invent.sale
Demand · Assortment

Build the assortment matrix around what each location actually sells

An assortment usually accumulates by history: whatever was brought in once stays. Different neighbourhoods end up with the same list, a dead tail occupies shelf space, and missing items are invisible entirely.

who this is for

commercial director, category manager

layer
Demand

What is wrong today

  • The dead tail holds both shelf and cash, yet shows up in no report as a problem.
  • Locations with different demand get an identical matrix.
  • Revenue lost to an item you never stocked is nowhere counted.

How it works

A category × location matrix

Every item gets a role: core, local, test, delist, or missed opportunity. The role is argued from sales data, not from opinion.

A builder that shows consequences

Remove an item and you see the capital and shelf space it frees. Add one and you see expected revenue, the confidence behind it, and the cannibalisation of neighbours.

Delisting as a process

Delisting is not a checkbox but a sequence: stop purchasing, sell down the remainder, return to the supplier where the contract allows.

the working screen

What it looks like in use

Every item is named by its role in the category — and by the number that justifies that role.

invent.sale · Demand · Assortmentwindow: last 90 days
Full categoryDelisting candidates18Gaps7
ItemLocations sellingShare of categorySubstituteDecision
Nourishing cream 75 ml12 of 149.4%nokeep
Nourishing cream 30 ml4 of 140.8%yes, 75 mldelist
Wet wipes, 60 pcs14 of 146.1%nokeep
Shampoo 400 ml, type 23 of 140.4%yes, type 1delist
18 candidates · 4.2% of shelf, 0.9% of salesAssemble the category decision
The “substitute” columnNothing is delisted blind. You see whether the shopper moves to the neighbouring item or leaves the store.
Sales are counted per location, not on averageAn item that carries one location out of fourteen stops looking like a network-wide failure.
The cost of the decision, at the bottomYou immediately see how much shelf frees up and how much sales it costs. The argument becomes arithmetic.

An anonymised example: locations are lettered and the dataset is shared across every screen on this site.

What it is built on

  • Dead stock and the cash frozen in it are computed per location today — that is half of the future matrix.
  • Sales history at item × location is already collected, slow movers included.

Who stays in control

A matrix version is approved by a person and stored with a date and an author. Supplier and shelf constraints are set by the network and never overridden by a model.

How it is measured

Against locations whose matrix did not change: revenue per item, share of dead tail, and capital released.

what to read next

Show us one critical process. We will show how it runs here.

We look at your cycle: how an order is assembled today, who decides, where time leaks and what the system takes over.

Start with one critical process

Connect replenishment, supplier ordering or another first workflow, then expand across one platform.

Assortment — invent.sale