Retail Frontline Performance: Store Context Card

1. Frame a store question that can change a decision

Replace a broad request such as explain why Store A is behind with a question about one work moment. Name the operational owner, the available choices and the date by which the comparison must inform a decision.

• Work moment: opening, customer handover, replenishment, returns, closing or another defined operation.
• Decision: what the owner may test, adjust, investigate further or leave unchanged.
• Comparison group: stores, shifts or cohorts with a stated reason to compare them.
• Period: exact start and end dates, including promotions, holidays, launches or disruptions.
• Out of scope: employment, safety, customer or process issues that follow another required route.

2. Complete one store context card per setting

Do not place two results beside each other until their units and operating conditions are visible. A percentage without its numerator, denominator and period is not a comparable store result.

• Measure: numerator, denominator, unit and source owner. For example, incomplete eligible handovers divided by all eligible handovers in the stated period.
• Demand and load: customer volume, cases, deliveries, queue pattern or other relevant work volume.
• People and coverage: staffing, scheduled hours, role mix, tenure, shift overlap and absences as permitted and relevant.
• Format and assortment: store format, channel, product range, campaign, return policy or customer mix that could affect the work moment.
• Process and tools: current instruction, device, system screen, access, escalation route and recent change.
• Employee context: permitted accounts, missing voices and what those accounts do not establish.

3. Separate observation from the explanation to test

Write the observed pattern in one sentence, then list at least two explanations that could fit it. A recurring employee account can make a work condition worth checking, but it does not establish an individual manager's capability or the cause of a store result.

• Observation: the bounded measure and context that differ, or the defined result that is the same.
• Explanation A: a process, access or workflow condition to check.
• Explanation B: a demand, staffing, assortment, format or measurement difference to check.
• Discriminating check: the smallest permitted record, observation or conversation that could weaken one explanation.
• Human review: the store, regional or process owner who can validate the local context.

4. Turn the comparison into one local test

Choose the smallest adjustment that fits the local work and does not bypass an approved process. Agree the baseline, eligible cases, support, side effects and stop rule before testing. A practice may later be prepared for transfer only after the local evidence and responsible review justify that separate decision.

• Test: one locally approved change, owner, participating shift or store and start/end dates.
• Implementation check: whether the change was available and used in the eligible work cases.
• Review measures: the same numerator and denominator definitions for the eligible test cases, plus workload, quality, access or customer side effects.
• Employee follow-up: a permitted route to describe what helped, remained unclear or was unavailable.
• Decision rule: continue, adapt, stop or gather specified missing evidence, with a responsible human owner.

5. Record what the comparison cannot show

Keep the context card with the review. State when the stores, periods or definitions were too different for a direct comparison, when a group was not represented, and when a metric moved without a confirmed explanation. This makes a later test more useful than a retrospective story about who performed better.

Why this guide is useful

Use this original context card when a store metric, operational incident or recurring employee account suggests a work moment worth checking. It comes before a local operational test. The store-practice card documents a practice ready to share, while the store-manager conversation pack prepares the human discussion about a pattern. This page establishes whether the initial comparison is meaningful enough to take either next step.

Compare the conditions before comparing the result

The UK Government's Magenta Book recommends defining the intervention, its context, assumptions and proportionate evidence before deciding whether a change made a difference. Apply that discipline to a store question: record the measure and its denominator, then test a practical explanation locally. A KPI identifies where to look; it does not determine what a manager, employee or store should do next.

Concrete example

Fictional example: a UK-US retailer compares handovers of open customer orders at closing in two stores during the same four-week promotion period. Store A records 24 orders without a named next owner among 120 orders requiring a closing handover, or 20%. Store B records 18 among 90 eligible orders, also 20%. The raw counts differ, but the stated rate does not. The context card still records that Store A had one extra closing colleague, Store B had a smaller accessory assortment and their escalation screens used different labels. In structured conversations, both teams say the next owner is sometimes unclear, while only Store B mentions that the screen label is hard to find. The network does not call either manager stronger or weaker. Store B tests one locally approved screen-label change for one eligible shift pattern, then checks completion using the same numerator and denominator definitions within that shift pattern, any added workload and employee accounts at the review. The result can support a local continue, revision or stop decision; it does not prove the label caused a network-wide performance difference.

Sources

• UK Government: The Magenta Book, evaluation guidance
• Lontra retail industry guide