Five stages, and one where most stall
Every capability centre that becomes genuinely valuable passes through the same five stages. Knowing which one you are at tells you which problem to solve, and stops you applying the wrong fix for a year.
The wrong fix costs a year and some credibility
A centre stuck between stage two and three needs something entirely different from one stuck between one and two. Most interventions fail because the diagnosis was never made.
Capability centres do not improve gradually. They move between states, each with a specific condition that has to be met before the next becomes possible, and they stall at the boundaries rather than in the middle. The stall is almost always at the same place, and for the same reason.
The value of naming the stages is diagnostic. It tells you what to change, what evidence would show it worked, and roughly how long that should take. It also tells you when a centre is being asked for something it is structurally not yet able to give, which is the most common source of unfair disappointment on both sides.
What each one actually means
The exit condition is the useful column. It is what has to be true before the next stage becomes available, regardless of how long you have been at this one.
| Stage | What the centre decides | Exit condition | Typical duration |
|---|---|---|---|
| 1. Establishment | Nothing yet | Team hired, access granted, first change shipped | 1 to 2 quarters |
| 2. Delivery | How to build a specified thing | Completes work without repeated clarification | 1 to 2 quarters |
| 3. Ownership | How to solve a stated problem | Resolves ambiguity locally instead of escalating | 3 to 6 quarters |
| 4. Outcomes | What to build for a given result | Sets its own priorities within an outcome | 4 to 8 quarters |
| 5. Contribution | What matters, and proposes it | Not an exit; a state the centre holds | Ongoing |
The jump from stage two to stage three is the one that matters and the one that fails. Everything before it is logistics; everything after it depends on it.
What actually happens at each stage
Establishment
Entity, compliance, infrastructure and the first senior hires. Nothing meaningful is delivered and nothing should be measured except readiness. Measuring output here produces activity theatre and teaches the team the wrong thing about what is valued.
Delivery
The centre executes specified work correctly. This feels like success and it is a waypoint. A centre that settles here permanently is a cost arrangement, and it will be assessed as one when the savings case runs out.
Ownership
The centre owns named systems and resolves ambiguity locally. This requires higher seniority and the removal of onshore approval gates, both of which make visible metrics worse before they make anything better. This is where centres stall.
Outcomes
The centre is given a result rather than a system, and decides what to build to achieve it. Requires genuine business context, which most organisations never deliberately invest in transferring.
Contribution
The centre originates work the business adopts. It has an opinion, and there is a route for that opinion to be acted on. Few centres reach this, and those that do took three to five years.
Why stage two to three is where it breaks
This transition is not harder technically. It is harder politically, and it is the only one that requires somebody onshore to give something up.
Crossing into ownership requires two changes at once. The seniority mix has to rise, which makes cost per head worse immediately and visibly. And onshore approval gates have to be removed, which means named individuals give up control over decisions they are currently accountable for.
Both changes make the reported numbers worse for roughly two quarters before anything improves. Escalation rate falls first, at around quarter two. Genuine ownership follows at about quarter four. An organisation that evaluates the change after one quarter sees higher cost and no benefit, and reverses it precisely before the evidence arrives.
The practical protection is to say this in advance. State that cost per head will rise, that delivery metrics will be flat for two quarters, and that escalation rate is the leading indicator to watch. A dip that was predicted reads as a plan; the same dip unannounced reads as a failure.
What has to change
Seniority weighting rises, approval gates come off, and the centre is given a named system with a transfer date rather than a promise of more responsibility later.
What to watch
Escalation rate, monthly. It is the metric that cannot be improved without the centre genuinely knowing more and being permitted more.
What it will look like
Cost per head worse, delivery flat, escalation starting to fall by the end of quarter two. If escalation is flat at six months, the gates were never actually removed.
How long to allow
Four quarters. Judging it at one is the single most common way a working transition gets reversed.
Which stage are you actually at?
Stated intent is a poor guide, because almost every centre describes itself as a partner. These questions produce a more honest answer.
- When the team hits genuine ambiguity, does it decide or does it wait?
- How many systems does it own with no onshore approval gate? A number, not a description.
- Has escalation rate fallen over the past year, or only been discussed?
- What did the team build last quarter that nobody asked it to build?
- What specifically would have to be rebuilt if the centre closed tomorrow?
- Has the business ever adopted something the centre proposed?
Most centres are one stage behind where leadership believes. That gap is not dishonesty. It is that stage three is defined by the removal of approval gates, and gates tend to persist informally long after they have been formally withdrawn. The question that settles it is whether the team waits.
The distribution is not what most people expect
Establishment and delivery are comparatively easy and comparatively quick. Almost all of the difficulty, and almost all of the population, sits at one boundary.
Illustrative distribution based on the pattern we see rather than measured research. The shape is the point: the population concentrates just before the transition that requires onshore control to be given up.
The concentration at stage two is not a talent problem. Stage two is a comfortable place to be: the centre is competent, delivery is predictable, and nobody is failing at anything. There is no natural pressure to leave it, because a centre performing well against the targets it has been given does not look like it needs changing.
The pressure arrives later and from a different direction, when the savings case runs out and somebody asks what the centre would cost to replace. A centre at stage two has no good answer to that question, and by the time it is asked, building one takes four quarters that nobody has budgeted.
Frequently asked questions
Can a centre skip stages?
It can start at stage three, which is the strongest argument for building small and senior. A team staffed with genuine seniority and given ownership from day one never has to make the hardest transition. What cannot be skipped is the credibility that stages three and four produce, which is why stage five is unreachable directly.
How long should the whole journey take?
Centres that reach stage five typically take three to five years. Stage two to three, the critical transition, takes three to six quarters when treated as a deliberate governance change and never completes when treated as something that will happen naturally.
What if we have been at stage two for three years?
That is common and it is recoverable. The intervention is specific: raise the seniority mix, transfer one consequential system with a date and no approval gate, and change what appears first in the review. Expect four quarters before it is visible.
Is stage five realistic for us?
For most centres, no, and that is a reasonable outcome. Stage four is a strong and durable position. Stage five requires the receiving organisation to change as well, and that is usually the binding constraint rather than the centre s capability.
Can different teams in one centre be at different stages?
Frequently they are, and treating the centre as a single unit hides it. Assess by team. A centre with one team at stage four and three at stage two has a very different problem from one where everything sits at stage three.
Do centres ever go backwards?
Yes, and usually without anyone deciding to. A new onshore leader reinstates gates, an incident triggers a control response that is never relaxed, senior attrition removes the judgement the ownership rested on, or cost pressure thins the seniority mix. Reassess the stage annually.
What is the single best indicator of progress?
Falling escalation rate alongside rising ownership. Almost every other positive metric can be produced without the centre actually improving; this pair cannot.
Where does a Nano GCC start?
At stage three, by design. Small, senior and scoped around ownership from the first hire removes the transition that most centres never complete. That is the whole structural argument for the model.
Related reading
Find out which stage you are actually at
Tell us what your centre owns and how often it escalates. We will give you an honest assessment of the stage, the specific intervention that matches it, and what evidence would show it worked.