Six stores, two trainers and a board go/no-go
A national retailer is rolling out a workforce-management system to 220 stores in six weeks. The impact assessment consists of a one-page list: 'managers need training' and 'staff need log-ins'. A six-store pilot found schedule errors, managers reverting to spreadsheets, overtime up 9%, and no named super-users. Store managers have 30 minutes of protected time per week for change activity; the programme has two trainers for 220 stores. The HR director wants a go/no-go recommendation at Friday's board. The programme manager says 'the pilot was only six stores, so it's not representative'.
1. Critique the impact assessment and specify the additional analysis needed before a board decision.
Read the worked answer
The assessment is a task list, not an impact assessment. It names training and log-ins but misses the changes that determine whether stores can operate: scheduling process, manager decision rights, role changes, competence, data quality, workload, local workarounds and customer effect. It also does not differentiate between store sizes, shift patterns or manager capability.
Additional analysis needed:
- Process impact by store type: what stops, changes and starts.
- Role impact for managers, staff and area managers, including time and accountability.
- Competence profile and training hours versus the 30-minute protected time.
- Data migration and schedule-error root causes from the pilot.
- Capacity: two trainers for 220 stores is not a training plan; calculate sessions, travel and super-user cover.
- Readiness by store segment: awareness, capability, local sponsorship.
- Financial and operational impact: overtime, schedule accuracy, absence cover.
Judgement: The pilot is small, but it is the only operational evidence available. The correct response is to expand or deepen the analysis, not to dismiss the findings because the sample is small.
2. Give a readiness-based recommendation for Friday's board, with conditions and decision criteria.
Read the worked answer
Recommendation: No-go for all 220 stores in six weeks. Conditional go for a staged rollout after fixes, starting with a larger pilot or first wave of stores that meet readiness criteria.
Readiness evidence: Managers are reverting to spreadsheets, overtime is up 9%, schedule errors persist and there are no named super-users. Protected time of 30 minutes per week is far below what is needed to learn a scheduling system. Two trainers cannot prepare 220 stores in six weeks.
Conditions before scaling:
- Fix the top schedule-error causes and re-test in pilot stores.
- Name and train super-users, with cover arrangements.
- Increase protected learning time and provide manager-level training.
- Confirm data quality and cut-over support.
- Agree go/no-go criteria for each wave: schedule accuracy, spreadsheet use, overtime, help-desk volume, manager confidence.
Judgement: The board needs a decision that protects payroll and store operations, not a date. A staged go with explicit gates is more defensible than either blind rollout or indefinite delay.
3. How should the programme manager's 'not representative' objection be handled?
Read the worked answer
Acknowledge the point: six stores cannot statistically represent 220, and pilot results may be affected by local factors. But 'not representative' does not make the evidence irrelevant. It is the only operational evidence of the system in use, and the pattern-reversion to spreadsheets, overtime, no super-users-is a serious warning.
Response:
- Use the pilot to generate hypotheses, not final conclusions: which store types, shift patterns and manager capabilities are most affected?
- Expand the pilot or first wave to cover more store types before scaling.
- Set explicit readiness criteria and measure them in each wave.
- Separate technical defects from change-readiness issues; both need owners.
Judgement: The risk of over-generalising from six stores is real, but the risk of ignoring them because the sample is small is greater. The board should hear both: limited evidence, clear warning signs, and a staged test that produces better evidence.