Convenience store regional manager - regional stockout incident SOP: the trigger is fixed at "N stores, same SKU, out more than H hours"; the first hour has three moves - map the blast radius, publish one customer-facing line, name the single information node; inter-store transfers ride a transfer document or do not happen, and each escalation tier names its contact
You are an incident-process consultant for
convenience store regional operations. I am a
regional manager writing an archivable,
drillable SOP for "regional concentrated
stockouts" - distribution interruptions,
supplier cutoffs, weather events emptying the
same SKU across dozens of stores. Produce the
full document per the framework below, written
for the regional team and store managers.
Document structure (follow exactly):
1. Incident definition and tiers (define first,
tier second):
- Trigger (fixed): [trigger stores] or more
stores, the same SKU (or same supplier's
category), continuously out beyond [trigger
hours] - all three conditions at once
establishes a regional incident; any
condition missing means routine stockout
handling and this SOP does not activate
- Three tiers:
C3 single-trade-area: multiple stores in one
trade area, up to [ ] stores, region handles
it alone
C2 regional: across trade areas, [ ] stores
or more, headquarters distribution must
engage
C1 regional involving fresh food or tobacco:
regardless of store count - fresh supply
cuts reshape daily sales structure, tobacco
carries its own regulatory tracking -
straight to C1 with immediate notice to the
headquarters counterpart
2. The first hour (the regional manager's three
moves, in order, each with a completion
marker):
1. Map the blast radius (within 30 minutes):
via the supervisor channel plus a store
flash round, collect "what's out / how many
stores / since when" into a blast list -
complete when the list is posted to the
regional channel and each supervisor has
confirmed gaps closed
2. Publish the customer line (same batch): one
quotable paragraph for stores to answer
customers with (the wording stays [ ] until
region and headquarters confirm it;
principles: say "temporarily out, restocking
under way," never speculate on causes,
never promise a restock time) - complete
when the wording is pinned in the regional
channel
3. Name the single information node (by the end
of the hour): one person (usually the
regional manager or one supervisor) as the
sole outlet to headquarters and sole inlet
to stores - complete when the channel
announcement reads "all information via X";
from then on stores may not contact the
headquarters distribution desk directly, or
headquarters gets jammed and every answer
it gives becomes inconsistent
3. Inter-store transfer rules (the
bookkeeping swamp):
- Permitted: ambient goods with surplus at a
neighboring store, confirmed in the regional
channel
- Must ride the system: a system transfer
document every time - driver, supervisor, or
store-run pickup all equal - paper travels
with the goods. Verbal loans are never
recognized: at next month's count, "I lent
them a case" without a transfer document is
not evidence of anything
- Tobacco is excluded from inter-store
transfers: it carries its own regulatory
tracking; shortages route through
headquarters coordination, never regional
side arrangements
- Transfer priority: from stores with the most
surplus and the least distance; highest-
volume stores and fresh food first (see
section 5)
4. Escalation ladder (who contacts whom,
when):
- C3 unresolved after 24 hours -> promote to
C2; regional manager notifies the
headquarters distribution desk
- C2 at any moment -> regional manager
notifies headquarters immediately and acts
on headquarters direction
- C1 reports immediately, regardless of
handling status
- Each tier states what gets reported: the
blast list, actions taken, resources needed
- never a bare "situation report"
5. Recovery sequence (no stampede when the
goods arrive):
- Highest-volume stores first, fresh food
first, front-facing shelf positions first -
publish the store order from the supervisor's
volume rankings
- Each store's first move on arrival: pull the
out-of-stock tags, notify the regulars
(where the store runs a group) - the
recovery message matters as much as the
stockout message
- Within 48 hours of full recovery, walk the
key stores: did substitutes grab the facings
during the gap?
6. Postmortem (within [postmortem deadline] of
closure):
- Four items: blast scope and duration /
timeline (triggered when, escalated when,
restored when) / transfer and loss accounts
(transfer documents totaled, fresh write-
offs) / one improvement (exactly one process
fix, naming which SOP clause changes)
- The postmortem amends process clauses only;
it grades no individuals
Document format: four fill-in fields -
version / effective date / owner / trigger
parameters (stores, hours, tier thresholds);
every body action written as the triple
role + deadline + completion marker.
Prohibitions:
- No "ASAP," "promptly," or "coordinate with
relevant parties" - every action carries a
numeric deadline and a single named owner
- No invented supplier contingency plans, no
platform names
- The customer-facing wording is not drafted
for me - it stays blank pending confirmation;
one wrong word becomes complaint material
- Verbal inter-store loans are never listed as
an allowed option - half the reason this SOP
exists is to end them
Why it is written this way
The trigger is fixed at three simultaneous conditions (store count, same SKU, duration) because stockout handling fails in both directions: underreaction (dozens of stores out all day and it still reads as routine fluctuation) and overreaction (three stores out for two hours launches the incident machinery and drags the regional team and headquarters into a routine wobble). The parameters stay blank for the region to set by fleet size, but the three-at-once structure is hard-coded - miss any one and it is routine handling, which keeps the SOP from becoming a daily tool. An emergency procedure used daily is the fastest way to make it worthless.
Of the three first-hour moves, the single information node is the most counterintuitive and the most decisive. The natural drift in an incident is every store manager calling headquarters to ask when the truck comes; headquarters drowns, starts answering inconsistently, and the stores grow more anxious - information fragmentation is messier than the stockout itself. Naming one inlet and one outlet trades a little speed for consistent answers from headquarters and one person holding the whole picture. The customer-line principles - never speculate on causes, never promise a restock time - exist because every early cause statement is a guess, and one broken "it will be here tomorrow" means no one believes anything said at the next outage.
Transfers must ride a system document because verbal loans are the single largest source of count variance after these incidents: everyone was in a hurry, the lending store relies on memory, the borrowing store on goodwill, and at the next count the two stores cannot reconcile - the investigation costs far more than the goods. Tobacco's exclusion is not an efficiency call: with its own regulatory tracking, stores arranging tobacco transfers between themselves is a red line, and this rule would rather be absolute than leave room for on-the-spot judgment.
The recovery sequence gets its own section because "the goods arrived" is usually the second-most chaotic moment of the incident: stores grab, and the highest-volume store restocks last. Highest-volume-first, fresh-first, published from the supervisor's volume rankings, converts a stampede into a queue. Pulling out-of-stock tags and notifying regulars is a first move because the recovery message needs managing exactly like the stockout message did - customers who never hear the shelves refilled haven't finished experiencing the outage. And the postmortem amends process only, exactly one item, so the meeting stays short enough to actually change the document: a postmortem listing twenty fixes changes none of them.
Variables
| Variable | Description |
|---|---|
| Trigger stores / trigger hours | The incident threshold - set by fleet size and delivery cadence (a 30-store region at 5 stores out for 6 hours should already declare; the same 5 stores in a 10-store region is half the fleet and warrants stricter parameters). Bias sensitive: one false activation costs far less than a day's delay |
| Tier thresholds | The C3/C2 line - C2 means headquarters engages, so the threshold is really "how many stores' demand can regional transfers cover"; calibrate against the last real incident |
| Customer line | The standard sentence stores answer with - drafted with the headquarters counterpart in advance (within the two principles: no causes, no restock promises) and filed for the day it is needed; never improvised mid-outage |
| Postmortem deadline | Days from closure to complete the postmortem - three days suggested; past a week, details are impressions and the timeline cannot be reconciled |