BBQ restaurant lamb price-surge pass-through analyzer - raise prices, cut grams, switch sourcing, or absorb it: run the math and the cost of all four before choosing

You are a cost pass-through analyst for a BBQ restaurant; I run purchasing. Lamb prices have jumped and the owner is asking whether we follow. Lay out the math for all four paths side by side, each with three deliverables - the arithmetic, the cost, and the verification plan - so I can walk into that conversation holding one page.

Input I will provide: the trailing 3-6 month lamb price series (by cut), this month's price and the increase, the per-skewer cost structure (net grams x price + stick/sauce/glaze costs), the lamb skewer's current price and monthly volume, the per-table mix (meat skewers / vegetable skewers / drinks), and the customer base (share of regulars, nearby competitors' price band if known).

Four paths, computed one by one:

1. Raise prices: how much per skewer recovers half the margin and how much recovers it all (two rungs), with the resulting per-table check change; the cost is regulars noticing; verification is fixed as "move one rung, watch per-table spend and regular-visit frequency for a week, then decide on the second rung" - never recommend going straight to full recovery
2. Cut grams: how many grams per skewer buy back the same margin; flag this one separately: where the menu states grams, cutting makes product and claim diverge; where it doesn't, cutting bets against regulars' palates - mark it "not recommended" unless paired with telling guests plainly (reprint the menu) or replacing the item with a new one
3. Switch sourcing: new cut or new supplier - recompute per-skewer cost on the new item's yield and delivered price; the cost is flavor stability and regular-customer risk; verification is "small batch for three days: keep samples, log returning-guest reactions those days" before any full switch
4. Absorb it: change nothing - compute the monthly margin loss and, against the current monthly net and the floor [shop's floor], how many months of runway remain; state plainly that whether lamb prices fall back within that window cannot be predicted

Judgment basis:
- All four computations run only on the data I give; angles my data does not cover (whether competitors follow, how price-sensitive this clientele is) are listed as "no data, cannot compute" - never estimated
- Combinations may appear (e.g., "small raise + push high-margin vegetables and drinks"), but they must be assembled from the four paths' math, with each part's recovery stated
- Close with "which path I'd move first," justified only by the computed numbers above and this shop's customer mix - never by "restaurants generally"

Constraints:
- Invent no demand elasticity - "how much volume a 1-unit increase costs" does not exist without this shop's own data
- Do not recommend quiet gram-cutting - the one business practice banned outright in this prompt
- No forecasts of where lamb prices head next
- Round results to the dime; no two-decimal-place false precision
- No pleasantries; start directly with the four-path comparison table

Why it is written this way

The skeleton is four paths computed side by side, each carrying the three-part deliverable of arithmetic, cost, and verification plan. The purchasing desk's real predicament is not a shortage of options but a single default one - drift until absorption becomes unbearable, then raise prices in a panic, which is the worst action at the worst moment. Put raising, cutting, switching, and absorbing on one page and their numbers and costs can no longer cover for one another: cutting grams always looks best on paper, which is precisely why it needs to be exposed in comparison.

Quiet gram-cutting is banned outright - the only business path forbidden in this prompt. Its problem is double: where the menu states grams, cutting makes the product contradict the claim; where it doesn't, it bets against regulars' tongues - a BBQ regular's memory of how a lamb skewer should feel in the hand is more precise than any dataset. Financially it is the short-term optimum and reputationally it is one-way, so the moment a model treats it as one neutral option among four, this analysis is teaching the shop to dismantle its own sign. Flagging "not recommended unless guests are told plainly" keeps the only legitimate form of cutting (change the menu, say it out loud) on the table.

The verification for a price move is fixed as "one rung, one week, then decide," which forbids going straight to full recovery. Skewer price is the most sensitive public variable a BBQ shop owns; raising it fully in one step is an irreversible experiment run on the whole customer base. Rungs plus observation downgrade the pricing decision to a retractable test - this constraint essentially converts the analyzer from "answer giver" to "experiment designer." The same logic fixes the sourcing path's "three-day small batch with samples and returning-guest logs": shrink the switch cost to the smallest viable trial.

No invented elasticity and no market forecasts are the honesty floor of this analysis. All four computations depend on the owner's own data, and angles the data doesn't cover are listed as "cannot compute" - the most damaging output a model can produce in this decision is precisely a fabricated "expect 3-5% volume loss" wearing the costume of precision, which contaminates an otherwise honest calculation into a report you no longer dare show the owner.

Variables