Stabilising a Post-Merger IT Group in 100 Days

Diagnosis: this is a decision-rights problem wearing a communication costume.

• Two CABs, two service desks, two monitoring stacks: every cross-estate change needs two approvals, so lead time roughly doubles. The 2.1 score on 'I know who decides' is measuring that structure, not morale.
• The steering committee has not met since March. Escalations have nowhere to go, so decisions stall or get made informally and re-litigated. The two ERP slips are more likely decision latency and lost legacy knowledge than engineering effort.
• The CIO from the larger firm communicating by email and never visiting the 900 km site confirms the smaller firm's staff in believing this is an acquisition, not a merger. That belief is the mechanism behind 18% attrition against a 9% market rate: people leave when the future is illegible, and half of the attrition is avoidable.

Days 1–30:

1. One decision forum. A single interim change authority under one chair, with a six-item decision-rights table published immediately and a fixed cadence.
2. Reinstate the steering committee fortnightly for 60 minutes, one-page status format, standing rule that no decision leaves without an owner and a date.
3. Site visit in week 2 - a full day, listening sessions without local management filtering, and a commitment list published within 48 hours with names against items.
4. Sixty-day scope freeze on non-critical work so the programme stops thrashing.
5. Integration principles on one page: one service desk, one CAB, one monitoring stack, with target dates. People tolerate bad news; they cannot tolerate ambiguity.
6. Start retention conversations against a critical-role list before the next resignation sets the frame.

Rule: a RACI that covers everything is used by no one. Define six decision types only. The two most contested:

Decision A - approval of a standard change affecting both estates

Role	Party
Accountable	Single combined change manager
Responsible	Technical approver from the owning estate
Consulted	Affected service owners; security if the data classification changes
Informed	Both service desks; the retired second CAB

Decision B - emergency change during an ERP cutover

Role	Party
Accountable	Integration programme director
Responsible	ERP application owner and infrastructure lead jointly
Consulted	Business process owners for affected finance processes
Informed	Steering committee; internal audit (log only)

Why joint responsibility works in B but not A: during a cutover the two estates cannot act independently, so one accountable owner with two responsible parties matches the work. In steady-state standard change, joint responsibility would recreate the two-CAB deadlock under a new name.

Getting it accepted:

1. Draft it with twelve people in one room - both change managers, both service desk leads, the ERP owners, security - not by email.
2. Resolve conflicts live and escalate at most two items to the CIO as options, not open questions.
3. Publish one page with a named owner per decision type and a 30-day review date.
4. Enforce by routing: approvals arriving through the retired channel are returned. Structure changes behaviour faster than exhortation.
5. Review at day 30 on a single measure - median decision lead time - with a baseline captured today.

Do not spread CNY 3M across 140 people - that is about 21k each, which buys no loyalty and creates an entitlement.

Allocation logic (18-month horizon, deferred with clawback):

1. Find critical roles by single-point-of-failure analysis tied to the two ERP migrations: who alone holds the legacy data model, the interface contracts, the regulatory reporting logic? Expect 20–30 roles, not 60.
2. Segment. Eight to ten 'must not lose' people get 150–200k each over 18 months, deferred and paid in thirds with clawback on voluntary exit. The next fifteen get 80–100k. That totals roughly 2.4–2.6M, leaving about 0.4M as a spot pool for unexpected resignations.
3. Frame the award as a conversation, not a letter: what the person is accountable for in the combined organisation, which role they will hold, who they will report to. Money without role clarity loses to a competitor offering both.

Non-financial levers, which address the survey scores directly:

• Named roles in the target operating model published within 30 days. A score of 2.1 says people cannot see themselves in the future.
• The decision-rights fix from task 2. The largest driver of frustration is not pay; it is re-litigating the same decision for months.
• Visible project choice on the ERP migration for high performers - the work people would rather do.
• CIO presence: repeated site visits and a monthly all-hands with real answers. Monthly email from 900 km away is itself a retention risk.
• Audit the appointment pattern. If exit interviews show the three managers left because senior roles went to larger-firm candidates, paying to retain the next tier treats the symptom. Fix selection first, or the pool buys eighteen months of the same problem.