A tariff-driven price increase opened a 90-day window. Retailers loaded at the old price.
Ninety days later the deductions arrived — calculated on a basis nobody documented.
Illustrative fixture data — Blue Harbor Seafood Co. is a fictional company
MAR 3
Price increase announced: +9.5%, effective Jun 1. Tariff pass-through on imported tuna.
MAR 3 – MAY 31
90-day window open. Retailers may buy unlimited volume at $28.40 per case.
JUN 1
New price live at $31.10. Shipments collapse — demand was pulled forward.
JUL – AUG
Promotional deductions arrive against forward-bought inventory.
Forward-bought volume
—
cases shipped above consumption, inside the window
Deferred price capture
—
— per case × forward-bought volume
Claims over commitment
—
— of — claims exceed what was agreed
Commitment on record
—
provable as of the date each claim was made
Why the increase happened — and why it keeps happening
The window is not an accident of this one price increase. It is a recurring feature of the current
cost environment, and it reopens every time a manufacturer moves price.
Tariffs land on landed cost, not on the shelf. Manufacturers work through existing
inventory and supply contracts before repricing — a 12 to 18 month lag. That is why the repricing
wave is arriving through the middle of 2026 rather than when the duties were set.
The increases are concentrated and simultaneous. Because everyone's contracts roll at
roughly the same time, categories reprice together — produce, packaged goods, imported specialty
foods, and beverages within the same few months.
The numbers are real. Food prices ran +3.8% year over year in August 2026.
Fresh lettuce was +32.1% and tomatoes +19.5% against June 2025.
P&G alone estimated roughly $1B in pre-tax tariff cost for FY2026.
Frequency is the change, not magnitude alone. Price increases used to be annual. When
they become quarterly, the forward-buy window stops being an event to manage once a year and becomes a
standing condition — and each one writes another set of accruals nobody can produce.
Shipments spike inside the window, then collapse below consumption for a full quarter. Neither number
reflects consumer demand — both reflect the buy-in. Q2 looks excellent. Q3 craters.
Shipments to retail
POS consumption
Forward-buy window
What it is worth — put their revenue in
Move the assumptions and the numbers recalculate live. Nothing here is pre-baked: every figure below is
computed from the four inputs on the left.
Trade spend—
Estimated leakage—
Recovered, annually—
Enterprise value at exit—
Recovered trade dollars fall straight to EBITDA — the money was already spent and already belonged to
the business. No new customers, no new products, no incremental sales investment. Enterprise value is
the recovered figure carried at the exit multiple.
Every commitment below is on the ledger
The fingerprint in each row is the sealed record of what was agreed, timestamped before the
window opened. That is what makes a claim testable months later — and it is what a system that
merely stores the number cannot do.
Every claim matched against the commitment on record at the time it was made. Click any row for the evidence.
Retailer
Program
Committed
Claimed
Variance
Ledger
Status
Disputes opened$0
Naomi — grounded
Naomi — refused
Midwest Grocers' $47,900 claim cannot be assessed. No commitment document was ever written to the ledger
for program MWG-Q2-DISPLAY — the deal was agreed verbally on a call.
It is reported as unprovable: not clean, and not fraud. The system will not assert what it
cannot ground. That gap is itself the finding — it is the fourth-largest claim of the period, and there is
nothing on record to test it against.