ESL Acceptance Checklist: 30 Checkpoints from Pilot Store to Full Rollout
You pay for an ESL project in two halves: on delivery and on acceptance. Most disputes are not about wrong goods — they are about acceptance criteria nobody wrote into the contract: how battery life is proven, what refresh speed counts as passing, what offline rate triggers remediation. This guide splits acceptance into four phases — pre-contract, pilot, staged rollout, closeout — with a checklist you can bring straight into the meeting room.
1. Pre-contract: turn claims into testable terms
Datasheet numbers are measured under ideal conditions. The job before signing is not to distrust the numbers but to put them into verifiable clauses. The usual suspects: battery life (must state whether it is an estimate model at a given daily refresh count or a measured value — "≈5 years estimated at 10 refreshes/day" is comparable; bare "5 years" is not), refresh speed (separate per-label and whole-store batch figures), gateway capacity (labels per gateway, and how expansion works beyond it), and whether the label quantity includes spares.
Run your own baseline first with the ESL Savings Calculator (/esl-calculator.html) from your facing count and repricing frequency, then compare quote structures: some include gateways and three-year warranty, others unbundle them into subscriptions — over three years the total gap can reach about 20% (estimate, quote-dependent).
2. Pilot store: six hard measurements
- RF coverage: run refresh tests at 3 positions each in the farthest shelf ends, behind stockroom walls and next to chillers; log failures — they determine gateway count and placement. Never test only within line of sight of the gateway.
- Batch refresh time: trigger a full-store batch repricing and time until the last label updates; write this measured figure into the acceptance baseline for later batches.
- Dual-price and promo layouts: verify member price, strikethrough original price, promo countdowns item by item after push; check long product names and CJK/multilingual characters for truncation.
- Offline behaviour: disconnect the gateway from the cloud, confirm labels hold the last price and auto-resync on reconnect — this is the most common real-world failure mode.
- Cold and sunlight: if you have cold-chain positions or sunlit shelving, test low-temperature models and UV ageing in the real environment during the pilot; do not accept lab reports as substitutes.
- Backend permissions: store staff / area supervisor / head office roles — what each can change, whether approval flows work, whether the audit log is exportable.
3. Staged rollout: go-live criteria per batch
Staged deployment contains problems in a small area — provided each batch has explicit go-live criteria. Four checks per batch: installation accuracy (audit 5% of facings; any facing-label binding error counts as wrong), batch repricing completion (full trigger; unupdated labels resolved within 24 hours), online rate (tracked daily in week one; below roughly 98% triggers remediation — set the exact threshold contractually by store size and gateway density), and staff competency (each store independently performs a batch repricing and a label swap).
The first two weeks after go-live are when anomalies cluster: fresh battery voltage spread, metal shelving blocking specific bays, and POS field-mapping mismatches all surface at once. Put "two weeks of on-site support after go-live" into the contract — far cheaper than arguing over email afterwards.
4. Closeout: retention, warranty and documentation
Acceptance clauses only have teeth if money is attached. A workable structure: release payment for batch two after pilot acceptance; retain 10–15% of the total until 30–60 days after full go-live (contract-defined) with online rate and battery spot-checks passing. Battery life cannot wait for real results, so define a spot-check method: read the battery curve of a 3–5% sample against the estimate model; deviation beyond the threshold (e.g. 20%) triggers remediation.
Documentation handover is the most skipped item and the one that hurts operations later: gateway configuration and key ownership, facing-to-label binding export, ownership of the backend super-admin account, integration API documentation, and warranty scope (labels vs gateways separately, accidental damage vs wear). Any item missing — hold the retention.
FAQ
How do you accept ESL battery life at handover without waiting two years?
You don't wait. The contract defines the estimate model (daily refresh count, temperature conditions); at acceptance you sample 3–5% of labels, read battery level and refresh logs, and compare the discharge curve against the model. Deviation beyond the agreed threshold (e.g. 20%) triggers remediation or warranty terms. Write all three — sample ratio, read-out method, threshold — into the contract.
How many labels are enough for a pilot store?
Covering the worst positions matters more than the count. Typically 100–300 labels, but they must include the farthest RF corners, chillers, metal shelving and promo end-caps, plus one full promo cycle in production (with at least one ad-hoc price change). A pilot installed only on the showcase shelf by the checkout tells you nothing.
Can we withhold the final payment if acceptance fails? What should the contract say?
Yes — provided the contract defines "pass" as measurable numbers: thresholds and re-test arrangements for batch repricing completion, online rate, refresh latency and battery spot-check deviation, with remediation deadlines and graduated deductions. Retaining 10–15% until KPIs pass 30–60 days after go-live is the common structure; a clause saying only "pay on acceptance" without metrics protects nothing.
Want this in your stores?
A 20-minute live demo with a full plan and quote for your facings.