We’ve all sat in that meeting.
Operations defend the CMDB like the last wall between uptime and chaos. Enterprise Architects arrive with capability maps and target states. The PMO turns up with a spreadsheet temple of cost categories and gates.
Each discipline has a valid purpose - run the now, design the next, fund the change - but over time their tools hardened into bunkers. Detail piled on detail. Rules grew teeth.
The work became less about seeing truth together and more about protecting the system that protects us.
That’s control theatre: models and processes that look like control yet deliver very little of it.
The result?
- Slow decisions
- Stale data
- Duplicated modelling
- Portfolio bets made without live service visibility
We don’t need more completeness. We need shared visibility.
CMDBs, EA platforms, PMO cost models - these are lenses on one system. Keep the borders up and truth gets blurry.
Three Lenses, One System
These are not competing worldviews; they’re altitudes on the same model.
Operations / CMDB
Current-state fidelity for availability, change, compliance, audit. The horizon is now.
Enterprise Architecture
Capabilities, principles, patterns, target states, roadmaps. The horizon is next.
Portfolio / PMO
Funding, risk, benefits, governance. The horizon is investment.
More lenses exist - later articles cover these - but these three dominate the struggle.
How Turf Wars Show Up
Here are typical manifestations of turf protection:
Detail Gravity
Each team adds more fields to “prove value”. Signal: Attribute completeness dashboards. Cost: Faster staleness, slower impact assessment.
Gatekeeping as Risk Management
Controls and handoffs multiply. Signal: More checkpoints than decisions. Cost: Latency and workarounds.
Map Worship Over Territory
Slides and screenshots treated as truth. Signal: Meetings about the model, not the service. Cost: Bad bets, brittle change.
Metric Vanity
Counting CIs, artefacts, or gates. Signal: Big numbers, small insight. Cost: Misplaced effort, busywork.
Integration Tax
Translation layers everywhere. Signal: Monthly reconciliation marathons. Cost: Drift, duplication, friction.
Diagnosis: Control theatre. We’ve optimised for protecting our system rather than seeing the system.
Minimal Viable Truth (MVT): The Smallest Model That Makes Big Decisions
Completeness is the wrong goal. Decision-ready truth is the right one.
Define an MVT - the smallest, freshest model that reliably answers the ten canonical questions:
- What is it?
- Who owns it?
- Where does it run?
- What service does it enable?
- Who uses it?
- What does it depend on, and who depends on it?
- What controls and risks apply?
- What does it cost (order-of-magnitude is fine)?
- What changes are in flight or planned?
- How healthy is it right now?
Use the single post-it method: If the answer doesn’t fit on one side of a post-it, it’s too much.
If your model requires side files and hallway conversations, you don’t have visibility - you have trivia.
One Service Graph, Many Contracts
Unify on a single service graph and let roles view it through their own lenses.
Practically:
Stable IDs and simple relationships
Standardise on two verbs:
- A contains B
- A consumes B
Layered detail
Start with MVT, then add overlays for cost, risk, controls, contracts.
Event-driven freshness
Incidents, changes, deployments, gates automatically update the graph.
Federation over centralisation
Discovery, cloud inventories, EA repos, CI/CD, finance feed the graph.
Finance as an overlay
PMO cost models attach to the same IDs - not a separate spreadsheet temple.
APIs and webhooks
Push events, pull overlays, avoid batch reconciliation theatre.
The aim is simplicity, timeliness, clarity - and a graph that everyone can trust.
Metrics That Matter
Drop vanity metrics like:
- items discovered
- fields completed
- diagrams published
Instead measure:
- Freshness: % entities updated in n days
- Coverage: % top services with complete dependency chains
- Decision speed: time to impact assessment
- Cross-lens traceability: incidents ↔ capabilities, investments ↔ services
- Portfolio coupling: % business cases tied to live, owned services
- Outcome adoption: how often the model is used in CAB, QBR, portfolio
If these trend upward, your model is earning trust. If not, you’re adding nouns, not value.
Governance That Scales
When aligned to MVT concepts, governance becomes light and scalable:
- Single product owner for the service model
- Shared backlog for gaps, overlays, integrations
- Guardrails for “enough” - detail must justify its decision value
This ends multi-discipline sprawl. The model becomes a product, not a battleground.
In Practice
Everyone uses the same model - but through the lens suited to their discipline.
Following the chain of relationships across roles reveals real dependencies and risks. It surfaces truths normally hidden by turf boundaries.
It becomes Truth instead of Turf.
The OSM Angle: Making One System Usable
OSM treats the enterprise as a network of services, with the service model as the backbone of how value flows.
OSM enables roles to use their own lens without forking the truth.
Core Principles
- Service is the unit of value
- One backbone, layered views
- Two simple relationship verbs
- Event-driven freshness
- Decision over description
How OSM Resolves Turf Wars
- Shared IDs and language
- Automatic traceability
- One product-owned model
- Overlays instead of silos
The Service Intelligence Base (SIB)
The SIB is a practical realisation of this approach - a lightweight service graph with overlays for risk, cost, controls, and patterns.
Federated, event-driven, role-specific views. No rip-and-replace required.
Working Practices
- CAB and QBR use the same views
- PMO overlays create immediate run/change splits
- EA patterns expressed as to-be overlays with measurable deltas
What OSM Is Not
- Not a heavyweight framework
- Not a vendor platform
- Not another silo
It’s a way to unify what you already have - with clearer contracts and shared truth.
Turf Wars Are Futile
CMDB, EA and PMO aren’t rivals; they’re lenses on one service model.
Turf wars create control theatre that slows decisions and hides truth.
Unite around a service graph. Define Minimal Viable Truth. Wire events for freshness. Measure decision speed and traceability, not attribute counts.
Start small. Use it in anger. Retire vanity metrics.