Six Operating Companies, Three Months and One Governance Body the Client Chairs

Most programme designs answer the governance question by adding a body. There is a working group that does the work, and above it a steering group that receives an account of the work, and the second one exists largely because the second one always exists. Nobody has to argue for it. Arguing against it is what costs political capital, which is why almost nobody does.

A proposal I worked on took the other option and wrote the reason down. It was a competitive response for the founding engagement of a European composite insurance group's group-level data analytics function: six operating companies plus the group centre, three phases, three months end to end. On the planning slide, in the middle of a Gantt chart, sits one sentence of governance design. The project is steered by a core team led by the head of the analytics centre of expertise, with agreed representatives from the countries in scope, and the proposal assumes the centre's sponsor fulfils the need for a steering group in this phase.

That is the whole apparatus. One body, chaired by the client, with the sponsor named as a standing substitute for the layer above it.

Everything below is a design in a bid document. No visit was run, no core team was convened, and no sponsor stood in for anything in the copy I hold.

The chair is the person who inherits the output

A discovery engagement does not produce a system. It produces a plan that somebody else has to fund, staff and run after the consultants have gone. In this case the somebody was already known by name: the head of the analytics centre of expertise, whose function the engagement was being run to establish.

Put that person in the chair and the handover problem disappears, because there is no handover. The roadmap does not arrive at the client at the end of month three as a document to be received, reviewed and gradually renegotiated. It accumulates in a body they have been running since month one. Every ranking argument, every country that wanted its own thing, every candidate that was cut, all of it happened in front of the person who will later be asked to deliver what survived.

The alternative shape is the familiar one, and it fails in a familiar way. The supplier chairs a working group, produces the analysis, and presents it to a client steering group at the end. What the client owns is then a recommendation, and a recommendation is a thing you can agree with without being committed to.

There is a second effect that matters more at group level than at country level. The countries are in the room. This is a group with operating companies across several markets, and the fastest way to produce six local roadmaps that cannot be added together is to interview six countries separately and aggregate the results centrally. Representatives on the steering body means the normalisation argument, whose optimism gets trimmed against whose, happens with all parties present rather than in a spreadsheet at the centre afterwards.

What the arithmetic does to a second body

Three months, three phases. Preparation in the first month, the six country visits compressed into the second, aggregation and the final report in the third. Each operating company gets one standardised three-day visit, and the anatomy of that visit, four stages of which only one happens with the team on site, is worked through in Only One of a Country Visit's Four Stages Happens While the Team Is There. What matters here is not what the visit contains but what its position in the calendar does to anyone trying to govern it.

Count the sittings a monthly steering group would get. Three, at best, and probably two that fall inside the working period rather than on its edges. The first lands during preparation, when there is nothing yet to steer except an approach the tender already bought. The second lands somewhere in the middle of the visit run, by which time three or four countries have been visited and their sessions have closed. The third lands after the aggregation is drafted.

A steering group meeting monthly would sit for the second time in the middle of the visit run, which is after every question it existed to settle has already been settled in a room in another country.

That is not a body with levers. It is a body receiving reports, and a governance layer that can only receive reports adds calendar friction, pack preparation and a second set of opinions arriving after the decisions they concern. The plan removes it and spends the recovered attention on a core team that meets continuously across all three months as the project management mechanism, which is the only cadence fast enough to matter when six markets are covered inside a single month of the plan.

6

Operating companies plus the group centre, in scope

3

Months from preparation to final report

1

Governance body in the design

0

Separate steering groups proposed

The joint this design cannot carry

I would not write this up as a pattern to copy without naming what it gives away, because a client-chaired body with nothing above it has one obvious weakness and the source does not address it.

The chair is the head of the function being created. The output is a ranked roadmap and an investment case for that same function. There is no independent party in the design whose job is to ask whether the roadmap is shaped by what the group needs or by what the new centre is well placed to deliver. Those two things overlap heavily at the start, which is exactly why the gap is easy to miss.

The sponsor is the intended counterweight, and here the drafting is thinner than the rest of the plan. The core team gets a drawn line across every month of the Gantt. The sponsor gets a sentence in an assumption. A counterweight without a cadence is an escalation path, not a check, and an escalation path only works when somebody chooses to escalate, which in practice means the chair choosing to escalate a decision they made.

For a phase that produces a plan, I think that is an acceptable trade. Nothing irreversible is being spent, the countries in the room are a partial check on central self-interest, and the investment case still has to survive a board that was never inside this process. For the phase after, where the plan becomes budget and headcount, it is not acceptable, and the phrase in this phase is the source acknowledging as much.

What I would write differently now

Two changes, one procedural and one about what the meeting is for.

Put an expiry date on the assumption rather than a phase name. In this phase is unambiguous while everyone remembers which phase they are in, and ambiguous the moment the discovery work rolls into a first funded build with the same people and the same standing meeting. The sentence I would add says that the sponsor stand-in lapses at the point the first item of the roadmap draws money, and that the body which replaces it has a named budget lever rather than a reporting line.

The second change is about the standing meeting. A core team meeting running continuously across three months absorbs most of its time in status: where the visits are, what came back, what the ranking looks like. Almost none of that needs a meeting now. A candidate log, its scores and the current ranking maintain themselves between sessions once the session output has a schema. What that frees is not calendar, it is agenda. A body that arrives already knowing the state of the ranking can spend its time on the two or three candidates where countries disagree, which is the only part of the work that needed all of those people in the same call.

Neither change touches the part worth taking. The person who will own the plan should chair the body that makes it, the people who will execute it should sit on that body, and a second layer whose meetings land after the decisions is not governance you are missing.

Source: a single planning slide from a competitive proposal written for the founding engagement of a European composite insurance group's group-level data analytics function. The slide carries the phase plan, the country visit schedule and one sentence of governance design: a core team led by the client's own centre of expertise head with country representatives, and a stated assumption that the centre's sponsor fulfils the need for a steering group in this phase. It is a proposal, not a delivered engagement. No core team was convened, no visit was run and no roadmap was aggregated in the copy I hold, and every count above is a planned count. The reading that the sponsor's missing cadence is the weak joint, the arithmetic on what a monthly steering group would and would not catch, and both of the changes proposed at the end, are mine and are not on the slide.