A single-store, zero-to-low-CapEx test sequence, structured so that every dollar of real capital spend is gated behind evidence from the phase before it — not committee confidence.
Four phases, each gated by the results of the one before it. No phase begins until the prior phase's KPIs are met.
The pilot only produces a usable signal if the store is representative, not cherry-picked. These are the screening criteria before Day 0.
| Criterion | Threshold | Why It Matters |
|---|---|---|
| Sales Performance Tier | Mid-tier store (not top or bottom decile of the fleet) | A top performer would mask the treatment effect; a bottom performer risks confounding results with store-specific decline. |
| Mall Traffic Class | Class A or B regional mall | Matches the traffic profile of the majority of the ~450-store fleet retained post-bankruptcy, for a generalizable result. |
| Store Format | Standard footprint (not a flagship or outlet format) | Flagship and outlet stores have different lighting infrastructure and traffic patterns that would distort comparisons. |
| Management Tenure | Store manager in-role 6+ months | A pilot run by a manager still ramping up introduces execution-quality noise into the result. |
| Existing Fixture Compatibility | Current track lighting rail compatible with LED swap | Preserves the "CapEx: Zero" constraint of Phase 1 — no electrical infrastructure work. |
One store. Existing fixture aim adjusted to targeted LED tracking over product displays. In-store audio loop swapped to the reference program. No new hardware, no contractor visit, no capital request — this phase tests direction, not infrastructure.
Dwell time and conversion targets must both be met. If only one is met, extend Phase 1 by 30 days before proceeding. If neither is met, halt and revert to standard lighting/audio at no further cost.
Shift a defined share of associate hours from passive folding/restocking tasks to active floor styling and customer engagement. No headcount change — this is a scheduling and task-priority reallocation within existing labor budget.
Units-per-transaction target must be met without an increase in associate turnover — a labor model that boosts sales by burning out staff is not a model worth scaling.
The first phase requiring real capital: one self-service return kiosk and single-swipe POS terminal, installed only if Phases 1 and 2 have both cleared their gates. This is the most expensive and least reversible step in the pilot, which is why it's last.
All three KPIs must clear before any fleet-wide capital request is drafted. This phase is explicitly not authorized until Phases 1–2 succeed — capital follows evidence, never the reverse.
A single consolidated review of all pilot data against every gate above, presented to the capital committee as a go/no-go/expand-scope decision — not a fait accompli. Three outcomes are legitimate: full-fleet rollout, expansion to a second pilot cohort (5–10 stores) for a longer validation window, or termination of the program.
A clean pass across Phases 1–3 supports a capital request scoped to Pillar I and Pillar III only, for stores matching the Phase 00 selection profile — not an automatic mandate for every initiative in this proposal. Pillar II (inventory strategy) has its own independent validation path via buying-cycle results, tracked separately from this pilot.
What happens if a phase doesn't work — stated up front, not improvised after the fact.
Revert fixture aim and audio program to prior state. Zero capital was spent, so there is nothing to unwind financially — only the operational reset.
Restore prior task scheduling. Any associate feedback suggesting the reallocation model itself (not the atmosphere shift) drove turnover gets flagged for a separate review before any retry.
Kiosk hardware is redeployed to a second candidate store for a fresh 90-day test rather than written off — a single-store failure doesn't indict the concept, but it does block scaling until a second data point exists.