful variants
- Authoritative events and minimum data
- Roles, decisions and escalation times
- Acceptance criteria and baseline measurement
To sequence in waves
- Additional sites, activities or populations
- Rare variants that do not condition the pilot
- Automations whose rule is not yet stable
- Advanced dashboards after source reliability
To decide explicitly
- Map flows by business criticality
- Assign each master datum to an owner
- Define expected latency, recovery and idempotency
- Test degraded modes before deployment
From direction to measured improvement
- 1Navigator
- 2Personalised review
- 3Scoping workshop
- 4Wave-based deployment
- 5KPI measurement
Each stage produces a decision or evidence reusable in the next. Deployment remains business-led and verifiable.
Measure to decide, not to fill a dashboard
Every KPI needs a definition, a source, a cadence and an associated decision.
KPIFormulaSourceDecision
Critical data qualityComplete, valid and current data ÷ data checkedMaster data and control logsAddress rules and ownership at source
Integration latencyAvailability time − event creation timeExchange logs and timestampsSecure the flows that condition operations
Decision lead timeDecision date − signal dateAlerts, decisions and actionsShorten the path from signal to action
Service rateDemand served to commitment ÷ total demandDemand, commitments and actualsAlign priority, capacity and promise
06 · Business scenario
One critical flow as architectural proof
- Situation
- A company is preparing several solutions and fears a stack of fragile connectors.
- Approach
- IT selects one critical flow, formalises contracts and owners, instruments exchanges and tests incident, recovery and coherence.
- Target outcome
- Architecture choices are proven through real use before scaling.
Useful answers before the first discussion
01Where should we start in practical terms?
Choose a representative scope spanning “Business capabilities” and “Authoritative master data”, then document one normal case and one frequent exception. The first flow should be important enough to matter but contained enough to observe end to end.
02Do all existing systems need replacing?
No. Scoping starts with decisions, authoritative events and ownership. It then determines what should be retained, integrated, replaced or simply better governed, wave by wave.
03What data should be prepared before a workshop?
Prepare a few real cases, the volumes shaping the operation, the main variance reasons and the sources used to measure Critical data quality. The quality of examples matters more than the quantity of documents.
04How should the right pilot scope be chosen?
Choose a scope with an available owner, accessible data, a meaningful exception and a measurable outcome. Avoid both an overly simple case that proves nothing and an overly broad scope that dilutes learning.
05How can scope changes during the project be limited?
Make assumptions, interfaces, variants, exceptions and acceptance criteria explicit. Every new request can then be classified: essential to the wave, suitable for a later wave, or outside the objective. Discussion focuses on impact rather than intuition.
Next step
Turn this guide into a personalised review
Complete Navigator to position your context, then use the review as the starting point for a scoping workshop.
- 1Navigator
- 2Personalised review
- 3Scoping workshop
- 4Wave-based deployment
- 5KPI measurement
Complementary guides
Guides by suite
EMS Logistics suite
Transform transport & logistics
Connect order, warehouse, transport, delivery and invoicing around the same events and responsibilities.
- Service promise
- Warehouse execution
- Transport plan
Guides by suite
EMS Industry suite
Connect planning to the shop floor
Build decision continuity across demand, materials, capacity, execution, quality, maintenance and cost.
- Demand and schedule
- Materials and capacity
- Execution and WIP
Guides by suite
EMS Commerce suite
Unify omnichannel commerce
Connect catalogue, availability, order, payment, fulfilment, delivery, return and customer relationship.
- Catalogue and pricing
- Available-to-promise inventory
- Order and allocation