Assortment Planning

Assortment planning is the retail/fashion merchandising discipline of deciding which products — styles, categories, sizes, colours — to carry, in what variety and quantity, for each store, channel and region across a season. Sources position it as the bridge between financial targets and customer demand: it takes the budget set by Merchandise Financial Planning (MFP) and the in-period buying limit of Open-to-Buy (OTB) and converts them into a concrete product mix, which is then distributed to locations by allocation and replenishment. Toolio frames the stack as a hand-off chain: "MFP sets financial targets and inventory budgets, assortment planning selects the product mix, and allocation distributes those products across stores and online channels" (Toolio, vendor). The core tension it manages is breadth versus depth — how many distinct options to offer against how many units of each — under a fixed budget.

Sourcing note: Coverage this run is web-only — the Reddit (reddit-research) and YouTube transcript (Apify) source MCPs were unavailable (see gaps). The web sources are dominated by retail-planning software vendors (Toolio, o9, Board, Centric, daVinci, StyleMatrix, Leafio, Nextail, Blue Yonder, invent.ai); every benchmark and outcome figure carries a conflict of interest and is stamped with the source named. No Baymard/NNG/McKinsey/Deloitte-grade independent research on assortment planning specifically was retrieved this round.

Where it sits in the planning stack

MFP is the financial layer that budgets the assortment; the merchandise financial plan is "one of the main inputs to assortment planning — from the sales target per category and past performance, planners determine the breadth of the assortment" (Toolio, vendor).
Open-to-Buy (OTB) is the in-season control that compares actual sales and inventory against forecasts so planners can adjust buys; Centric describes OTB as "the in-season pulse of merchandise planning" sitting under assortment planning (Centric Software, vendor).
o9 frames MFP as operating "at the intersection of finance, merchandising and operations so that every buying decision ladders up to revenue and margin targets" (o9 Solutions, vendor) — i.e. assortment is where those laddered targets become product decisions.
Board characterises 2026 retail planning as shifting "from functional tools to cross-functional continuous orchestration across merchandising, supply and finance," claiming few platforms unify financial, merchandising and operational planning in one model (Board, vendor, 2026).

Breadth vs depth (the core mechanic)

Sources are consistent on the definitions:

Breadth (variety) = the range of categories/styles offered; depth = the number of units/variations (sizes, colours) per style. "Buying depth covers all required sizes and colours for each SKU" (StyleMatrix, vendor).
A broad assortment carries many product types but fewer choices in each (a department store); a deep assortment focuses on one category with many variations (e.g. a sneaker store with 50+ styles) (Leafio, vendor).
The failure modes are symmetric: "underbuying depth causes lost sales from stockouts in popular SKUs, while overbuying breadth ties up capital in slow-moving or niche options" (DotActiv, vendor).

Wedge planning / store-tiered breadth-depth

daVinci Retail describes "wedge planning": retailers cluster stores by sales volume, then "high-sales stores receive wider breadth and greater depth, while lower-sales stores receive a narrower assortment with fewer styles" (daVinci Retail, vendor). It adds that multiple demand dimensions refine the assortment simultaneously — "a style may perform best in high-sales markets, warm climates, and conservative customer bases at the same time" (daVinci Retail, vendor).

Localisation, clustering and omnichannel

Store clustering = segmenting stores by shared traits (sales patterns, size, climate, customer profile) and planning at cluster level; localisation then tailors the assortment per location/market (Toolio, vendor).
nul.global notes omnichannel raises the bar: "the same product must perform across physical stores, e-commerce and wholesale simultaneously" (nul.global, vendor).
Zara is the recurring localisation exemplar (third-party, non-vendor): it feeds real-time store sales to HQ and adjusts a product "performing in Seoul but not Rome" inside the current production window, with ~two store deliveries per week and ~24h shipping to European stores / 48h to North America (accelingo). It adapts assortment to climate — Scandinavia gets deeper wool coats/knitwear early, Southeast Asia breathable linen/cotton near year-round, and Southern Hemisphere markets run opposite seasons (gtelocalize).

Endless aisle / long-tail online assortment

Endless aisle lets in-store customers access the retailer's entire catalogue (kiosks/tablets) so an out-of-stock item, size or colour can still be ordered (NewStore, vendor).
spice Technology Group claims endless aisle can push SKU count "to 50,000 or 100,000 from the same ~300 vendors" by making the extended vendor community's items sellable online or via kiosk (spice TG, vendor) — illustrative, single-source.
Manhattan Associates flags endless-aisle ordering and "100% global inventory visibility across DCs, stores and suppliers" as 2025 omnichannel focal points enabling store-to-store orders (Manhattan, vendor, 2025) (as-of 2026).

AI / tooling landscape (flag vendor bias throughout)

A defined AI-assortment vendor category now exists — IDC published an "IDC MarketScape: Worldwide Retail AI-Driven Assortment Planning Solutions 2025 Vendor Assessment" (IDC, 2025, paywalled — only the title was accessible). Named tools and their (self-reported) claims:

Blue Yonder announced (2026-01-09) AI agents for MFP and Assortment Planning that "identify profit risks, recommend actions and build optimized assortments based on trend analysis," plus Micro Space Planning for shelf/display plans (Blue Yonder/BusinessWire PR, 2026 — promotional, unverified).
o9 Solutions positions itself as a unified platform spanning merchandising, demand forecasting, supply chain and revenue management with AI-driven assortment planning, price optimisation and promotions in one system; reported (Q2 2025) multiple fashion firms going live on assortment planning, demand/supply planning, inventory optimisation and OMS (o9, vendor, 2025).
RELEX connects demand forecasting directly to replenishment, production and space/assortment management to reduce planning silos (via Viewpoint Analysis, 2026 — promotional framing).
Nextail predicts micro-demand and auto-optimises inventory allocation across stores, warehouses and e-commerce; named in Gartner's 2024 Market Guide for Retail Assortment Management Applications (RAMAs) for short-life-cycle products (Nextail PR citing Gartner, 2024).
Toolio, Centric, daVinci, StyleMatrix, Leafio, invent.ai, First Insight, SAS — also named in the assortment/merchandise-planning space (web-source brief); concrete outcome data was not retrievable for most.

The "single unified platform" positioning overlap. o9, RELEX and Board each frame themselves as the rare vendor that unifies financial + merchandising + operational planning, while Board simultaneously states "few platforms unify" the three. This is competitive positioning, not a data contradiction — but it means no single vendor's "we are the unifier" claim can be taken at face value. (Board vs o9, both vendor, 2026)

Benchmarks / quantified results

Hard quantified outcomes were thin and almost entirely vendor- or academic-sourced:

[!unverified] 40% profit uplift. invent.ai cites a women's-apparel catalogue case where an optimised assortment method raised profits "by at least 40%" via reduced markdowns and fewer stockouts. The figure appears to originate from an academic paper, "Assortment planning in fashion retailing: methodology, application and analysis" (ResearchGate), of uncertain/likely pre-2022 vintage. Treat as a dated, single-source illustration, not a current benchmark. (invent.ai, vendor)

Merchandise planners "typically hold markdown budgets and may plan for ~30% of the assortment to be marked down at end-of-season," or set a margin allowance for markdown losses (Toolio, vendor) — connects assortment depth decisions to Markdown Optimisation and the markdown reserve in Merchandise Financial Planning (MFP).
A "test, read and react" approach — predicting Sell-Through Rate to dynamically adjust purchase orders — is cited as a way to minimise over-ordering, size stockouts and markdowns (daVinci Retail, vendor).
Toolio publishes assortment case studies (Knix, McGee & Co.) describing reduced excess inventory and improved sell-through, but the specific percentages sit behind case-study pages not retrieved this run (Toolio, vendor — metrics not extracted).

Key terms

Term	Meaning (as sources use it)
Breadth (variety)	Range of distinct categories/styles offered (StyleMatrix, Leafio)
Depth	Units/variations (sizes, colours) carried per style (StyleMatrix)
Wedge planning	Clustering stores by volume, then scaling breadth+depth up for high-volume stores, down for low (daVinci)
Store clustering	Grouping stores by shared traits to plan assortment at cluster level (Toolio)
Localisation	Tailoring the assortment per location/market/climate (Toolio, Zara examples)
Endless aisle	Letting any channel sell the full catalogue / extended-vendor SKUs beyond physical stock (NewStore, spice TG)
Option / SKU count	The number of distinct choices planned within a category (web-source — mechanics confirmed conceptually, no worked methodology retrieved)

What practitioners report

[!unverified] No Reddit or YouTube practitioner signal was collected this run — both source MCPs (reddit-research, Apify transcripts) were unavailable. The lived-experience angles remain unfilled: how assortment decisions actually get made (spreadsheets vs systems), breadth-vs-depth debates, store-clustering/localisation pain, planning-tool "worth-it" sentiment for small/mid retailers, and online long-tail / endless-aisle tradeoffs. See Reddit — Assortment Planning 2026-06-26 (gap-only) and YouTube — Assortment Planning 2026-06-26 (metadata-only). Five conference/vendor videos were identified for transcript retrieval on a future run — most on-topic: a PI Apparel Merchandise Planning USA 2025 panel on AI in merchandising.