What we detect, how we score it, and what we refuse to guess at.
Founders are allergic to noise — and rightly so. This page is the contract: precise on what your brief will contain, equally precise on what it won’t.
What we detect
Out-of-stock events on hero SKUs, by colourway and size.
Silent price drops — discounts pushed without a banner or email.
Size-curve collapse on a colourway (the M / L / XL backbone going to zero).
New SKUs that go live, including those held back from collections.
Collection re-orderings — which products they’re promoting to the top.
Agent-surface check — UCP capability drift, payment-handler changes.
What we don’t claim to detect
Sales velocity (we don’t have order data; we don’t guess from inventory deltas).
Revenue, AOV, conversion rates — anything that requires the merchant’s admin.
Marketing intel (ad spend, creative tests, attribution).
Returns, refunds, customer behaviour. We watch the catalog, not the customer.
How we score
Every detected event gets a 0–10 score combining visibility (was the SKU on the homepage), magnitude (how much of the stock or price moved), and surprise (was it telegraphed by a banner or email). The score determines where in the brief — if at all — it appears.
Hero-SKU OOS or size-curve collapse
A homepage-featured SKU or its size backbone going to zero. Headlines the brief.
Silent discount or hidden launch
Pricing or catalog change with no marketing push. Inline in the brief body.
Routine catalog drift
Standard restocks, expected price moves, collection re-sorts. Footer only.
Noise
Single-variant transient changes. Suppressed.
Anti-noise stance
A quiet day gets a quiet brief — three lines and a baseline. We don’t pad. We don’t send a daily “digest of everything that moved.” We don’t alert on a single-variant restock that could be a clerical typo. The signal-to-noise ratio is the product. If you cancel because we wasted your morning, the engine has failed — not your patience.