Every piece in the boutique that will sell it at full price
In luxury, one size of one style may sell a single unit per boutique per season, and a markdown costs more than margin. buffers.ai forecasts where demand is readable, buys each size at a service level you choose, places every piece where it's most likely to sell, and moves it to another boutique before it becomes a markdown.
Built forMerchandising, allocation and retail-operations teams at luxury and premium fashion houses running boutiques, e-commerce and wholesale accounts from shared stock.
| Size | Curve | Expected | Rounded | 70% service |
|---|---|---|---|---|
| 46 | 6% | 0.48 | 0 | 1 |
| 48 | 15% | 1.20 | 1 | 2 |
| 50 | 26% | 2.08 | 2 | 3 |
| 52 | 24% | 1.92 | 2 | 3 |
| 54 | 19% | 1.52 | 2 | 2 |
| 56 | 10% | 0.80 | 1 | 1 |
| Units bought | 8 | 12 | ||
Rounding buys 8 and skips size 46, though there's a 38% chance a customer asks for it. A 70% service level buys 12, with every size covered. Illustrative numbers.
Luxury breaks the averages mass retail plans with
Planning tools built for fast movers assume enough sales per SKU to find a pattern. Luxury has the opposite: few units, high value, and no free way out of a mistake.
Few units, high value
A style in one colour and size may sell once a season in a boutique. Each missing size is a lost full-price sale, and each extra piece is capital on a shelf.
Markdowns are a brand decision
Discounting isn't a free way to clear mistakes. Overbuying costs more than it does in mass retail, so the buy and the placement both have to be right.
Thin data at the level you buy
Style × colour × size × boutique is where the order is placed, and it's where the sales history is thinnest. Most of the calendar is zero.
Several channels, one stock
Boutiques, e-commerce and wholesale accounts all draw on the same limited units, and each piece can only be in one place.
Buy right, place right, then move what doesn't sell
Four decisions, each made with the data that exists at that level instead of an average that doesn't.
- Step 1
Forecast where demand is readable
One size of one colourway is too sparse to forecast, so buffers.ai forecasts the style family, where a champion model can be backtested, then splits it to colourway and size with a size curve pooled across the category and blended with the family's own sales.
Family Forecast & Size Curve - Step 2
Buy each size at a service level you choose
An expected demand of 0.48 units isn't an order. Each size gets the whole-unit quantity that covers its demand with the probability you set, instead of a rounded average that quietly drops the outer sizes.
How service-level buying works - Step 3
Place every piece in the right boutique
First Allocation sends a new style to each boutique in the sizes its own customers buy, within the maximums you set. Online and wholesale stock is reserved so boutique replenishment can't take it.
First Allocation - Step 4
Move it before you mark it down
Mid-season, Store-to-Store Rebalancing moves pieces that aren't selling where they are to boutiques that are short of the same style and size, so they get another chance at full price without a new purchase.
Store-to-Store Rebalancing
From the first shipment to the last piece
The same Replenishment capabilities the module page describes, picked for a network of boutiques with shared, limited stock.
First Allocation
Send new products to stores in the right sizes from day one
When a new product arrives at the warehouse, buffers.ai handles its first distribution to stores. You set rules with the maximum quantity per store and department. buffers.ai builds each department's size curve automatically from its sales history, with no hand-maintained size ratios. It then splits each store's allocation by that curve and decides how many units of each size to send.
Jersey Basic T · Ladies
356,005 units · 1,907 SKUs
Max per size 6 = 23 units
- XS 4
- S 6
- M 6
- L 4
- XL 3
- XXL 0
- Function
- First distribution of new products, split by size curves built from sales history
- Optimizes for
- A size mix that matches each department's customers from the first shipment
Assortment
Adjust your assortment dynamically
Adjusts your product assortment as sales data comes in, surfacing best sellers and matching the mix to what customers are actually buying. Also used to set the initial assortment when a new store opens.
Tops · 6 stores
Tail: under 2 units in 8 weeks
- Best sellers added 3
- Tail dropped 2
- Function
- Real-time, data-driven product-mix adjustment per store
- Optimizes for
- Best-seller identification, new-store product fit
Omnichannel/Online
Balance inventory across online sales, physical stores and wholesale
Splits inventory between online, physical stores and wholesale customers according to your policy, so no channel runs short while another sits overstocked. Wholesalers are supported as a channel alongside your own stores and online sales.
Denim Jacket · central warehouse
1,000 units on hand
- Stores 620
- Online 250
- Wholesale 130
- Function
- Inventory allocation between online, physical stores and wholesalers
- Optimizes for
- Cross-channel availability, policy compliance
Store-to-Store Rebalancing
Move stock from the store where it sits to the store where it sells
The units one store isn't selling are often exactly what another store is missing. buffers.ai finds stock sitting above target at one store while another store is short of the same product and size, and turns it into store-to-store transfers. Slow sellers get another chance to sell at full price instead of waiting for a markdown, and short stores are covered without a new purchase. It matters most for slow-moving, high-value products, where a single unit in the wrong store is a lost sale in one place and dead stock in another.
Wool Coat · size M · 4 stores
Warehouse sold out · no reorder this season
- Moved store-to-store 3
- New purchase 0
Store 01 sold none of its four in eight weeks; stores 02 and 03 sold out of the size.
- Function
- Store-to-store transfers from stores holding excess stock to stores that are short
- Optimizes for
- Full-price sell-through without new purchasing
The model suggests. Merchandising decides.
Luxury placement is partly commercial, so the rules that matter to a boutique override the statistics.
A display piece in every boutique
Display is a commercial decision. Set a display quantity and the boutique always holds it, whatever the model would suggest for a slow seller.
No broken size runs
Set a replenishment threshold and a shipment waits until enough of the size run is available, instead of sending a single size that can't sell on its own.
Stop chasing the tail
When a style falls below the sales threshold you set at a boutique, replenishment stops sending it there.
See the method before you sign
A planning decision on a piece that retails in the thousands should be explainable. buffers.ai publishes how it forecasts, how it scores, and what it costs.
Named models, published benchmark
Five named forecasting engines backtested per family, with the results of 21,600 backtests published by model and horizon.
Measured on whether the size was there
At one unit a season, average error rewards forecasting zero. The size-split holdout reports fill rate, the share of customers who found their size, alongside the error. Why that matters.
Prices on the website
Per-store and per-catalog rates with an instant calculator. Each module is licensed on its own.
Fits the systems you run
Connectors for SAP, Oracle, Microsoft Dynamics and custom ERPs; cloud or on-premise. Integrations.
Frequently asked questions
Test it on a season you've already sold
Send a past season's sales by style, colour, size and boutique. We'll show you your size curves, the size-level buy, and where stock would have moved between boutiques.