All use cases
Allocation2026-06-26

Allocation that finally knows the network

A 14-store apparel network still split each season with a rule written when it had 8 stores. Solya re-allocated on what every store had actually become.

Outcome

+9 pts full-price sell-through

Network

14 stores · lifestyle apparel

Measured outcomes

+9 pts

Full-price sell-through in tier-A stores

−15%

In-season inter-store transfers needed

Day 1

Live allocation the day stock landed

What's wired up

Systems connected

MERCH

Merchandising system

3 yrs sell-through by store

ECOM

E-commerce platform

Online demand

WMS

Warehouse (WMS)

Allocation & dispatch

Before · After

Before

Allocating yesterday's network

From a flagship to small regional stores, every store still got the same season split by a rule written when the chain had 8 stores. Tier-A stores ran out of bestsellers by week 4; tier-C stores still held week-1 stock at week 12.

After

A distribution that fits each store

Solya tagged every store on revenue tier, customer profile and category strength, then proposed a per-store distribution under MOQs, capacity and a 12% reserve. The team reviewed, adjusted a few lines, and shipped it live the day stock landed.

The challenge

It was FW 2024. The network ran from a flagship store down to small regional-town shops. But the same buy and the same allocation rules had been in place since the chain was 8 stores.

The new allocation lead, three months in, ran the numbers. Tier-A stores ran out of bestsellers by week 4. Tier-C stores were still sitting on week-1 stock at week 12. The pattern repeated every season: total revenue looked fine, but the missed opportunity was huge.

When the season hit the warehouse, the team split it with a rule written in 2017. Each store got a quota proportional to last year's revenue, plus a few manual nudges for known bestsellers.

That rule ignored everything that had changed since — store size, customer profile, online cannibalization, local events. They were allocating yesterday's network for tomorrow's season. Allocation had become a copy-paste job, not a decision.

What we changed

Through its data layer, Solya ingested three years of sell-through per store, category and week. It then tagged every store on three dynamic dimensions: revenue tier, customer profile (seasonal vs steady) and category strength.

Allocation logic became a function of those tags, not last year's revenue alone. The intelligence layer proposed a new per-store, per-tier distribution: more stock to high-velocity stores, less to slow regional ones. It ran under MOQs, store capacity and a 12% reserve held back for week-3 rebalancing.

How decisions get made

The team didn't rubber-stamp the model. They reviewed the proposed split, adjusted a handful of items where they knew something the data didn't, and approved the rest.

The 12% reserve gave them a live lever: stock held back at the warehouse, ready to flow to whichever stores showed the strongest early sell-through.

Where it lands

Approved allocations flowed through the orchestration layer straight into the WMS for dispatch, and the plan shipped live the day stock landed. Sell-through was monitored from week 1, and each season's outcome was logged to retune the model.

For the first time, the allocation matched the network it served. Tier-A stores stopped stocking out of bestsellers, tier-C stores stopped sitting on dead stock, and in-season transfers dropped.

What changed

  • +9 points of full-price sell-through in tier-A stores
  • 15% fewer in-season inter-store transfers needed
  • Live allocation on day 1, with dynamic store tags driving every line

Related: see how the seasonal buy plan feeds this allocation upstream, or how real-time network performance watches each store from week 1.

More use cases

From the blog