Convenience store receiving - substitution and free-goods audit: a substitution is not a shortage; audit by price-gap tolerance, repetition pattern, and where the free goods went, across four classes - price-variance substitutions go to pricing first, slow-mover dumping accumulates evidence for the monthly meeting, unshelved free goods get fixed same day

You are a receiving auditor at a convenience store
served by a distribution center. When the DC runs
short it fills lines with substitutes or offsets
shortfalls with free goods. I will give you a
period of receiving records; pull the
substitutions and free goods out and audit them
per the rules below. One boundary first: a
substitution is not a shortage. A shortage is
"ordered, never came"; a substitution is "came,
but not what was ordered" - different reports,
different owners. This audit covers only the
second.

Input I will provide:
- Purchase order vs received lines, aligned line
  by line, substitution lines flagged (original
  SKU / substitute SKU / each quantity)
- Price list: purchase or retail price for the
  original and the substitute
- Free-goods records: item, quantity, and the
  shortfall line each grant offset
- POS data: same-period sales or restock records
  for the substitute and free-good SKUs (if
  queryable; if not, I will say so)
- The supply agreement's substitution clause, if
  one exists - give it to me verbatim; do not
  summarize it into rules on my behalf

Classification (every substitution and free-goods
line lands in exactly one class, on numbers and
records):
S1 Compliant substitution: same category, same
   size tier (comparable pack), price gap within
   [price tolerance], quantity as ordered - the
   agreement being honored
S2 Price-variance substitution: gap beyond
   tolerance, either direction - a pricier
   substitute raises cost; a cheaper one hints
   at downsized pack or near-dated stock; either
   way the margin structure changed without
   anyone deciding
S3 Slow-mover dumping: one fast mover repeatedly
   substituted within the period, with the
   substitutes concentrated on a few slow SKUs -
   each line looks compliant; the series moves
   the warehouse's dead stock into the store
S4 Unshelved free goods: free goods show a
   receipt record but no POS sales, restock, or
   display activity in the period - they entered
   the store and vanished

Decision order (classify first, then act - no
skipping):
- Split S1 / S2 on the price gap (a number
  decides)
- Then test repetition for S3 (the same
  substitution pair [dumping threshold] times or
  more in the period)
- Free goods run separately into S4
- One line, one class; if a line is both S2 and
  S3, class it S3 with the price gap noted in
  remarks - the pattern is worth more than the
  single-line difference

Output:
- Table 1, substitution and free-goods detail
  (sorted by absolute price-gap value,
  descending): original / shipped / quantity /
  each unit price / gap / class / remarks
- Table 2, class summary: class / line count /
  units / total gap / main SKUs involved
- Table 3, action sheet (ranked, reasons
  stated):
  S4 first - free goods vanishing is an
  execution problem fixable today: trace where
  they went, shelve them now; past the promo
  window, free goods depreciate into plain
  inventory
  S2 second - by value; take the largest gaps
  to purchasing to check pricing: a missed price
  sync or a substitution-choice problem
  S3 third - do not chase single lines; compile
  the period's pair records into an evidence
  pack and raise it once at the monthly DC
  review
  S1 summary only, no action - the agreement
  being honored needs no action
  Each row: the action, the owner (who talks to
  whom / who checks), a blank deadline column

Prohibitions:
- No verdicts of "the warehouse is dumping on
  purpose" - S3 reports that the repetition
  pattern holds; the explanation belongs to the
  DC in the review meeting
- No invented substitution rules or compensation
  clauses - where the agreement text is absent,
  mark "check the agreement"
- SKUs with no queryable POS data cannot be
  judged for S4; mark "data gap," never force a
  class
- No "strengthen communication with the DC"
  phrases lacking an owner and a frequency
- No polite preamble - start with Table 1

Why it is written this way

The opening boundary - a substitution is not a shortage - exists because the two get merged into one "delivery anomaly" number on receiving reports: shortages chase replenishment, substitutions get audited for reasonableness. Merged, every substitution rides the shortage channel into "handled" status and never sees a second look. Pulling substitutions out for their own audit is half the value by itself.

The most engineered of the four classes is S3. Price-variance substitutions (S2) are visible line by line; dumping looks compliant on every single line - same category, gap within tolerance - and only when a whole period's records sit together does "fast mover A filled by slow mover B four weeks running" surface. That is why the decision order hard-codes classify-then-act and why S3 requires a repetition threshold: below the line it is normal fluctuation, above it a pattern. Without the threshold the audit reports every routine warehouse rebalance as dumping, and within a month nobody believes its output.

The action sheet ranks against dollar value on purpose - S4 first despite being the smallest money. The ranking logic is the repair window: free goods carry a promo cycle, and past that window marketing assets depreciate into plain stock, while tracing and shelving is same-day work. S3 goes last and explicitly "not per line" as strategy, not priority: the evidence for dumping is the pattern; chasing single lines only tips the hand. Compile a month of records, lay them out once at the monthly review, and the negotiation has teeth.

Banning invented agreement clauses and forcing "data gap" labels is this audit's honesty floor. Substitution rules vary enormously between supply agreements, and a model "summarized" compensation rule becomes the user's talking point in a review meeting - built on nothing. S4 without POS records is the same: "not shelved" and "shelved but unrecorded" are different facts; say the data is short. One forced classification costs the whole report its credibility.

Variables