Operating model design

Operating model design is the method for building or rebuilding how an organisation’s work is arranged: which capabilities exist, who owns each decision, where every step sits, and which systems carry it. It is the work of producing the model, not the model.

That line is worth holding. The business operating model page describes the artefact itself, covering people, process, technology, locations, data and governance.

This page is about the work of producing that artefact: what you inventory first, how you decide where work sits, and how you test a design before you commit to it.

Design is also not the calendar. An operating rhythm is the set of meetings and reviews that runs a model once it exists, and no cadence rescues accountability the design put in the wrong box.

Key takeaways

• Design is the method; the operating model is the artefact the method produces.
• Start from capabilities, not from the current organisation chart.
• Decision rights carry the most weight and attract the least attention.
• A design is unproven until one real flow has run through it — on live work, not on paper.

How it works

Operating model design works in a fixed order: establish what the organisation must be able to do, map how that work actually flows today, decide where each step should sit, name the decision rights, then test the design on one real flow before rollout.

The inventory comes first because job titles mislead. Organizational structure tells you who reports to whom, not what the company must be capable of doing, and designing around the chart reproduces the chart.

Mapping is the step teams rush. Process mapping shows the route a task really takes, and an end-to-end process view catches the handovers where work waits, which is usually where the cost hides.

Sequencing decides whether the redesign survives contact with people. Move the boxes before the systems and the decision rights follow, and staff rebuild the old routes informally inside a month, because nothing supports the new ones yet.

Placement is the commercial decision. Running a step in house, in a shared centre or at a provider moves supervision, working hours and cost at the same time, and a design recording only the headcount move will miss two of the three.

Testing comes before committing — run one live case through the new arrangement, with the named owners actually deciding, and you will find the handover that the diagram hides. Paper walkthroughs almost never do.

Public bodies formalise the method. The UK Cabinet Office’s government functional standard GovS 001 sets expectations for managing functions across government while leaving flexibility for each function on how the standard is met.

The work is also a recognised trade. The US Bureau of Labor Statistics counted 1,077,100 management analysts in 2025, whose listed duties run from onsite observation through to conferring with managers to ensure changes are working.

Examples

Redesigns are easiest to read in cases where one decision forced everything else to move. Three shapes come up repeatedly across sectors, and each one breaks a different component.

A bank reorganises around customer journeys. Account opening stops being four departmental steps and becomes one owned flow, which means the approval thresholds and the systems of record all have to be renegotiated at once.

A software firm moves from functional teams to product teams. The structure change is easy to announce — the hard part is that budgets, hiring approval and release sign-off were all built around the old functions.

A multinational consolidates procurement into shared services. The design holds only where country teams give up their local approval limits, and it fails quietly wherever one market keeps its own.

A hospital group redesigns outpatient scheduling. The capability is unchanged, but moving booking into one team shifts who may overrule a clinic’s diary, which is the decision nobody wants to write down.

In all three the structural change is the small part. What takes the time is agreeing the approvals, the data and the hours that have to move with it.

Related terms

• Business operating model: the artefact itself, where design is the method that produces it.
• Process mapping: recording how work flows today, the input a redesign depends on.
• Process analysis: testing a mapped process for cost, delay and failure points.
• Shared services centre: one common answer to the placement question, with its own governance load.
• Organizational structure: the reporting lines, which a model constrains rather than copies.

FAQ

How long does a redesign take?

A design for a single function usually takes eight to sixteen weeks, while the transition that follows runs far longer. Treat the design and the transition plan as two separate pieces of work with two separate owners.

What is the most common mistake?

Changing the structure and leaving decision rights untouched — the result is the previous behaviour wearing new job titles. It is why so many programmes report plenty of activity and very little movement.

Do you need external help?

Not always, but an outside view helps where the answer affects whoever is drawing the design. Internal teams rarely recommend removing their own layer, and rarely get thanked when they do.

How do you test a design before committing?

Pick the awkward cases rather than the clean ones, since the clean ones work under any design. A complaint, a cancellation and an exception will show you where the new routing stalls.

Where does outsourcing fit?

Sourcing is a placement decision inside the design, not a separate exercise bolted on afterwards. Deciding it late forces a second redesign, because the supervision, hours and handovers all have to be reworked around the provider you eventually pick.

Compare delivery models and provider options across Outsource Accelerator before you commit a step to an offshore site.