The Governance Board That Keeps Redesigning
A utility's architecture review board meets fortnightly. It has 11 members from IT, three from the business, and is chaired by the Head of Infrastructure. In the last quarter it reviewed 34 submissions, rejected 19, and average time-to-decision was 27 days. Business stakeholders have started routing around the board, getting sign-off from a divisional CFO instead. Two business architects sit on the board but have no vote. The board's stated remit is 'ensure architectural compliance', but its actual decisions are dominated by infrastructure standards and licensing costs. The CDO has asked you to 'fix the board' without increasing its headcount.
1. Diagnose why stakeholders are routing around the board, using governance design concepts rather than blaming individuals.
Read the worked answer
Routing around is a rational response to a board whose decision rights do not match its stated remit. Three structural faults:
- Mandate/authority mismatch. The remit says 'architectural compliance' (a business-architecture question), but the composition (8 of 11 IT, chaired by Infrastructure) and the observed decision content (infrastructure standards, licensing) mean the board is in practice a technology standards board. Business submissions get judged on criteria that are not theirs, so they lose - and then stop coming.
- Decision rights are undefined. With 11 members and no explicit delegation, every submission is a full-board deliberation. 27 days is the arithmetic consequence, not a discipline problem. Rejections at 56% suggest the board is acting as a gate rather than a governance body - it is doing design review, which belongs earlier and lighter.
- The architects have voice but no vote. They can explain but cannot adjudicate, so the body's decisions cannot be architecture-led by construction.
Routing to a divisional CFO is not a governance failure by the business - it is the business finding the nearest body that can decide. The fix is to give the board a mandate it is actually composed to discharge, and to create a fast path for the rest.
2. Propose a redesign that fits the no-headcount constraint, and state what you would measure to know it worked.
Read the worked answer
Redesign around tiered decision rights, not around more people:
- Split the remit in two. A Business Architecture Review (business-led chair, architects voting) for capability, value-stream and investment-alignment questions; a Technology Standards forum (Infrastructure-led) for platforms and licensing. Same people, two agendas, different chairs. This costs zero headcount and removes the mismatch.
- Define delegation thresholds. Below a stated cost/risk threshold, decisions delegate to the domain architect with post-hoc notification. Only above-threshold items reach the full board. This is what collapses 27 days.
- Give the business architects a vote on the business-architecture forum. Voice without vote is why their input is discounted.
- Publish the criteria the board uses, so submitters can self-assess before submitting - this converts rejections into pre-submission fixes.
Measures (baseline now, review at 90 days):
- Median time-to-decision (target: materially below 27 days).
- Percentage of submissions originating from the business - the key health signal; routing-around shows up here first.
- Rejection rate, with a target band not a minimum: near-zero rejections may mean the board has become a rubber stamp.
- Count of decisions taken via delegation vs. full board.
Do not measure 'number of reviews' - it rewards the gate behaviour you are trying to remove.