The accounts you are about to lose, on one list, every morning.

Each account's own baseline

Every account gets a baseline built from its own order dates — how many days normally pass between its purchases. Orders placed within three days of each other cluster as one purchase, so a split shipment does not read as two. Below four clustered orders nothing is claimed at all; the account is simply marked new.

• Baseline learned per account, never an industry average
• Orders within three days cluster into one purchase
• Four clustered orders minimum before any baseline is claimed
• Account states: new, healthy, watch, slipping, gone quiet

One Opportunity List

The accounts that broke their pattern land on a single list, capped so it stays workable — fifteen by default, thirty at the absolute most. One row per account, even when several signals fire on it. Every row carries the sentence that explains the flag and a concrete action, not a suggestion to follow up.

• Four types: Recover, Retain, Reorder, Expand
• One row per account — extra signals raise priority, not row count
• Priority is named: high, medium, low
• Capped at fifteen by default, thirty hard maximum

Named levels, never a number

Drift is reported as watch, slipping, or gone quiet — three levels, each a literal claim about the gap since the last order. There is no composite figure to argue about and nothing is rated out of a hundred, so a rep can read a row and immediately know what it asserts.

• Watch — the rhythm is stretching, not yet overdue
• Slipping — genuinely past the account's own expected order date
• Gone quiet — well beyond anything this account has done before
• The same three levels in-app and in each person's daily digest

Your whole book in one view

Every account you sell to, with the state it is in right now and what it is worth to you. Open any one of them for the baseline it was measured against, the gap that broke it, the orders behind both, and the date this flag would have fired on your own history.

• Every account carries a state: new, healthy, watch, slipping, gone quiet
• Per-rep books, plus a coverage view for managers
• An evidence panel behind every flag, all of it deterministic
• Unlimited seats at every tier

Outcomes and coverage

A rep logs what happened after working an account and it leaves today's list. When that account orders again, the save is attributed to the call that came before it — so the work is measured against real returning revenue rather than asserted.

• Logging the outcome is the only thing anyone is asked to type
• Saves attributed when a logged call precedes a resumed order
• Coverage shows what is unworked, never a ranking of people
• Today's list broken out by type

Replay, before you talk to anyone

Export your order history, drop the file in, and the same rules run over your past: which accounts drifted, the date each one would have been flagged, and how many days of warning that would have bought you. It runs in your browser, the file is never uploaded, and it needs no account.

• The same rules as the live product, over your own history
• Flag date, stop date, and days of warning for every account
• Runs in your browser — the file never leaves your machine
• Free, no login, no limit

Deterministic detection

Levels, priority and reasons are arithmetic on your order history, run inside your account's database. Same data in, same list out.

Data isolation

Row-level security enforces complete tenant isolation across all data.

You own your data

Your order history is yours. We don't use it for any purpose beyond your account, and never to train a model.

Alpha Stage