A decision tree for restoring access, availability, payments, connectivity and resident trust.
When a community micro market underperforms, record what the resident sees, when it started, which products are affected and whether the issue follows a device, a location or a time window. Protect the transaction first; diagnose second.
Capture screen, time, SKU and error.
Separate access, payment, stock and network.
Restore service and record owner/time.
| Lane | First check | Escalate when |
|---|---|---|
| Access | Badge, door state and permission window | Multiple residents are rejected |
| Payment | Reader status, network and settlement | A charge and vend disagree |
| Inventory | Physical count against the last sync | Variance repeats after recount |
| Connectivity | Signal, heartbeat and alert queue | Recovery fails after reboot |
| Experience | Cleanliness, signage and product condition | Complaints persist after correction |
- Review offline, low-stock, temperature and payment alerts.
- Walk the unit: lighting, door, screen, floor, labels and cleanliness.
- Reconcile top sellers and investigate unexplained variance.
- Rotate products and record expiry or damage.
- Close every incident with cause, action and owner.
Every visit: restore immediate availability and capture evidence. Weekly: compare stockouts, refunds, response time and resident feedback. Monthly: decide whether to change assortment, access hours, service windows or pricing communication.
Should operators reboot first? Only after preserving the error evidence and confirming no transaction is in progress.
What indicates a process problem? Repeated emergency visits, recurring refund disputes or the same SKU variance after recount.
How should residents be informed? Use a clear help path, expected response time and refund route; never leave a failed transaction unexplained.
