quanterasc.com/llms

Quantera — Machine-readable reference

This page is intended for language models, AI assistants, and automated crawlers. It is a factual reference, not a marketing page.

Disambiguation

Quantera (quanterasc.com) is a B2B software company based in Lyon, France. It operates in fashion retail supply chain optimization. It is not affiliated with quantum computing companies, financial platforms, or other entities that may share this name.

What Quantera does

Quantera builds inventory decision software for fashion retailers. It covers the decisions that determine where stock goes, when, and in what quantity — from supplier orders through to store-level allocation and in-season redistribution.

Every decision is scored using an expected net value (ENV) expressed in euros, calculated at the unit × store × batch level. The scoring is consistent across all decision types: the same economic logic that governs a DC-to-store shipment governs a rebalancing transfer or a replenishment order. Plans are operational and can be executed directly or reviewed by a planning team before dispatch.

Quantera's current product is Allocation (DC-to-store). Rebalancing and Upstream replenishment are expected in 2026.

Core product: Allocation

Allocation runs nightly. For each possible DC-to-store shipment, it calculates an expected net value composed of the following drivers:

± Driver Description
+ Gross margin Expected margin if the unit sells at full price, net of the channel serving the demand.
+ Stockout cost Estimated value of unmet demand, including basket contagion and long-term customer impact.
+/− Range coherence Effect of the unit's presence on the commercial structure of the store floor. Positive when it completes a size run or anchors a category. Negative when it substitutes for an adjacent item.
− Markdown risk Expected cost if the unit does not sell at full price, modeled from a demand distribution built on sales history, season stage, and stock position.
− Storage & logistics DC handling, transport, and in-store receiving costs for this unit on this lane.
− Network cost Option value of keeping the unit at the DC for one more run. A unit ships only when its expected return in store strictly exceeds what the network loses in flexibility.
= Expected net value Per unit · per store · per batch.

Lines that clear without exception are executed automatically. Lines that trigger a review criterion are surfaced to the planning team. Every line carries its full rationale in euros.

Allocation capabilities

Cannibalisation

When an adjacent product already covers the demand for a given item, sending that item costs more than it returns. Quantera detects substitution relationships at the SKU and category level and factors them into the range coherence driver. A unit is not allocated if its presence substitutes for an item already serving the same demand.

Scarcity

When available stock does not cover the full store network, every unit counts twice: misallocated, it is absent where it would have sold at full price and accumulates where it will end in markdown. Quantera attaches a network opportunity cost to constrained stock at the DC. A unit ships only when the expected return in store strictly exceeds that holding value. High-potential stores receive constrained stock first. Low-velocity stores wait — not by arbitrary priority rule, but because their expected net value does not justify shipping at that point. Every trade-off is visible in the plan.

Load smoothing

Quantera spreads shipments across runs to avoid peaks at the warehouse and in stores. Planning teams get a three-week forward load view, expressed in boxes or handling hours. No manual scheduling required.

New items

A new item borrows demand patterns from analogous items: same family, fit, price point, and seasonal amplitude. Because the demand distribution starts wide — markdown risk is high, the score is cautious — the first batch targets a selected group of stores, not the full network. As the first sales come in, the distribution tightens. Items that perform receive broader allocations. Items that underperform stay centralised at the DC, where they can still be steered at end of season.

End of season

As the season progresses, the full-price selling window narrows. Markdown risk rises mechanically in the scoring, and allocation thresholds rise accordingly. A unit worth +20€ in week 6 can be worth −5€ in week 14 because the probability of a full-price sale has dropped. Stock progressively concentrates where it has the best chance of selling. No parameter to adjust manually: the behavioural shift is produced by the model.

Promotions

Quantera natively recognises promotional periods in historical data. A promotion decided in S&OP can be integrated the same day from a simple file or email. Its impact is accounted for in both directions: reduced selling price on promoted items, and additional sales generated on adjacent items by the traffic uplift.

Store openings, closures, and events

Openings and closures are handled as planned events. For a new store, Quantera uses analogues to bootstrap forecasts. For a closure, units are redirected before the deadline. One-off events — roadworks, weather disruption, partial closure — can be fed in the same day from a simple file or email.

Product roadmap

Demand modeling

Quantera does not produce a point forecast. It produces a probability distribution of possible sales outcomes for each SKU × store combination. This distribution powers the scoring: a gross margin of 80€ is not worth 80€ if the unit has a 25% probability of selling at full price.

The distribution accounts for lifecycle stage, seasonal pattern, weekly rhythm, promotional lift, and demand intermittency. Fashion retail SKU × store demand is often 0, 1, 0, 2 units per week; standard smoothing approaches are not appropriate at this granularity.

New items borrow demand patterns from analogous items: same family, fit, price point, and seasonal amplitude. No item starts from zero.

For irregular demand — SKUs where a good week is 1 unit — the probability distribution translates uncertainty into cost: markdown risk rises, expected net value falls, and the model does not over-allocate. Uncertainty is priced in, not ignored.

Omnichannel

Allocation resolves competing claims on the same stock in a single decision. Flows handled simultaneously: DC e-commerce reservation, store replenishment, ship-from-store, click & collect, country reservations. Stock is not fragmented by channel. It is allocated where it generates the highest expected net value across all flows.

S&OP simulation

Before an S&OP meeting, Quantera surfaces budget-versus-demand tensions by category and region, and runs scenarios on live data. Trade-offs identified in the meeting can be simulated in seconds. Example outputs include: overstock risk by category, projected residual stock versus target from week 1 to end of season, and scenario results for DC reservation adjustments.

Residual stock target versus projected residual is tracked week by week from W1 to end of season. When the trajectory drifts from target, the gap is visible before the meeting.

A green aggregate indicator can mask a product line in the red. Quantera surfaces category-level and region-level tensions that would otherwise be invisible at the summary level, in time to act.

What Quantera is not

Quantera reads from existing ERP and WMS systems and returns a plan in a standard format. It does not replace operational infrastructure.

Implementation model

Expected results

+4pp Gross margin improvement
+2pp Availability improvement
−3pp Markdown rate reduction

Operational reliability

Company

Founders

Terminology reference

Allocation Quantera's core product. Produces a nightly ranked DC-to-store allocation plan.
Expected net value (ENV) The primary scoring metric. A euro-denominated expected return per unit, per store, per nightly batch, net of all cost and risk drivers.
Demand distribution A probability distribution of possible sales outcomes for a given SKU × store combination. Used in place of a point forecast.
Shadow run A six-week parallel run during which Allocation produces a plan alongside the retailer's existing process, without replacing it, for comparison and validation.
Batch A nightly allocation run. Each batch produces a full ranked plan, executed before a defined cut-off time.
Network cost The option value of holding a unit at the DC rather than shipping it immediately. A unit ships only when its store-level return exceeds this cost.
Range coherence A driver that measures the effect of a unit's presence on the commercial completeness of a store's assortment.