Business Capability Model
Definition
Business Capability Model
A business capability model is the defined, levelled taxonomy of what an organisation is able to do, with each capability named, described and owned. The model is the schema; a map is one view of it — confusing the two produces pictures nobody maintains.
A model is durable infrastructure rather than a deliverable — built once, corrected occasionally, and used repeatedly by cost analysis, sourcing decisions, application portfolios and investment planning.
Size is the other discipline. A first level running beyond twenty capabilities usually means processes have crept into the taxonomy, and processes multiply in a way that abilities do not.
Its discipline lies in the definitions. Each capability needs a one-line description precise enough that two people can agree whether a given activity belongs inside it, which is harder than naming the capability itself.
Without that precision the model quietly stops working. Costs attributed to ambiguous capabilities cannot be compared, and two teams will map the same spending to different boxes.
Key takeaways
- The model is a defined taxonomy; a capability map is a rendered view of it.
- Every capability needs a name, a description, an owner and a level.
- Mutually exclusive definitions are what make cost and system attribution reliable.
- One model serves many uses: costing, sourcing, portfolio and investment decisions.
How it works
Models are built top down and validated bottom up. A first level is drafted from the business model, decomposed into second and third levels, then tested by attempting to place real activities, systems and costs into the boxes.
The placement test is what finds the defects. An activity that fits two capabilities means the definitions overlap — and one that fits none means something is missing from the taxonomy.
Stability is a design goal rather than a happy accident. A model rebuilt at every reorganisation loses the comparability that made it worth building, so structural changes are made rarely and deliberately.
Ownership is assigned capability by capability. The owner arbitrates whether a borderline activity belongs inside their boundary, and unowned capabilities are where definitions drift first, because nobody is asked to defend them.
| Attribute | Purpose | Consequence if missing |
|---|---|---|
| Name | Shared vocabulary | Teams use different words |
| Description | Boundary between capabilities | Attribution becomes arbitrary |
| Level | Consistent depth of detail | Uneven decomposition |
| Owner | Accountability for judgements | Nobody resolves disputes |
| Linked systems | Ties the estate to the model | Portfolio work loses its anchor |
Architecture standards supply the method and the vocabulary. The Open Group’s TOGAF standard is described by its publisher as ensuring “consistent standards, methods, and communication among Enterprise Architecture professionals”.
Public guidance pushes toward the same reuse. The UK Technology Code of Practice asks organisations to “avoid duplicating effort and unnecessary costs” by sharing and reusing technology, data and services across boundaries.
Examples
A model pays for itself when several unrelated exercises draw on the same definitions. The three cases below use one model for costing, sourcing and system attribution.
A services group attributes every cost line to a capability. That single mapping makes key performance indicator (KPI) reporting comparable across business units for the first time.
An insurer links each application to the capabilities it supports. Duplication becomes visible immediately, and process design outsourcing can be scoped by capability rather than by system.
A retailer reports performance by capability rather than by department. Its balanced scorecard survives two reorganisations because capabilities did not change when the org chart did.
Related terms
The model underpins several disciplines that consume its definitions rather than produce them. The entries below all sit downstream of it, using the taxonomy the model defines to scope their own work.
- Six sigma: improvement work scoped against a capability the model names.
- Transformational outsourcing: handing over whole capabilities rather than tasks.
- Program manager: plans delivery against the capabilities being changed.
- Business intelligence outsourcing: reporting built on the model’s structure.
FAQ
What is the difference between the model and a map?
The model is the underlying definition set. A map is a picture generated from it, usually coloured to answer one question at one moment.
How many levels should a model have?
Three for most organisations. A fourth is built only where a specific decision needs it, because deeper levels are expensive to keep current.
How is a capability worded?
As a noun phrase describing an ability, such as claims settlement or demand planning. Verb phrases tend to describe processes instead.
How is the model validated?
By placing real activities, costs and systems into it. Anything that fits two boxes or none reveals a definition that needs work.
Who owns it?
Enterprise architecture, with named business owners per capability. A model without business owners becomes an architecture artefact rather than a shared reference.
Do industry reference models help?
As a starting point, yes. They shorten the first draft, but they need local adjustment or the definitions will not match how the organisation actually operates.
Read more architecture and operating model guidance at Outsource Accelerator.
Independent
