The Stage-Gate Pilot That Never Left the PMO
Northbridge Mutual, a 2,400-employee insurer, launched six 'agile' teams after an executive offsite. Eighteen months later the teams still produce quarterly business cases, obtain steering-committee approval before each release, and report percent-complete on a 40-page plan. The CIO says the company is 'doing agile' because teams hold daily stand-ups and use sticky notes. A recent incident illustrates the tension: a team discovered a simpler claims-routing design that could cut handling time by 20%, but waited seven weeks for the architecture board to approve a change request; by then the window had closed. Team members now describe retrospectives as 'therapy sessions with no authority.' Two high performers have resigned, citing bureaucracy. The newly appointed head of agile transformation asks you to advise her before the next board meeting. She has executive sponsorship but no budget to replace the PMO, and the CFO is skeptical of anything that weakens financial control.
1. What evidence in the case shows that Northbridge has adopted agile ceremonies without an agile mindset? How would you make that distinction concrete for the CFO, who equates control with stage gates?
Read the worked answer
Evidence of practice without mindset: stand-ups and sticky notes coexist with quarterly business cases, steering-committee approval for each release, percent-complete reporting, and a seven-week architecture-board wait. The team deferred a 20% handling-time improvement to an approval body, and retrospectives are described as having no authority-a sign that the team does not believe it can change its own system. Resignations of high performers suggest the environment rewards compliance over learning.
For the CFO, do not argue 'trust versus control.' Define the mindset as a set of decision behaviours: (1) decisions are made close to the work within explicit guardrails; (2) funding is released against outcomes and reviewed at cadence, not against activity completion; (3) experiments and fast feedback are expected, so uncertainty is managed rather than hidden; (4) leaders ask what was learned, not only what was delivered. Then show the cost of the current control system: seven weeks of decision latency, a lost 20% improvement, and replacement hiring. Propose a bounded experiment: one value stream gets a delegated decision budget and outcome-based reviews, with financial guardrails and a 90-day review. That gives the CFO control by design rather than by gate.
2. Recommend three interventions the head of transformation should propose to the board, and explain the sequence and expected resistance.
Read the worked answer
Change one decision right before changing anything else. Pick the claims-routing value stream and give its product owner authority to approve changes below a stated risk and cost threshold, with the architecture board publishing guardrails and a 48-hour exception path. This directly addresses the seven-week wait. Expect resistance from the architecture board and PMO; counter with a risk-tiered model and a 60-day decision-latency measure.
Replace percent-complete reporting with outcome and flow reviews. The board should see lead time, decision latency, and evidence of customer or operational value, not a 40-page plan. Expect CFO resistance; pilot the reporting for one quarter and keep statutory financial controls separate.
Coach leaders and protect retrospectives. Give teams a small improvement budget and require leaders to respond to one systemic impediment per retrospective. Expect middle-manager anxiety; define what managers are accountable for in the new model.
Sequence: decision rights first, because training without authority recreates cynicism; then reporting, then leadership coaching. Do not start with an agile maturity assessment.
3. The CFO asks: 'If we remove stage gates, how do we stop teams from wasting money?' Write the response you would give, including what you would not say.
Read the worked answer
I would not say 'trust the teams' or 'agile means no planning.' I would say: we are not removing financial control; we are moving it from pre-approval of activity to bounded delegation plus faster evidence. A stage gate is one control design, and here it is producing seven-week decision latency and an estimated lost 20% improvement. We can keep the controls that matter-annual funding envelopes, separation of duties, risk thresholds, audit trails-and change the cadence.
Concretely: set a delegated spend threshold per value stream; require a one-page outcome case with leading indicators and a kill-or-continue review every 60 to 90 days; use a portfolio kanban to limit work in progress; and publish decision latency and value evidence. The CFO retains the right to stop funding at review points. I would offer a 90-day controlled pilot with a pre-agreed measure: if decision latency does not fall or risk incidents rise, we revert. That makes the change falsifiable rather than ideological.