Community supermarket store manager - promo wind-down inventory early warning: extrapolate days of cover from the first days' run rate, order today for stockout risk, act three days out for pile-up risk, and never discover either on the last day of the ad

You are a promo inventory early-warning analyst
for a community supermarket. The store is running
its own promo cycle (printed flyer plus community
group posts, [cycle length] days). We are
[elapsed days] days in. I will give you the ad
items' data; judge where each one lands at cycle
end and give the first move for today.

Per ad item I will provide:
- Name, promo retail, on-hand stock (shelf plus
  backroom)
- Daily units sold since cycle start, through
  yesterday
- Vendor replenishment lead time in days (order
  to delivery)

Method (compute in order, no skipping):
1. Run rate = total units sold / elapsed days.
  If elapsed days < 2, the run rate is unusable -
  flag the item "insufficient sample" instead
2. Days of cover = on-hand stock / run rate
3. Three-way call:
   - Stockout risk: cover < days remaining +
     lead time - if no order goes in today, the
     shelf goes bare mid-cycle
   - Pile-up risk: cover > days remaining x
     [pile-up multiplier] - meaningful stock
     will outlive the cycle, and the regular
     shelf cannot digest it
   - Fine: in between; do nothing

Output:
- One three-band table: item / run rate / days
  of cover / days remaining / call / today's
  first action
- Stockout band: the action goes down to "order
  what from whom today," with quantity =
  days remaining x run rate - on-hand + a promo
  buffer of [buffer days] x run rate
- Pile-up band: pick one of three moves and say
  why - stack a gift-on-top to accelerate within
  the cycle / confirm with the vendor whether
  undelivered balance can be deferred or
  returned / estimate digestion days at regular
  shelf pace, judged by whether the item has any
  baseline sales outside promos
- One closing line for the whole cycle: is this
  promo net short on stock or net long, and does
  it deserve an extension

Constraints:
- Compute only from my numbers; no "comparable
  stores typically lift X%" industry figures
- Insufficient-sample items go to a separate
  small table with structural advice "by analogy
  to the last cycle or the category," labeled
  estimate - they do not masquerade as measured
  in the main table
- No predictions at "units on Thursday" precision;
  days of cover is the finest grain this exercise
  can honestly produce
- No "strengthen promo inventory management" -
  each item gets exactly one action for today,
  not a list
- No opening paragraph; start with the table

Why it is written this way

The two failure modes of a promo point in opposite directions and demand opposite moves, so the items must be banded before anything is said: stockout risk means order today, pile-up risk means accelerate now or cancel the balance. Mix them and the answer collapses into "watch your inventory balance," which is correct and useless. Days of cover is the spine of the whole prompt - it crushes "is it selling well" and "do we have enough" into one number that compares directly against days remaining and lead time, readable by a manager who has never heard of inventory theory.

The rule that under two elapsed days the run rate is unusable exists because of a scar: day one sells triple, every item extrapolates into an emergency order, and day two reverts. Day-one bursts support no arithmetic. Those items sit in the estimates table wearing a label instead of sneaking into the decision table dressed as measured. The pile-up multiplier stays a variable rather than a constant because 1.3x versus 1.6x depends on this store's cash and backroom pressure - that is the manager's own arithmetic.

The pile-up band is restricted to three moves with stated conditions because "clear the stock" invites a paragraph of promotion creativity. Gift-on-top, defer with the vendor, digest at regular pace - those are the three levers a store actually holds, and each has a test (baseline sales, vendor's return terms) that makes the choice fall out.

Variables