The Room Where the Model Changes: Auditing Service Governance Forums

Service operating models are usually described in artefacts: a target operating model deck, a responsibility matrix, a process catalogue, a service handbook. Yet the model that actually operates is often negotiated elsewhere. It is shaped in the daily huddle, the weekly performance review, the incident postmortem and the monthly change board. These forums decide what gets attention, what gets escalated, what gets tolerated and what gets fixed. If their agendas, attendees and rituals diverge from the designed model, the meeting becomes the real operating model.

That is not necessarily a failure. Forums are where judgement is applied to circumstances. But when the divergence is unexamined, the organisation can find itself governed by habit rather than design. A service owner may hold accountability on paper while a weekly review chaired by finance quietly sets the priority order. A postmortem may produce thoughtful actions that nobody with budget authority ever sees. A daily huddle may reward speed of closure while the operating model asks for end-to-end resolution.

Why the review is part of the model

Governance forums allocate three scarce things: attention, decision rights and legitimacy. Attention determines which signals are seen. Decision rights determine who can commit resources or accept risk. Legitimacy determines which actions carry authority after the meeting ends. An operating model that does not account for these forums is incomplete.

This matters especially in service operations, where work crosses teams, suppliers and channels. The designed model may specify a single point of accountability, but the review forum may reward functional escalation instead. It may state that customer outcomes take precedence, but the standing agenda may spend the first twenty minutes on cost per contact. None of this requires bad intent. It requires only that the forum has evolved separately from the model.

A practical audit of service review forums

The aim is not to audit meeting discipline for its own sake. It is to compare what the operating model says should happen with what the forum rewards, escalates and ignores. A four-week sample is usually enough to expose the pattern.

Start by collecting the formal evidence: terms of reference, agendas, attendee lists, information packs, decision logs, action trackers and postmortem outputs. Then observe at least two sessions of each significant forum. Interviews should follow, particularly with people who attend but rarely speak. The most useful output is a decision ledger: date, forum, decision taken, decision owner, relevant model reference, whether the decision was documented, and whether the resulting action closed.

Five tests to apply

1. Decision rights test. Do the decisions taken in the room match the charter? Are attendees empowered to make them, or do they defer to absent authority? Do escalations bypass the designed route?

2. Agenda weighting test. Compare time spent with model priorities. If customer harm, risk and capacity are stated priorities but cost dominates the standing agenda, the forum is teaching a different model.

3. Attendance test. Note who is absent and what their absence signals. If the accountable service owner is regularly represented by a delegate, accountability may be fictional. If suppliers attend only when there is a problem, the forum may be managing incidents rather than the service relationship.

4. Information test. Review the pack. Do the metrics reward local optimisation over end-to-end outcomes? Does a green status depend on excluding known defects? Are frontline and customer signals present, or only operational statistics?

5. Ritual test. Listen to the language. A postmortem that is called blameless but opens with 'who missed the alert' is not blameless. A review that treats every exception as a one-off is not learning.

Common patterns of drift

Several patterns recur. Review drift occurs when a forum gradually acquires decisions that belong elsewhere. Escalation theatre occurs when the meeting appears to decide, but the real decision is made beforehand or afterwards. Metric myopia occurs when the pack measures what is easy to count rather than what the model needs to manage. Action graveyard occurs when postmortems and reviews generate actions without owners, dates or funding.

These patterns are not merely inefficient. They create a second operating model that competes with the first. Staff learn to satisfy the room rather than the design. Risk accumulates in the gap.

Fixes that make the model visible again

First, publish a short forum charter linked explicitly to the operating model. It should state purpose, decisions the forum is authorised to take, quorum, required inputs and expected outputs. Avoid generic terms of reference that could describe any meeting.

Second, separate the forums by decision type. Performance review, incident review, change prioritisation and risk acceptance are different conversations. Combining them usually means one dominates and the others become standing items.

Third, maintain a decision log and action register owned by the model owner, not the meeting secretary. Review it monthly for closure and for drift. A decision that cannot be traced to a model role or accountability is a candidate for redesign.

Fourth, rebalance the agenda around the model's stated priorities. Start with customer harm and risk, then service performance, then cost and change. This is not a moral ordering; it is a way of preventing the easiest metric from setting the tone.

Fifth, invite people by decision rights rather than seniority. A forum that needs a supplier to commit to recovery should include someone who can commit the supplier, not only someone who can describe the contract.

Finally, make cadence an explicit choice. A daily huddle suits flow management. A weekly review suits performance correction. A monthly board suits risk acceptance and investment. A postmortem has a different purpose again. Write down why each forum exists and what would happen if it did not.

A short sequence for the audit

In week one, collect documents and map forums to model decisions. In week two, observe sessions and record time allocation, attendance and language. In week three, build the decision ledger and interview a sample of attendees and absentees. In week four, play the findings back to the forum itself. The output should be a one-page reconciliation: where the forum reinforces the model, where it contradicts it, and which three changes will be made.

The meeting after the metric is not a peripheral ritual. It is where the operating model is ratified, amended or eroded. Auditing it is a governance act, not an administrative one.