The ERP Rollout That Mapped the Org Chart, Not the Influence

A grid built from the register would look precise but would encode the org chart, not the change. Power/interest classifies by formal position and assumed interest; it does not test either. The register omits supervisors and planners who hold expert and informational power, and whose shadow spreadsheets are evidence that the nominal sponsors cannot enforce the process. It also treats 'plant director' as one stakeholder when the director's endorsement and the scheduler's behaviour point in opposite directions.

Before design freeze I would:

• Rebuild the long list from process touchpoints (procure-to-pay, scheduling, shift handover), not hierarchy, and add roles identified by snowball interviews ('who do you ask when this breaks?').
• Record for each stakeholder: interest, influence, attitude (support/neutral/oppose), impact of change, and evidence source.
• Run 10-15 short interviews across plants and functions to test the list and surface informal influencers.
• Map the works council and inspectorate separately as legal/consultative stakeholders with high legitimacy and delay power.
• Use a salience or influence/impact view alongside power/interest, but present the grid as a working hypothesis with a named owner and review date, not a final sign-off.

The output for the board should be a one-page table showing confidence and gaps, plus the decision it supports: who needs dialogue before design freeze, not who gets which newsletter.

Treat this as a compliance and trust risk, not a communications task. The works council is a high-legitimacy stakeholder with the power to delay or challenge implementation; the inspectorate has high power and low ongoing interest, so it needs accurate, timely information rather than a place on the community grid.

Actions:

1. Notify the works council in writing within days, propose a scoping meeting, and ask what information and timetable it needs. Do not present a decided design as consultation.
2. Take employment-law advice on consultation triggers; shift-handover changes and task redesign may require formal consultation before implementation.
3. Give the inspectorate a factual update: what changes, which roles, what consultation is planned, who owns it. Keep a single point of contact and a log of commitments.
4. Add both to the stakeholder register with attitude, influence, legal basis and engagement level (consult/negotiate), not just 'inform'.
5. Re-plan the design freeze: separate decisions that can proceed from those that touch consultation. This may mean freezing only the non-contentious scope.

Tell the board: engaging now may cost two to three weeks but reduces the risk of a stop notice, a formal dispute or a design that has to be reopened. Delaying consultation does not remove the obligation; it compresses it into a worse moment.

Completeness is a judgement backed by evidence, not a count. I would run five tests:

• Process test: walk each affected process end to end and list every role that touches it. If a role is missing from the register, the register is incomplete.
• Dependency test: ask 'who can stop, delay, fund, veto or quietly undermine this?' and 'who bears the cost?'
• Network test: snowball interviews and a short question in the pulse survey ('who do you ask when a process breaks?') to surface informal influencers.
• Boundary test: include external parties with legal or contractual rights - works council, inspectorate, key suppliers, customers - even if they seem remote.
• Feedback test: ask a sample of registered stakeholders who else should be included; review complaints, incidents and audit findings for names.

Then record confidence per entry and a review trigger. No register is ever final before design freeze; it should be versioned and reviewed at each gate. The sign-off should be on the process and the known gaps, not on a claim of completeness.