The Visible Win That Breaks the Blueprint
A retail group is running a programme to replace three legacy order-management systems with a single platform across four regions. The blueprint requires a common product and pricing data model to be agreed before any regional migration, and the data-model design workshops are scheduled to finish in week 14. In week 3 the sponsor returns from a board meeting and asks the programme manager to "land something visible" - a pilot migration in the smallest region by week 10, using that region's existing product codes plus a temporary mapping layer. The region's operations manager supports the idea because it would prove the platform works. The programme's headline benefit is a 12% reduction in order-processing cost, which depends on the common data model being adopted in all four regions.
1. What judgement should the programme manager make about the sponsor's request, and on what basis?
Read the worked answer
Do not answer yes or no. Reframe the request against the principles.
Two principles are pulling apart:
- Design and deliver a coherent capability - the blueprint capability is a common data model. A pilot built on local codes plus a mapping layer is a different capability; it optimises one region and creates a permanent exception.
- Focus on benefits and threats to them and add value - there is real value in an early proof point, and refusing everything looks like dogma.
The test to apply: an early deliverable is acceptable only if it is either (a) a strict subset of the blueprint capability, or (b) explicitly disposable - no production data, nothing downstream depending on it.
So the programme manager should offer alternatives rather than refuse:
- Migrate one non-critical product category end-to-end in the small region using the draft data model (subset, not exception).
- Run a technical pilot of the migration tooling and reconciliation in a sandbox (disposable).
- Parallel-run the platform against live orders without switching traffic.
If the sponsor still insists on the mapping layer, the programme manager quantifies the benefit impact (does 12% survive four regions with mapping layers and one region off-model?) and puts the decision in writing to the programme board, with the sponsor as decision owner. The programme manager's job is to make the trade-off visible, not to absorb it silently.
2. The sponsor wants something to show the board in two weeks' time. What should the programme manager actually put in front of the board?
Read the worked answer
A one-page decision paper, not a status update. It should contain:
- The ask: approval of one of three options (subset pilot, technical pilot, deferred pilot).
- What each option proves and does not prove - an explicit line saying that none of them confirms the 12% benefit case, only the data model does.
- Benefit impact and reversibility - cost of the mapping layer, cost of removing it later, and whether the region would need re-migrating.
- The dependency: data-model workshops finish week 14; anything before that is based on a draft.
- Recommendation and rationale in two sentences.
Why this shape: the sponsor's underlying need is credibility with the board, and credibility is damaged far more by a pilot that has to be undone in tranche 2 than by a well-argued 'here is what we can prove by week 10'. Framing it as a decision also puts accountability for the trade-off where it belongs - with the board - while the programme manager keeps ownership of the professional advice.