# OODARIS AI: the Agentic OS for Retail Decisions > The main content of oodaris.ai in one Markdown file, taken from the English pages. Every page also exists in French, Spanish, Turkish, German, Italian and Portuguese at https://oodaris.ai//... Index: https://oodaris.ai/llms.txt Last updated: 2026-10-29 ## Overview (https://oodaris.ai/en) Built in retail. Not built for. The Agentic Operating System for Retail Decisions. It watches the business overnight, explains what moved, proposes the fix, and writes the approved action back into your systems. Merchandise financial planning and open-to-buy run in production at ALDO Group. Five capabilities are available to new clients today, and five more are coming soon. OODIE is the agent your team talks to: ask it anything about OODARIS, run scenarios and simulations, or change a value, always with approval. ## The OODARIS loop (https://oodaris.ai/en/research) OODARIS builds on the OODA loop. Observe, Orient, Decide and Act think fast and act. Reflect, Iterate and Serve think slow: they compare what was predicted with what happened, adjust, and serve the new version back to your systems, which feeds the next Observe. The tag on each step names the product's job there. OODA: think fast, then act. RIS: think slow, then learn. - Observe: Watch sales, inventory and trends as they change. (In the product: Monitor) - Orient: Read the context and find the opportunity. (In the product: Understand) - Decide: Choose the best action for the goal. (In the product: Recommend) - Act: Carry out the decision. By default, its owner approves it first. (In the product: Coordinate · Execute) - Reflect: Compare what was decided and predicted with what actually happened, and simulate the paths not taken. (In the product: Simulate · Learn) - Iterate: Adjust the model and the plan where reflection found a gap. (In the product: Adjust) - Serve: Serve the new version to the systems and APIs that use it. (In the product: Publish) - Serve feeds the next Observe. ## Agentic OS: the ten capabilities (https://oodaris.ai/en/agentic-os) Every night, OODARIS checks plan, stock, sell-through and price, from the demand forecast to the markdown. By morning, the person who owns each decision has the cause and a proposed fix, with its expected margin. By default, nothing is written back to your systems until they approve it. Merchandise financial planning and open-to-buy run in production at ALDO Group. Five capabilities are available to new clients today, and five more are coming soon. ### The ten capabilities in OODARIS OODARIS proposes the decisions in each capability, and each decision has an owner. All ten run on one data model, so each one you add reuses the same data and integrations. Each stage has its own page, with every decision in it. - Plan (https://oodaris.ai/en/decisions/plan): Demand forecasting: Demand curves by style, store and week, with what moved them. MFP & OTB: Season targets for sales, margin, receipts and inventory, and the budget left to buy by department, channel and month. Assortment planning & optimization: Which styles go to which store clusters, and how deep. - Buy (https://oodaris.ai/en/decisions/buy): Purchase orders & procurement (coming soon): Order splits, alternate suppliers or new terms when a supplier slips. Sizing & prepacking (coming soon): Size curves and pack builds for each store cluster. - Flow (https://oodaris.ai/en/decisions/flow): Allocation planning & optimization: Units by store and size run. Replenishment (coming soon): The daily replenishment order, within pack, minimum-order and vendor rules. Transfers (coming soon): Store-to-store transfers before a size run breaks. - Sell (https://oodaris.ai/en/decisions/sell): Pricing, markdown & promotions: Price ladders inside MAP, MSRP and margin floors, markdown depth and timing, and offer depth and audience for each campaign. Order fulfillment (coming soon): Routing each online order to the warehouse or store best placed to ship it. ### Demand forecasting (https://oodaris.ai/en/decisions/plan#demand-forecasting) - When: Before the season, then every week - Today: The same forecast is rebuilt in several spreadsheets before the Monday meeting, and what moved it is hard to show. - With OODARIS: OODARIS updates the demand curve for each style, store and week, and shows what moved it: store and online sales, the marketing calendar, weather or a local event. The plan, buy and allocation agents work from the forecast once it is approved. - Approved by: Merchandise planning - Written back: The approved forecast, which the plan, buy and allocation agents use ### Merchandise financial planning (MFP) (https://oodaris.ai/en/decisions/plan#mfp) - In production at ALDO Group. Replaced a legacy planning system and 20+ spreadsheets. - When: Before the season, then every week - Today: The plan lives in spreadsheets and closes at month-end, so a margin miss is explained weeks after it happened. - With OODARIS: The plan re-projects as actuals land. A variance arrives with its cause by store cluster and the decision that would close it. - Approved by: Merchandise planning, with merch finance - Written back: The plan of record that open-to-buy is built from ### Open-to-buy (OTB) (https://oodaris.ai/en/decisions/plan#otb) - In production at ALDO Group. Feeds ALDO Group's own reports and dashboards. - When: Through the buy - Today: When a cost changes, OTB is rebuilt by hand, and buyers commit receipts against last week's numbers. - With OODARIS: A cost change, such as a new tariff on the autumn buy, rebuilds OTB and the margin impact by department before the next purchase order goes out. - Approved by: Merch finance, with the buying lead - Written back: OTB by department and channel, the budget buyers commit against ### Assortment (https://oodaris.ai/en/decisions/plan#assortment) - When: Range build - Today: Last season's winners are clear. Which options belong in which stores next season gets argued out in the line review. - With OODARIS: Each option is ranked by local demand, space, margin and exit risk for every store cluster before the range is committed. - Approved by: The assortment planner, with the buyer - Written back: The approved range by cluster, which allocation plans from ### Allocation (https://oodaris.ai/en/decisions/flow#allocation) - When: Launch and in-season - Today: A broken size curve shows up in the weekly report after the best doors have sold through their core sizes. - With OODARIS: Breaks are caught door by door, and each one arrives with a pack change, a transfer plan and its commercial impact. - Approved by: Allocation - Written back: Transfer orders and pack changes, written to your ERP ### Pricing and markdown (https://oodaris.ai/en/decisions/sell#markdown) - When: Full price through season exit - Today: Price ladders and markdown cadence follow rules written seasons ago, and the first cut is often deeper than the stock needs. - With OODARIS: OODARIS proposes price ladders that follow elasticity, inside your MAP, MSRP and margin floors. It proposes markdown depth and timing to hit sell-through. - Approved by: Merchandise planning, inside the margin limit merch finance set - Written back: Price conditions in SAP S/4HANA ## How it works: OODIE and the levels of autonomy (https://oodaris.ai/en/how-it-works) Ask OODIE. Approve what OODARIS proposes. By default, OODARIS proposes and a person approves each change before it is written back. You can also keep it read-only, or let it act on its own inside limits the decision's owner sets. OODIE is the agent your team talks to: ask it anything about OODARIS, run scenarios and simulations, or change a value, always with approval. How much OODARIS does is your call. Prescriptive is the default: OODARIS proposes, and a person approves before anything is written back. - OODIE, the agent your team talks to. What OODARIS does: Answers questions about anything in OODARIS, runs scenarios and simulations, and changes a value when you ask. What your team does: Asks why a number moved, tries a scenario, or asks OODIE to change a value. Who approves: Every change to a value needs approval. Limits, record and rollback: Nothing changes without approval. - Advisory, read-only signals. What OODARIS does: Shows what moved and why. Nothing is proposed or written. What your team does: Reads the signals and decides what to do about them. Who approves: Nothing to approve. Limits, record and rollback: Read only. Your systems don't change. - Prescriptive (Default), human approval required. What OODARIS does: Proposes the next best action, with its reason and expected impact. What your team does: Reviews the evidence, then approves, overrides or rejects the proposal. Who approves: The decision's owner, before anything is written back. Limits, record and rollback: Every approval and override goes on the record, and each writeback has a rollback path. - Autonomous (Available), acts within policy. What OODARIS does: Acts on its own, inside guardrails the decision's owner sets. What your team does: Sets the limits and the scope, then reviews what was done and why. Who approves: No approval per action inside the limits. The owner approves the limits. Limits, record and rollback: Every action is logged with its reason and can be reversed. ## Agents (https://oodaris.ai/en/agents) Twelve agents watch your plan, buys, ranges, stock and prices. Each one proposes a change with its evidence. By default, the person who owns the decision approves it before anything is written back. Each specialist works in one stage, plan, buy, flow or sell, and the AI Orchestrator connects them. Approvers follow your own approval model; each card names the usual approver. ### Plan (Demand forecasting, MFP & OTB, Assortment planning & optimization) - MFP Planner (https://oodaris.ai/en/agents#mfp-planner): watches actuals against plan for sales, margin, receipts and inventory turns. Proposes: A re-projected plan version, with each variance explained. Approved by: Merchandise planning, with merchandise finance. - Demand Intelligence Agent (https://oodaris.ai/en/agents#demand-intelligence): watches store and online sales, marketing calendars, weather and local events. Proposes: Updated demand curves by style, store and week, with what moved them. Approved by: Merchandise planning. - Market Insights Analyst (https://oodaris.ai/en/agents#market-insights-analyst): watches competitor catalogs, prices and availability, plus regional economic signals. Proposes: A weekly market read for each category, with competitor moves flagged. Approved by: Merchandise planning reviews it before it changes a plan. - OTB Controller (https://oodaris.ai/en/agents#otb-controller): watches open-to-buy by department, channel and month, against receipts, orders and sales. Proposes: Budget moves between departments and channels when demand shifts. Approved by: Merchandise finance, with the buying lead. - Assortment Curator (https://oodaris.ai/en/agents#assortment-curator): watches sell-through, attribute performance and space, by store cluster. Proposes: Adds, keeps and drops for each cluster, with the reason for each. Approved by: Assortment planning, with the buyer. - Trend Intelligence Scout (https://oodaris.ai/en/agents#trend-intelligence-scout): watches social, search and influencer signals, matched to your product attributes. Proposes: Early demand scores for new-to-market styles, by region. Approved by: Buying and assortment planning. ### Buy (Purchase orders & procurement, Sizing & prepacking) - Procurement Navigator (https://oodaris.ai/en/agents#procurement-navigator): watches supplier lead times, fill rates, costs and risk signals. Proposes: Order splits, alternate suppliers or new terms when a supplier slips. Approved by: Buying and sourcing. - New Product Catalyst (https://oodaris.ai/en/agents#new-product-catalyst): watches early sales of new styles against their launch plan. Proposes: Buy sizes for each launch wave, test groups, and when to chase or cut. Approved by: The buyer. ### Flow (Allocation planning & optimization, Replenishment, Transfers) - Inventory Optimizer (https://oodaris.ai/en/agents#inventory-optimizer): watches stock, sell-through and weeks of cover by store, warehouse and size. Proposes: Transfers and safety-stock changes before a size run breaks. Approved by: Allocation. - Replenishment Director (https://oodaris.ai/en/agents#replenishment-director): watches daily sales and stock on hand and on order, against reorder points. Proposes: The daily replenishment order, within pack, minimum-order and vendor rules. Approved by: Replenishment or allocation. ### Sell (Pricing, markdown & promotions, Order fulfillment) - Pricing Strategist (https://oodaris.ai/en/agents#pricing-strategist): watches sell-through, weeks of cover, price elasticity and competitor prices, by cluster. Proposes: Price ladders, and markdown depth and timing for each cluster, with the expected margin. Approved by: Merchandise planning. Merchandise finance clears the margin limit. - Promotions Director (https://oodaris.ai/en/agents#promotions-director): watches promotion calendars, uplift and cannibalization, by segment and channel. Proposes: Offer depth, timing and audience for each campaign, with the forecast uplift. Approved by: Marketing, with merchandising. ### Across all ten capabilities - AI Orchestrator (https://oodaris.ai/en/agents#ai-orchestrator): watches every proposal the specialist agents make. Proposes: The order of work, and a resolution when two agents disagree. Approved by: It approves nothing. By default, it holds each writeback until the decision’s owner approves. ## Architecture, integrations and security (https://oodaris.ai/en/platform) A decision layer over the systems you run. OODARIS reads from your ERP, POS and planning data. By default it writes back only what a named person approved. Your systems of record stay the source of truth. Every decision moves through the same five stages. An architecture review is about the contracts between them: what each stage receives, what it is allowed to do, where a person signs, and what goes on the record. ### Where OODARIS fits in your stack OODARIS reads from the systems you run today and writes changes back to the system that owns them. Nothing is migrated, and your systems of record stay authoritative. - Reads from: Merchandising and ERP: SAP ECC or S/4HANA, Microsoft D365, Oracle Retail, NetSuite, APTOS. Stores and commerce: POS, Shopify. Data platforms: Data warehouses such as Redshift, object storage such as S3. Everything else: Custom sources through APIs, event streams or file exports. - Ten capabilities: From demand forecasting and MFP & OTB to replenishment, pricing and order fulfillment, across plan, buy, flow and sell. - Named approval: By default, nothing is written back until the decision's owner approves. If the owner chooses Autonomous, OODARIS acts inside guardrails they set. Every action is logged with its reason and can be reversed. - Writes back: Plan updates: the re-projected open-to-buy (OTB) and merchandise financial plan (MFP); Price conditions and markdowns, for example in SAP S/4HANA; Transfer orders and pack changes. Each writeback uses the target system's own APIs or event paths and is logged with the decision that caused it. - To start a pilot, we need: POS or ERP access, through an API or a file export; Sales transaction history; Inventory data, ideally at SKU level. ### How data, access and approvals are handled These are the practices OODARIS is built and run with. The architecture review goes through each one against your own requirements. - Access control: Identity and role-based access decide who can view, propose, approve and execute each decision. Data is encrypted. - Tenancy and environments: Each customer's data stays inside its own tenant boundary, and environments are kept separate. - Data handling: OODARIS reads the data each decision needs and changes your systems only through writeback: approved by a person, or run inside limits the decision's owner set. Your systems of record stay the source of truth. - Approval trail: Every recommendation, approval and override is recorded with who made it, when and why. Policy checks and confidence thresholds run before anything reaches an approver. - Traceability: One trace ID links the source inputs, the agent and model runs, the approval, the writeback and the outcome. - Rollback: Each writeback has a rollback path, and a rollback goes on the same record as the decision. ### What technology leaders ask first The questions technology, AI and data leaders bring to an Agentic OS evaluation, and what to review together. - How does this fit the enterprise we already run? (Chief Information Officer) It sits above the systems you keep, and it can replace the planning tools you no longer need: at ALDO Group, OODARIS replaced a legacy planning system and more than 20 spreadsheets. It reads from your ERP, POS and planning data and, by default, writes back only approved actions, through those systems' own interfaces. A pilot reaches its first live cycle in 8–12 weeks. What to review together: Integration with existing enterprise systems; Security, access, and operating ownership; A phased rollout with clear decision gates. - How does a decision run in production? (Chief Technology Officer) Through five stages with a contract between each: data in, a bounded task, a recommendation, a policy check and approval, then writeback through APIs and events. Every step carries a trace ID, and every writeback has a rollback path. What to review together: APIs, events, and service boundaries; Failure handling and recovery paths; Observability across the decision lifecycle. - Where does the agent stop and a person decide? (Chief AI Officer) Where you set it. By default it stops at the approval: agents check, explain and propose, and a named person approves before anything is written back. At the Autonomous level, OODARIS acts on its own inside limits the decision's owner sets, and every action is logged with its reason and can be reversed. The policy limits and confidence thresholds your teams set are checked first, and every override is recorded with a reason. What to review together: Evaluation criteria for each agent task; Approval boundaries and human intervention; Monitoring, feedback, and model changes. - Can we trust the data behind each decision? (Chief Data Officer) Products, locations, calendars and measures are mapped to one model with lineage to the source record. Each recommendation lists the sources it used and its stated confidence. What to review together: Source ownership and shared definitions; Data quality, freshness, and lineage; Access policies and accountable stewardship. ## Customers: ALDO Group (https://oodaris.ai/en/customers) ALDO Group is a backer of OODARIS and a customer. Merchandise financial planning and open-to-buy run in production there, and five capabilities are available to new clients today. ### Capability by capability OODARIS runs all ten capabilities on one data model. Start with any one; each one you add reuses the same data and integrations. - Demand forecasting (Plan): Demand curves by style, store and week, with what moved them. - MFP & OTB (Plan) (In production at ALDO Group): OODARIS replaced a legacy planning system and more than 20 spreadsheets. Planners now do everything in one app, where OODARIS explains what moved and proposes the fix. Open-to-buy runs in OODARIS, built from the plan of record. Its output feeds ALDO Group's own reports and dashboards, as well as OODARIS. "Our merchandise financial planning used to live in spreadsheets and month-end. Now MFP and the AI run in the same loop. Targets re-project as actuals land, and the legacy planning system was retired in July 2026." (ALDO Group · Merchandise planning) - Assortment planning & optimization (Plan): Assortment ranks every option by local demand, space, margin and exit risk for each store cluster. - Purchase orders & procurement (Buy, coming soon): Order splits, alternate suppliers or new terms when a supplier slips. - Sizing & prepacking (Buy, coming soon): Size curves and pack builds for each store cluster. - Allocation planning & optimization (Flow): Allocation catches broken size curves door by door and proposes pack changes and transfers. - Replenishment (Flow, coming soon): The daily replenishment order, within pack, minimum-order and vendor rules. - Transfers (Flow, coming soon): Store-to-store transfers before a size run breaks. - Pricing, markdown & promotions (Sell): Pricing and markdown is built as an agentic system: price ladders inside MAP, MSRP and margin floors, and markdown depth and timing set to hit sell-through. - Order fulfillment (Sell, coming soon): Routing each online order to the warehouse or store best placed to ship it. ### Measured results - +4–8% gross-margin dollars, from the same inventory - −8–12pp discount depth, with units still cleared - +1–3pp margin rate, from more deliberate discounting - 8–12 weeks to a first live cycle, from kickoff to the first live decision cycle Ranges measured in pilots and in production at ALDO Group, against an agreed control. The ranges are not guarantees. ### What an engagement looks like Start with one decision, on your own data. 8–12 weeks to a first live cycle. - Foundations: Pick the decision, connect the source data and agree the KPIs. - Configure: Map your data, set the approval rules and tune OODARIS to your price ladders. - Live cycle: Planners run a real decision in OODARIS, with named approvers and controlled writeback. - Measure: Read the result against the control, then decide whether to scale. ## Research: Foundations of Agentic AI for Retail and Foundations of Agentic AI Systems (https://oodaris.ai/en/research) Foundations of Agentic AI for Retail: Designing, Building, and Operating Retail AI Agents, from First Principles to Production. By Fatih Nayebi (Founder and CEO of OODARIS; Assistant Professor, McGill University). Third edition · Gradient Divergence · October 2026. ISBN 9781069422644 (paperback) · 9781069422651 (hardcover). 430 pages. A practical guide to implementing AI agents in retail operations. OODARIS turns that research into software. By default, a person approves each decision before it is written back. The book introduces the BDI (Belief-Desire-Intention) framework and the OODARIS loop for retail AI: Observe, Orient, Decide, Act, Reflect, Iterate, Serve. It covers the theory and how to put it into practice. Editions (the first and second verified 2026-10-05 against ISBN registries and the Amazon listings; the third verified 2026-10-27 against the Amazon listings): - Third edition (current): Foundations of Agentic AI for Retail - Third Edition: Designing, Building, and Operating Retail AI Agents, from First Principles to Production. Gradient Divergence, October 2026. 430 pages. Paperback, ISBN 9781069422644, https://www.amazon.com/dp/1069422649. Hardcover, ISBN 9781069422651, https://www.amazon.com/dp/1069422657 - Second edition: Foundations of Agentic AI for Retail - Second Edition: Operating Models, Decision Frameworks, and Retail‑Grade Architectures. Gradient Divergence, January 2026. Paperback, 444 pages, ISBN 9781069422620. https://www.amazon.com/dp/1069422622 - First edition: Foundations of Agentic AI for Retail: Concepts, Technologies, and Architectures for Autonomous Retail Systems. May 2025. Paperback, ISBN 9781069422606. https://www.amazon.com/Foundations-Agentic-Retail-Technologies-Architectures/dp/1069422606 Foundations of Agentic AI Systems: One book in two volumes, about agents that act and what an agent must show before it is allowed to. Gradient Divergence, October 2026, hardcover (verified 2026-10-27 against the Amazon listings): - Volume I: Models, Learning, and Decisions, from First Principles. 455 pages, ISBN 9781069422668. https://www.amazon.com/dp/1069422665 - Volume II: Agents, Operations, and Governance, in Production. 419 pages, ISBN 9781069422682. https://www.amazon.com/dp/1069422681 ## Company (https://oodaris.ai/en/about) The Agentic OS for Retail Decisions, made in Montréal. Backed by and in production at ALDO Group. Founded by Fatih Nayebi, author of Foundations of Agentic AI for Retail. ### Where OODARIS comes from Built in retail: the people behind OODARIS were building AI inside ALDO Group before the company existed. - Before OODARIS: Years in AI, and one gap in retail. Fatih spent years building AI products and putting them into production. In retail he kept meeting the same gap: merchants, planners and allocators spend a lot of their time getting to a decision, across spreadsheets, handoffs and static rules, before they can make it. - ALDO Group: AI inside a retailer. As VP of Data & AI at ALDO Group, Fatih built AI for the retailer's own decisions. - McGill University: Teaching how agentic AI is built. He teaches at McGill, where students learn to design, build and run agentic AI systems. - 2025: Foundations of Agentic AI for Retail. The ideas behind OODARIS are written up in Fatih's book. A second edition followed in January 2026, and a third in October 2026. - OODARIS AI: A company to close the gap. Fatih founded OODARIS AI in Montréal to close that gap, with funding from ALDO Group, which is also a customer. - The team: A team from ALDO Group. He brought a team from ALDO Group, some of them his former students at McGill. - Today: In production at ALDO Group. Merchandise financial planning and open-to-buy run in production at ALDO Group, across its four brands. Five capabilities are available to new clients today, and five more are coming soon. Mission: Turn retail data into decisions, every day. OODARIS watches the business overnight, explains what moved and proposes the fix; by default, its owner approves before anything is written back. Vision: Every retail decision made with AI, from the season plan to the last markdown: less waste, more value, and the right product in front of every customer. Founder: Fatih Nayebi, Founder & CEO. Fatih founded OODARIS and leads the company. He wrote Foundations of Agentic AI for Retail. He is an Assistant Professor at McGill University, where he teaches Designing & Building Agentic AI Systems, Deep Learning, Data Science, and Machine Learning in Production. The team building OODARIS spans agentic AI systems, operations research, data science, product and engineering. ## News (https://oodaris.ai/en/news) ### OODARIS launches the Agentic OS for Retail Decisions (March 24, 2026, Las Vegas, NV) OODARIS today launched the Agentic OS for Retail Decisions, software for retail teams that still run planning across spreadsheets, handoffs and disconnected systems. The company is backed by and in production at ALDO Group. In most retailers, targets are set in one place, assortment is decided in another, and allocation and pricing follow downstream. OODARIS brings that work into one system. Each night it checks the business, explains what moved and proposes a fix, and a planner approves the fix before anything is written back. At launch, OODARIS covers demand forecasting, merchandise financial planning, assortment, allocation, and markdown and pricing. Teams use it to review, simulate, explain, approve and track decisions. "Retail teams do not need more disconnected AI. They need a better way to turn financial intent, merchant judgment, and live business context into action. That is what we built OODARIS to do." (Fatih Nayebi, Founder and CEO of OODARIS, author of Foundations of Agentic AI for Retail) OODARIS combines retail data, shared business memory, workflow orchestration, machine learning and optimization in one system, and it works alongside the systems retailers already run. The company is introducing OODARIS during Shoptalk week in Las Vegas, where the team is meeting retailers and partners. About OODARIS: OODARIS builds the Agentic OS for Retail Decisions for merchants, planners and allocators. The company is based in Montréal and was founded by Fatih Nayebi, author of Foundations of Agentic AI for Retail. Media contact: press@oodaris.ai ## Careers: jobs in Montréal and across Canada (https://oodaris.ai/en/careers) Build the software retail planners act on every morning. OODARIS is the Agentic OS for Retail Decisions. Overnight it finds what moved in the business, explains why, and proposes a fix that a planner approves in the morning. We're a Montréal team, backed by and in production at ALDO Group. Full-time roles in Montréal or remote anywhere in Canada, with core hours on Eastern Time. - AI Engineer (https://oodaris.ai/en/careers/ai-engineer): Build the agents that run every night in production at ALDO Group, and the evaluations and tracing that show whether their proposals are right. Agent Platform; Montréal, Canada (preferred), or Remote, Canada. - AI Product Manager (https://oodaris.ai/en/careers/ai-product-manager): Own one of our planning decisions end to end: work with ALDO Group planners, decide what we build next and measure whether it gets used. Agent Platform; Montréal, Canada (preferred), or Remote, Canada. - Founding Account Executive (https://oodaris.ai/en/careers/founding-account-executive): Be our first salesperson: build the pipeline with fashion, footwear and specialty retailers, and run deals to a signed paid pilot alongside the founder. Sales; Montréal, Canada (preferred), or Remote, Canada. - Full Stack Engineer (https://oodaris.ai/en/careers/full-stack-engineer): Build the screens where planners review, question and approve what our agents propose, and the typed APIs behind them. Product Engineering; Montréal, Canada (preferred), or Remote, Canada. Apply on each role's page, or write to careers@oodaris.ai. ## Agentic AI for retail planning (https://oodaris.ai/en/agentic-ai-for-retail) Agentic AI for retail is software that watches the business, works out what should change and acts on it, inside limits people set. This guide explains what that means for merchandise planning: which planning decisions agents can run, how an agentic planning system is built, how people stay in control, and how to start. ### What agentic AI means in retail (https://oodaris.ai/en/agentic-ai-for-retail#what) An AI agent is software that can decide and act, not only predict. It works in a loop: it senses what is happening by reading data from databases, APIs and live feeds; it thinks by running that data through models and rules; and it acts by triggering a business process, such as raising an alert, drafting an order or calling another system’s API. Agentic AI is the practice of building systems out of such agents, each with a goal, the data and tools it needs, and the constraints it must respect. What makes a system agentic is not one clever model. It is the orchestration of several steps and tools toward a goal, with little manual handling in between. A demand forecast on its own is analytics. A system that forecasts, compares the forecast with the plan, works out the order that closes the gap, checks the order against budget and supplier rules, and sends it to the ERP is agentic. Figure: An agent’s loop: it senses, thinks and acts, and records each step. An agent senses (reads sales, stock and plans), thinks (runs models and rules), acts (raises an alert, drafts an order or calls a system) and records what it did. Acting is what makes it agentic. That distinction matters for retail because most retailers already have plenty of prediction. They have forecasts, dashboards and weekly reports. What they lack is the step from a number on a screen to a decision made in time, written into the system that executes it, and recorded so someone can check it later. Agentic AI is aimed at that step. Retail uses agents in many places. Customer-facing agents answer questions, recommend products and help with fit. Operational agents forecast demand, manage stock across stores and warehouses, and set prices. This guide covers the second group as it applies to merchandise planning, sometimes called retail planning AI: the decisions that set what a retailer buys, where it goes and what it sells for. ### Why merchandise planning suits agentic AI (https://oodaris.ai/en/agentic-ai-for-retail#why-planning) Merchandise planning has the shape of a problem agents handle well. The decisions recur every week. They draw on a lot of data: sales, stock, receipts, orders, prices, calendars and plans, by product, store, channel and week. They have hard constraints: budgets, margin floors, pack sizes, vendor minimums and space. And they depend on each other, so a change in one decision should flow into the next. Today, most of that flow runs through people and spreadsheets. A planner re-forecasts the plan, a buyer reads the new open-to-buy, an allocator sees a broken size run in the weekly report, a pricing analyst picks the markdown. Each handoff takes days, and each team works from a different extract of the data. The result is decisions made late, on numbers that have already moved, with the reasons scattered across email and workbooks. Agents change the timing. They can check every product, store and week overnight, find what moved, explain why, and have a proposed fix ready when the planner starts the day. They can pass a change from one decision to the next without waiting for a meeting. And because every step runs in software, every step can be recorded. ### The planning decisions agents can run (https://oodaris.ai/en/agentic-ai-for-retail#decisions) Merchandise planning is a chain of decisions, from the season’s money down to the price of a single style in a single store cluster. Each has its own data, cadence and owner, and each can have its own specialist agent. - [Merchandise financial planning (MFP)](https://oodaris.ai/en/agents/merchandise-financial-planning): the season’s targets for sales, margin, receipts and inventory, re-projected as actuals land. An agent watches actuals against plan, explains each variance and proposes a re-projected plan. - [Open to buy (OTB)](https://oodaris.ai/en/agents/open-to-buy): the budget left to buy by department, channel and month. An agent rolls OTB forward from the latest demand and proposes moving budget when demand shifts or a cost changes. - [Assortment planning](https://oodaris.ai/en/agents/assortment-planning): which options each store cluster carries, and how deep. An agent ranks options by local demand, space, margin and exit risk, and proposes adds, keeps and drops. - [Allocation](https://oodaris.ai/en/agents/allocation): packs, size curves and transfers, store by store. An agent watches stock and weeks of cover by size and proposes transfers or pack changes before a size run breaks. - [Replenishment](https://oodaris.ai/en/agents/replenishment): reordering continuing products as they sell. An agent drafts the daily order within pack, minimum-order and vendor rules. - [Pricing and markdown](https://oodaris.ai/en/agents/markdown-optimization): price ladders inside price and margin limits, and how deep and how soon to mark down. An agent models elasticity by cluster and proposes the ladder with its expected margin. The value comes from connecting them. An approved plan feeds open to buy. Open to buy limits what the assortment can buy. The approved range is what allocation plans from. Allocation and replenishment move the stock, and pricing clears what is left. When each decision has an agent and the agents share one data model, a change at the top of the chain carries through to the bottom instead of waiting for each team’s next review. The [retail planning glossary](https://oodaris.ai/en/glossary) defines the terms used across these decisions. Figure: The planning chain: each decision hands its result to the next. Six decisions in order, each handing its result to the next: financial planning, open to buy, assortment, allocation, replenishment, and pricing and markdown. A change at the top carries down the chain. ### How an agentic planning system is built (https://oodaris.ai/en/agentic-ai-for-retail#how-built) Under the agents, an agentic planning system has the same few layers, whoever builds it. 1. A data foundation. Sales, inventory, orders, plans, product attributes and calendars from every source, cleaned and mapped to one model. Retail data sits in POS, e-commerce, ERP and planning systems, and joining it is the first and often the hardest job. Agents are only as good as the data they read. 2. Prediction. Time-series forecasting for demand, and classification, regression and clustering models for questions such as which stores behave alike or which products are at risk. 3. Prescription. Optimization that turns predictions into decisions under constraints: how many units to send to each store given capacity, which markdown clears the stock at the least margin cost, how to split a budget. Prediction says what will happen; optimization says what to do about it. 4. Business rules and guardrails. The limits a decision must respect: never price below cost, stay inside the margin floor, respect pack sizes and minimums. Domain experts set them, and the agents check against them before proposing anything. 5. Agents and orchestration. Specialist agents, each responsible for one decision, and a coordinating agent that routes work between them, resolves conflicts and keeps a record. Each agent has the context it needs, such as product data, store data and past cases. 6. Integration. Agents read from and write back to the systems that run the business through their APIs and event streams, so a decision becomes an order, a transfer or a price change in the system of record. Building it as separate modules with clear interfaces keeps it maintainable. A forecasting module can be retrained or replaced without touching the replenishment logic, as long as the interface between them stays the same. Figure: The layers of an agentic planning system, from the data up. Six layers, top to bottom: integration, agents and orchestration, rules and guardrails, prescription, prediction, and the data foundation that every other layer reads. ### What changes for planning teams (https://oodaris.ai/en/agentic-ai-for-retail#teams) Agentic AI does not remove the planner, the buyer or the allocator. It changes where their week goes. Today much of it goes to gathering numbers: exporting reports, rebuilding the same forecast in several workbooks, and reconciling versions before a meeting can start. With agents doing the checking overnight, the variance is already found and explained when the planner starts the day, and the time goes to deciding. The skills that matter shift with it. Planners spend less time on spreadsheet mechanics and more on judgment: whether a proposal makes commercial sense, which constraint should bend, when to override and why. Their knowledge also becomes part of the system, as the rules, limits and approval paths the agents follow, so it is written down instead of living in one person’s head. That only works if the teams who own the decisions help shape the system. Projects that bring merchandise planners, buyers, allocators and finance in from the start, alongside the data and engineering people, get rules that reflect how the business really runs, and users who trust the proposals because they helped define them. A feedback channel after launch, where users flag proposals that look wrong, keeps that trust as the system changes. ### Keeping people in control (https://oodaris.ai/en/agentic-ai-for-retail#control) The question every planning leader asks first is who decides. Merchandise decisions carry real money, and a wrong markdown or an over-buy is expensive. A well-built agentic system answers that question with five rules. - Explain every proposal. A recommendation should arrive with its evidence: what moved, why, the data it used and how confident the system is. Planners trust what they can check. - Keep a person in the loop, especially at first. The owner of each decision reviews and approves before anything changes, and every override is recorded with its reason. - Set limits and escalate exceptions. When a proposal would break a limit, or the agent meets a situation it was not built for, it should flag it to a person instead of acting. - Record everything and make it reversible. Every proposal, approval and writeback should be traceable to its source data, and every change should have a way back. - Widen autonomy gradually. As an agent proves reliable on a decision, the owner can let it act on its own inside limits they set, and review what it did afterwards. Autonomy, in other words, is a setting that changes over time, not a property of the software. Teams usually start with agents that propose and people who approve, and move routine, low-risk decisions to autonomous handling once the track record is there. Figure: Autonomy is a setting, widened one decision at a time. Three levels: it shows what moved, and nothing changes; it proposes a fix that a person approves first; it acts inside set limits, and a person reviews afterwards. ### How OODARIS applies agentic AI to retail planning (https://oodaris.ai/en/agentic-ai-for-retail#oodaris) OODARIS is the Agentic Operating System for Retail Decisions, for fashion and footwear retailers. It watches the business overnight, explains what moved, proposes the fix, and writes the approved action back into your systems. Every night it checks plan, stock, sell-through and price across ten capabilities in plan, buy, flow and sell, from demand forecasting to order fulfillment. By morning the person who owns each decision has the cause and a proposed fix with its expected margin. Twelve specialist agents watch the plan, buys, ranges, stock and prices, and each proposes a change with its evidence. The AI Orchestrator connects them: it routes each task to the right specialist, validates incoming data and records where every number came from, reviews each proposal before it reaches a person, and replans when a tool or an agent fails. OODIE is the agent your team talks to: ask it anything about OODARIS, run scenarios and simulations, or change a value, always with approval. The [agents page](https://oodaris.ai/en/agents) lists every agent with what it watches and proposes. Every decision moves through the same five stages: assemble context, scope and route work, generate a recommendation, review and approve, and write back and learn. Confidence thresholds, human overrides and approvals, and decision records travel with the decision. One trace ID links the source inputs, the agent and model runs, the approval, the writeback and the outcome, and each writeback has a rollback path. How much OODARIS does is your call, in four levels: - OODIE answers questions, runs scenarios and simulations, and changes a value when you ask. Every change to a value needs approval. - Advisory shows what moved and why. Nothing is proposed or written. - Prescriptive, the default, proposes the next best action with its reason and expected impact, and the decision’s owner approves before anything is written back. - Autonomous acts on its own inside guardrails the decision’s owner sets. Every action is logged with its reason and can be reversed. OODARIS reads from the systems a retailer already runs, including SAP ECC or S/4HANA, Microsoft D365, Oracle Retail, NetSuite, APTOS, POS, Shopify and data warehouses, and maps products, locations, calendars and measures from every source to one retail data model with lineage back to the source record. Each writeback uses the target system’s own APIs or event paths and is logged with the decision that caused it. [How it works](https://oodaris.ai/en/how-it-works) and the [architecture page](https://oodaris.ai/en/platform) have the detail. Merchandise financial planning and open-to-buy run in production today, and five capabilities are available to new clients. In production, OODARIS replaced a legacy planning system and more than 20 spreadsheets for MFP, and open-to-buy feeds the customer’s own reports and dashboards. The [customers page](https://oodaris.ai/en/customers) walks through each decision. ### How to start: one decision, one pilot (https://oodaris.ai/en/agentic-ai-for-retail#start) Agentic AI projects fail for familiar reasons: no clear objective, data that is not ready, a big-bang rollout that never reaches production, and users who were not involved and do not trust the output. The way around them is to start narrow and measure. 1. Pick one decision where margin leaks first, and say what success means in business terms, such as fewer broken size runs or shallower markdowns for the same sell-through. 2. Check the data. A pilot usually needs sales transaction history, access to POS or ERP data through an API or a file export, and inventory data, ideally at SKU level. 3. Agree the control. Decide in advance what the result will be compared against, such as a matched group of stores or categories, so the outcome can be read honestly. 4. Involve the people who own the decision. Planners, buyers and allocators know the constraints the agents must respect, and they are the ones who will approve the proposals. 5. Run a real cycle with named approvers, read the result against the control, and only then decide whether to scale or add the next decision. 6. Keep it current. Retrain models as seasons and behavior change, and review the guardrails as the agents prove themselves. An OODARIS engagement follows the same path: foundations (pick the decision, connect the source data and agree the KPIs), configure (map your data, set the approval rules and tune OODARIS to your price ladders), a live cycle (planners run a real decision in OODARIS, with named approvers and controlled writeback), and measure (read the result against the control, then decide whether to scale). You can start with one decision; each capability you add after it reuses the same data and integrations. Figure: An OODARIS engagement: four stages, from one decision to the choice to scale. Four stages: foundations, configure, a live cycle with named approvers, and measuring the result against the control before deciding to scale. ### Questions to ask before you buy (https://oodaris.ai/en/agentic-ai-for-retail#evaluate) When you compare agentic AI for retail planning, these questions separate a working system from a demo: - Which planning decisions does it run end to end, and which only produce a report? - What does it write back, to which systems, and through which interfaces? - Who approves each change, and can the approval rules follow our own approval model? - Can every recommendation be traced to its source data, and every change rolled back? - What does it need from us to start, and how soon is the first live decision? - How will results be measured, and against what control? - Does it replace the tools we no longer need, or sit beside them as one more system? ### The research behind it (https://oodaris.ai/en/agentic-ai-for-retail#research) OODARIS founder Fatih Nayebi wrote [Foundations of Agentic AI for Retail](https://oodaris.ai/en/research), a practical guide to implementing AI agents in retail operations. The book introduces the BDI (Belief-Desire-Intention) framework and the OODARIS loop for retail AI: Observe, Orient, Decide, Act, Reflect, Iterate, Serve. OODARIS turns that research into software. Figure: The OODARIS loop: the name is the method. Seven steps, one per letter of OODARIS: observe, orient, decide and act make the decision; reflect, iterate and serve check it against what happened and learn. ## Merchandise financial planning (MFP) and the MFP Planner agent (https://oodaris.ai/en/agents/merchandise-financial-planning) Merchandise financial planning sets a season’s targets for sales, margin and inventory, then keeps the business on them as actuals come in. This guide covers how planners run MFP today, the arithmetic that holds a plan together, and how the OODARIS MFP Planner agent re-projects the plan as actuals land. ### What merchandise financial planning is (https://oodaris.ai/en/agents/merchandise-financial-planning#definition) Merchandise financial planning (MFP) is how a retailer sets financial targets for its merchandise and then manages the business against them. The plan states how much each department or category should sell, at what gross margin, with how much inventory, and therefore how much product has to arrive and when. It is usually built for a season, broken down by month or week, and set at the upper levels of the product hierarchy (division, department, class), often split by channel and region. MFP is a financial plan, not a product plan. It works in money, usually at retail value with cost alongside, rather than in styles and units. The [assortment plan](https://oodaris.ai/en/agents/assortment-planning) decides which products fill that money, and allocation and replenishment decide where the units go. MFP sets the envelope those decisions work inside. A merchandise financial plan holds these measures for each period and each level of the hierarchy: - Sales: planned sales at retail, often with units and average unit retail. - Gross margin: in dollars and as a rate of sales. - Markdowns and other reductions: permanent and promotional markdowns, employee discounts and shrink. - Inventory: beginning-of-period (BOP) and end-of-period (EOP) stock, at retail and at cost. - Receipts: the inventory that has to arrive for the plan to work. - Turn and cover: inventory turnover, weeks of supply and often GMROI (gross margin return on inventory investment). Two directions of planning meet in MFP. Top-down planning starts from company targets set by finance and leadership and spreads them across divisions and departments. Bottom-up planning starts from the merchants and planners closest to the product, who build each category from its history, trends and buying intentions. Much of the planning cycle is the work of reconciling the two until the totals agree. ### How planners run MFP today (https://oodaris.ai/en/agents/merchandise-financial-planning#today) MFP follows a calendar. Before the season, often six to twelve months ahead, the plan is built. Last year’s actuals are the starting point, adjusted for growth targets, store openings and closings, channel shifts and known events. Leadership sets the top-down targets, merchants and planners build the bottom-up view, and the two are reconciled in a series of reviews until the plan is signed off as the plan of record. In season, the work becomes a weekly cycle. Planners load the week’s actuals, compare them with the plan and re-forecast the rest of the season. Where sales run ahead, they look for more receipts. Where sales lag, they cut or delay receipts and bring markdowns forward so that inventory does not pile up. The re-forecast goes to a weekly business review with merchandising, finance and buying. Much of this still runs in spreadsheets, sometimes beside a dedicated planning system. Spreadsheets are flexible, but they bring familiar problems: versions drift between teams, links between workbooks break, a change in one department takes hours to reach the totals, and nobody can easily say why a number changed or who changed it. The month-end close is often where variances surface, so a margin miss gets explained weeks after the decisions that caused it. ### The arithmetic that holds a plan together (https://oodaris.ai/en/agents/merchandise-financial-planning#arithmetic) The measures in a merchandise financial plan are tied together by a stock-flow identity: what you start with, plus what arrives, minus what leaves, is what you end with. Rearranged, it gives the receipts the plan needs. ``` Planned receipts = planned sales + planned markdowns + planned EOP inventory − planned BOP inventory ``` Figure: The month below as a stock flow, at retail: what you start with, plus what arrives, minus what leaves, is what you end with (illustrative numbers). A waterfall chart. BOP stock of $900,000, plus $400,000 of receipts, minus $400,000 of sales and $40,000 of markdowns, ends at EOP stock of $860,000. All the values are at retail. Many teams use total reductions (markdowns plus employee discounts plus shrink) in place of markdowns alone. Two more measures come from the same numbers: inventory turnover, which is sales divided by average inventory, and weeks of supply, which is stock divided by the planned weekly sales rate. The [retail planning glossary](https://oodaris.ai/en/glossary) has short definitions of each measure. A worked example: one month for a women’s footwear department, at retail (illustrative numbers): | Measure | Plan | | --- | --- | | Planned sales | $400,000 | | Planned gross margin | 52% of sales, or $208,000 | | Planned markdowns | $40,000 | | Planned BOP inventory | $900,000 | | Planned EOP inventory | $860,000 | | Planned receipts | $400,000 + $40,000 + $860,000 − $900,000 = $400,000 | | Average inventory | ($900,000 + $860,000) ÷ 2 = $880,000 | | Turn for the month | $400,000 ÷ $880,000 = 0.45 | A turn of 0.45 a month is about 5.5 a year if every month looked like this one. If the next month’s plan is $100,000 of sales a week, the planned ending stock of $860,000 is 8.6 weeks of supply. The point of the arithmetic is that every measure moves the others. Suppose sales come in at $360,000, 10% behind plan, while receipts and markdowns land as planned. Ending inventory is then $900,000 + $400,000 − $360,000 − $40,000 = $900,000, which is $40,000 above plan. Cover rises, and next month’s receipts must fall by $40,000 to get back on plan, or markdowns must rise. A weekly re-forecast exists to catch that chain while there is still time to change the receipts. ### The WSSI: the plan week by week (https://oodaris.ai/en/agents/merchandise-financial-planning#wssi) In the UK and in many fashion businesses, the in-season plan is kept in a WSSI, short for weekly sales, stock and intake. It has one row per week, with opening stock, sales, markdowns, intake (receipts) and closing stock, each shown as plan, actual and re-forecast. Closing stock is opening stock plus intake minus sales minus markdowns, and it becomes the next week’s opening stock. Most WSSIs also show forward cover: how many weeks the closing stock would last at the planned sales rate. The same month as a four-week WSSI, plan only, in thousands of dollars at retail (illustrative numbers): | Week | Opening stock | Intake | Sales | Markdowns | Closing stock | | --- | --- | --- | --- | --- | --- | | 1 | 900 | 90 | 100 | 10 | 880 | | 2 | 880 | 110 | 100 | 10 | 880 | | 3 | 880 | 100 | 95 | 10 | 875 | | 4 | 875 | 100 | 105 | 10 | 860 | | Month | 900 | 400 | 400 | 40 | 860 | Because the WSSI shows the plan one week at a time, it shows a problem while there is still time to act on it. A run of weeks where actual sales trail plan and closing stock climbs is the signal to pull intake back or move a markdown forward. Read at the department level, it is also the bridge between the season plan and open to buy, which is calculated from the same columns. ### MFP and OTB: how they connect (https://oodaris.ai/en/agents/merchandise-financial-planning#mfp-and-otb) MFP and open to buy (OTB) use the same numbers for different jobs. The merchandise financial plan says what the business intends to sell, earn and hold. Open to buy turns that plan into a budget for buying: the receipts the plan needs, minus what is already on order. MFP is the plan of record, and OTB is the spending control built from it. When the plan is re-forecast, OTB moves with it. That is why a stale plan does damage beyond the planning team. If the plan still shows last month’s sales trend, OTB still shows last month’s budget, and buyers commit receipts against it. The [open to buy page](https://oodaris.ai/en/agents/open-to-buy) works through the OTB formula with an example. ### How the OODARIS MFP Planner agent runs it (https://oodaris.ai/en/agents/merchandise-financial-planning#oodaris) In OODARIS, merchandise financial planning is part of MFP & OTB, one of the ten capabilities the Agentic OS covers across plan, buy, flow and sell. OODARIS uses [agentic AI for retail planning](https://oodaris.ai/en/agentic-ai-for-retail): specialist agents watch the business, explain what moved and propose a change with its evidence. The MFP Planner is the specialist for this decision. - Watches: Actuals against plan for sales, margin, receipts and inventory turns. - Proposes: A re-projected plan version, with each variance explained. - Approved by: Merchandise planning, with merchandise finance. - Writes back: The approved plan version, to your planning system. It builds driver-based plans from the demand forecast, runs scenarios for margin, receipts and inventory turns, reconciles top-down targets with bottom-up plans, and records every override and the reason given. It works with financial planning systems, merchandise planning tools, and your data warehouse and BI. The difference from the month-end cycle is timing. With OODARIS, the plan re-projects as actuals land. A variance arrives with its cause by store cluster and the decision that would close it. Once approved, the plan is the plan of record that open-to-buy is built from, and the re-projected open-to-buy and merchandise financial plan are written back as plan updates. Two other agents work in the same crew. The Demand Intelligence Agent proposes updated demand curves by style, store and week, with what moved them, and its approved forecast is used by the plan, buy and allocation agents. The Market Insights Analyst prepares a weekly market read for each category, with competitor moves flagged, which merchandise planning reviews before it changes a plan. By default, the person who owns the decision approves each change before anything is written back. Teams can also keep OODARIS read-only, or let it act on its own inside limits the decision’s owner sets; [how it works](https://oodaris.ai/en/how-it-works) describes the four levels. Merchandise financial planning and open-to-buy run in production at ALDO Group. OODARIS replaced a legacy planning system and more than 20 spreadsheets there, and planners now do everything in one app, where OODARIS explains what moved and proposes the fix. [ALDO Group’s story](https://oodaris.ai/en/customers) has the detail. > Our merchandise financial planning used to live in spreadsheets and month-end. Now MFP and the AI run in the same loop. Targets re-project as actuals land, and the legacy planning system was retired in July 2026. > > ALDO Group · Merchandise planning ### Questions Q: What is the difference between MFP and OTB? A: Merchandise financial planning (MFP) sets the targets for sales, margin, inventory and receipts. Open to buy (OTB) is the budget left to buy within that plan: planned receipts minus what is already on order. MFP decides how much the business should own; OTB controls how much buyers can still commit. Q: What does a merchandise financial plan include? A: Planned sales, gross margin, markdowns and other reductions, beginning and ending inventory, receipts, and measures such as turn, weeks of supply and GMROI. Each is set by period (month or week) and by level of the product hierarchy, usually at retail with cost alongside. Q: What is a WSSI in retail? A: A WSSI (weekly sales, stock and intake) is the week-by-week view of the merchandise plan. Each row shows opening stock, sales, markdowns, intake and closing stock, as plan, actual and re-forecast, so planners can see a gap between plan and actuals while there is still time to change intake or markdowns. Q: How often should a merchandise financial plan be re-forecast? A: The season plan is usually set once and revisited at set points, but the in-season re-forecast is typically weekly. A monthly re-forecast tends to find a problem after the receipts that caused it have already landed. Q: What is the difference between top-down and bottom-up planning? A: Top-down planning spreads company targets from finance and leadership down the hierarchy. Bottom-up planning builds each category from its history and the merchants’ intentions. MFP reconciles the two until the totals agree. Q: Does the MFP Planner change the plan on its own? A: Not by default. It proposes a re-projected plan version with each variance explained, and merchandise planning approves it, with merchandise finance, before the approved version is written back to your planning system. Every override is recorded with the reason given. ## Open to buy (OTB) and the OTB Controller agent (https://oodaris.ai/en/agents/open-to-buy) Open to buy is the budget a buyer has left for a period: what the plan says should arrive, minus what is already on order. This guide gives the open to buy formula with a worked example, shows how buying teams manage OTB today, and explains how the OODARIS OTB Controller agent keeps it current. ### What open to buy means (https://oodaris.ai/en/agents/open-to-buy#definition) Open to buy (OTB, also written open-to-buy) is the amount of merchandise a retailer can still buy for a period without going over its inventory plan. It is usually stated in money, at retail or at cost, and sometimes in units, for a department, class or channel and for each month or week. When OTB is positive, the buyer has budget left. When it is negative, the business is overbought for that period: more is coming than the plan can absorb. OTB comes from the [merchandise financial plan](https://oodaris.ai/en/agents/merchandise-financial-planning). The plan sets the sales, markdowns and inventory the business intends to have. OTB turns those targets into a purchasing limit and subtracts the commitments already made. It is the control that keeps a buying team ordering against this season’s plan rather than last season’s enthusiasm. Getting OTB right matters for three reasons: - Cash: inventory ties up working capital, so buying past the plan strains it. - Margin: excess stock usually leaves through markdowns. - Availability: buying too little leaves stores short of what customers want, and those sales do not come back. A well-run OTB keeps some budget open in season, so buyers can chase what sells instead of being fully committed months before customers have voted. ### The open to buy formula (https://oodaris.ai/en/agents/open-to-buy#formula) The standard retail formula, for one period at retail value: ``` OTB = planned EOP inventory + planned sales + planned markdowns − BOP inventory − on order ``` The first four terms give the planned purchases, or receipts, for the period: the stock you want to end with, plus everything that will leave through sales and markdowns, minus what you hold at the start. Subtracting on order, the merchandise already committed to arrive in the period, leaves what is still open. ``` Planned purchases = planned EOP inventory + planned sales + planned markdowns − BOP inventory OTB = planned purchases − on order ``` Some teams count all reductions, not only markdowns: markdowns plus employee discounts plus shrink. Others plan at cost. The structure is the same either way. What matters is that every term uses the same basis (all at retail or all at cost) and the same period, and that on order is counted by the month the goods arrive, not the month the order was placed. The [retail planning glossary](https://oodaris.ai/en/glossary) defines each term. For a future month, BOP inventory is the planned figure. Once the month has started, teams use actual stock on hand and the re-forecast sales, which is why OTB for the current and next periods changes every week. ### A worked open to buy example (https://oodaris.ai/en/agents/open-to-buy#example) One month for a men’s footwear department, at retail (illustrative numbers): | Term | Value | | --- | --- | | Planned EOP inventory | $520,000 | | Planned sales | $180,000 | | Planned markdowns | $20,000 | | BOP inventory | $480,000 | | On order for the month | $150,000 | Planned purchases are $520,000 + $180,000 + $20,000 − $480,000 = $240,000. Subtract the $150,000 already on order and OTB is $90,000 at retail. Buyers commit purchase orders at cost, so the retail figure is converted with the cost complement. If the department buys at an initial markup of 55% of retail, each retail dollar costs 45 cents, and OTB at cost is $90,000 × 0.45 = $40,500. That is what the buyer can still commit for the month. Now suppose a slow start brings the month’s sales re-forecast down to $150,000. With the same ending inventory target, planned purchases fall to $520,000 + $150,000 + $20,000 − $480,000 = $210,000, and OTB falls to $60,000. A $30,000 drop in sales takes $30,000 out of the budget. If the buyer has already spent the original $90,000, the department is $30,000 overbought, and the choices are to cancel or delay orders, accept a higher ending stock, or plan deeper markdowns to clear it. Figure: The example’s planned purchases, split into what is already on order and what is still open, before and after the sales re-forecast (at retail, illustrative numbers). Two bars. At the start of the month, planned purchases of $240,000 are $150,000 on order and $90,000 open to buy. After sales are re-forecast to $150,000, planned purchases are $210,000: the same $150,000 on order and $60,000 open, with $30,000 of budget taken away. The example shows why OTB has to move with the sales re-forecast. A budget calculated at the start of the month and left alone would have let the buyer spend money the plan no longer supports. ### How buying teams manage OTB today (https://oodaris.ai/en/agents/open-to-buy#today) OTB planning is usually owned by merchandise planning or merchandise finance and used by buyers. Each week the planner loads actual sales and stock, rolls the re-forecast forward and reissues OTB by department and month. Buyers check it before raising purchase orders, and the weekly business review covers where the business is overbought and where there is room to chase. Most of this runs in spreadsheets fed by exports from the merchandising system or ERP. The workbook holds the plan, the actuals and the on-order file, and its formulas produce OTB. It works until something moves fast. A cost change, a supplier delay or a sales spike means rebuilding the numbers by hand, and in the meantime buyers commit receipts against last week’s numbers. Practices that keep OTB useful: - Hold part of the season’s budget back when buying pre-season, so there is money to chase in-season winners. - Track OTB by the month goods arrive, since that is when they count against the inventory plan. - Keep OTB at cost as well as at retail when margins differ a lot between vendors or categories. - Move budget between departments and channels when demand shifts, rather than letting one area sit overbought while another runs short. - Recalculate whenever a large order, a cost change or a new re-forecast lands, not only at month-end. ### Where OTB sits in the planning cycle (https://oodaris.ai/en/agents/open-to-buy#otb-and-assortment) OTB is the link between the financial plan and the product decisions. The merchandise financial plan sets the money. OTB says how much of it is still uncommitted for each month. The [assortment plan](https://oodaris.ai/en/agents/assortment-planning) decides which styles that money buys, and purchase orders commit it. When the plan changes, OTB is the first number that tells buyers what the change means for them. Because of that position, OTB is only as good as the plan beneath it. A plan that is re-forecast weekly gives buyers a budget that reflects the current trade. A plan that closes at month-end gives them a budget that is weeks old. ### How the OODARIS OTB Controller agent runs it (https://oodaris.ai/en/agents/open-to-buy#oodaris) In OODARIS, open-to-buy is part of MFP & OTB, one of the ten capabilities the Agentic OS covers, built from the plan of record that merchandise financial planning produces. It is part of how OODARIS applies [agentic AI to retail planning](https://oodaris.ai/en/agentic-ai-for-retail). The OTB Controller is the specialist agent for it. - Watches: Open-to-buy by department, channel and month, against receipts, orders and sales. - Proposes: Budget moves between departments and channels when demand shifts. - Approved by: Merchandise finance, with the buying lead. - Writes back: The approved budgets, to your planning or ERP system. It rolls the open-to-buy forecast forward from the latest demand, checks cash, capital and margin limits before proposing a move, and drafts the open-to-buy notes for the weekly business review. It works with finance planning systems, assortment plans, and the ERP ledger and inventory valuation. With OODARIS, a cost change, such as a new tariff on the autumn buy, rebuilds OTB and the margin impact by department before the next purchase order goes out. In the illustrative product view on the Plan page, the steps are: the new tariff is applied to autumn receipts, OTB is re-projected by department and channel, the margin impact is restated against the plan, and open purchase orders are held for approval. What is written back is OTB by department and channel, the budget buyers commit against. The Procurement Navigator works in the same crew. It watches supplier lead times, fill rates, costs and risk signals, and proposes order splits, alternate suppliers or new terms when a supplier slips. Buying and sourcing approve its changes, which are written to your ERP as purchase-order changes. By default, the person who owns the decision approves each change before anything is written back. Teams can also keep OODARIS read-only, or let it act on its own inside limits the decision’s owner sets; [how it works](https://oodaris.ai/en/how-it-works) describes the four levels. Merchandise financial planning and open-to-buy run in production at ALDO Group. Open-to-buy runs in OODARIS, built from the plan of record, and its output feeds ALDO Group’s own reports and dashboards, as well as OODARIS. [ALDO Group’s story](https://oodaris.ai/en/customers) has the detail. ### Questions Q: What is the open to buy formula? A: OTB = planned EOP inventory + planned sales + planned markdowns − BOP inventory − on order. The first four terms give planned purchases for the period; subtracting what is already on order leaves what can still be bought. Q: How do you calculate open to buy at cost? A: Either calculate OTB at retail and multiply it by the cost complement (1 minus the markup as a share of retail), or plan every term at cost from the start. With a 55% markup, $90,000 of OTB at retail is $40,500 at cost. Q: What does negative open to buy mean? A: The business is overbought for that period: the goods already on order exceed the purchases the plan supports. The usual responses are to cancel or delay orders, move budget from another department, accept a higher ending stock, or plan markdowns to clear the excess. Q: What is OTB planning? A: OTB planning is the process of setting and maintaining open-to-buy budgets: deriving them from the merchandise financial plan, splitting them by department, channel and month, updating them as sales and orders change, and deciding where to move budget when demand shifts. Q: What is the difference between OTB and MFP? A: The merchandise financial plan sets targets for sales, margin and inventory. OTB is the budget for buying that falls out of that plan once existing orders are subtracted. Q: Does the OTB Controller move budget on its own? A: Not by default. It proposes budget moves between departments and channels, after checking cash, capital and margin limits, and merchandise finance approves with the buying lead before the approved budgets are written back to your planning or ERP system. ## Assortment planning and the Assortment Curator agent (https://oodaris.ai/en/agents/assortment-planning) Assortment planning decides which products a retailer carries, in which stores and channels, and in what breadth and depth. This guide covers how planners build an assortment today, the arithmetic that limits a range, and how the OODARIS Assortment Curator agent proposes the range for each store cluster. ### What assortment planning is (https://oodaris.ai/en/agents/assortment-planning#definition) Assortment planning is the process of choosing the mix of products a retailer will offer in a season or period, and where each product will be offered. It answers questions such as how many styles to carry in each category, which ones, in which colors and price points, in which stores and channels, and how deep to buy each one. A few terms come up in every assortment conversation: - Option: a style in one color. A black leather loafer and a tan leather loafer are two options of the same style. - Breadth: how many different options a category carries. - Depth: how many units of each option are carried, including the spread across sizes. - Store cluster: a group of stores with similar sales volume, climate, space or customers that receive the same range. - Line review: the meeting where merchants and planners decide which options to keep, add and drop. Assortment planning sits between the money and the units. [Merchandise financial planning](https://oodaris.ai/en/agents/merchandise-financial-planning) sets the sales and inventory budget for a category. The assortment plan decides which products fill that budget in each cluster. [Allocation](https://oodaris.ai/en/agents/allocation) then decides how many units of each product go to each store, in which sizes. ### How assortment planning is done today (https://oodaris.ai/en/agents/assortment-planning#today) Before a season, merchants and assortment planners review last season by style and by attribute: what sold through at full price, what needed markdowns, what ran out early. They combine that with trend input, vendor offers and the financial plan to draft a range. Stores are grouped into clusters, usually by sales volume and climate and sometimes by customer profile, and each cluster gets a range sized to its space. The draft is then argued through line reviews. Each option is defended or cut, and the outcome is a range plan: which options go to which clusters, with planned units. Much of this lives in spreadsheets and slide decks, and the line review pack is often assembled by hand from several reports. The common weak points are well known to anyone who has sat through a line review: - Ranges are localized only at a coarse level, so every store in a large cluster gets the same products. - Space is checked late, and allocation finds there is no room for the depth that was bought. - Decisions lean on last season’s top sellers, which may not repeat. - The reasons behind adds and drops are not written down, so next season the argument starts again from scratch. ### Breadth, depth and space: a worked example (https://oodaris.ai/en/agents/assortment-planning#example) Space and depth together cap how many options a store can carry. Take a cluster whose stores have room for about 360 pairs of women’s boots, on the floor and in the stockroom. To present a full size run, each option needs at least one pair in each of six sizes and two pairs in the four core sizes: 10 pairs per option. ``` Maximum options = capacity ÷ minimum units per option = 360 ÷ 10 = 36 ``` If the category plan calls for 30 options at that depth, the range fits with room to chase. If merchants want 45, something has to give: fewer options, shallower depth that risks broken size runs, or space taken from another category. Working this out at the assortment stage is cheaper than discovering it during allocation. Figure: The cluster’s capacity in options: 360 pairs at 10 pairs per option, against a plan of 30 options and an ask of 45 (illustrative numbers). Two rows of option bars against a capacity line at 36 options. A plan of 30 options fills 30 places and leaves 6 to chase. An ask of 45 fills all 36 and leaves 9 options with no room. The main performance measure in a line review is sell-through: units sold divided by units received, for a set period. An option that received 1,200 units and sold 840 at full price in its first eight weeks has a full-price sell-through of 70%. Comparing sell-through by cluster and by attribute (color family, heel height, price point) shows which products travel and which are local. Eight-week full-price sell-through for three options in one cluster, and the line review outcome (illustrative numbers): | Option | Units received | Units sold at full price | Sell-through | Outcome | | --- | --- | --- | --- | --- | | Chelsea boot, black | 1,200 | 840 | 70% | Keep, buy deeper | | Leather loafer, black | 900 | 540 | 60% | Keep | | Leather loafer, tan | 900 | 270 | 30% | Drop in this cluster | Sell-through alone can mislead. A low number may mean the option was over-bought, sent to the wrong cluster, or broken in its core sizes early. Good line reviews read sell-through next to depth, size availability, margin and the markdown that would be needed to clear what is left. ### Building an assortment plan, step by step (https://oodaris.ai/en/agents/assortment-planning#steps) 1. Start from the money. Take the sales, margin and inventory budget for each category and cluster from the merchandise financial plan, so the range is sized to what the business can afford to hold. 2. Review the last comparable season. Rank options and attributes by full-price sell-through, rate of sale and margin, cluster by cluster, and note which results were distorted by stockouts or late deliveries. 3. Set the cluster structure. Decide which stores share a range, and check that each cluster’s space and depth limits are known before any option is chosen. 4. Draft the range. For each cluster, decide which options to keep, which to add from new lines and trend input, and which to drop, with breadth and depth that fit the space. 5. Review and approve. Take the draft through the line review with merchants, buyers and planners, and record the reason for each add and drop. 6. Hand over to buying and allocation. The approved range by cluster sets what buyers order within open to buy and what allocation plans to send. In season, the plan is not finished. Early sell-through on new options tells the team where to chase, where to stop reordering, and which clusters should swap options before the season ends. ### Measures assortment planners use (https://oodaris.ai/en/agents/assortment-planning#measures) - Sell-through: units sold as a share of units received, at full price and overall. - Rate of sale: units sold per store per week, which compares options with different store counts. - Gross margin and GMROI: what an option earns for the inventory it ties up. - Weeks of supply: how long current stock lasts at the current rate of sale. - Attribute performance: sales and sell-through by color, fabric, fit or price point, across options. - Exit risk: how much of an option is likely to need a markdown to clear by the end of its life. Short definitions of these measures are in the [retail planning glossary](https://oodaris.ai/en/glossary). ### How the OODARIS Assortment Curator agent works (https://oodaris.ai/en/agents/assortment-planning#oodaris) In OODARIS, assortment is one of the ten capabilities the Agentic OS covers, and part of how OODARIS applies [agentic AI to retail planning](https://oodaris.ai/en/agentic-ai-for-retail). The Assortment Curator is the specialist agent for it. - Watches: Sell-through, attribute performance and space, by store cluster. - Proposes: Adds, keeps and drops for each cluster, with the reason for each. - Approved by: Assortment planning, with the buyer. - Writes back: The approved range by cluster, to your assortment plan. It runs line reviews by store and by cluster, localizes ranges to regional demand, and exports the review for merchant and supplier meetings. It works with product information management, catalog and planogram tools, and market and trend data. With OODARIS, each option is ranked by local demand, space, margin and exit risk for every store cluster before the range is committed. The approved range by cluster is what allocation plans from. Two agents work alongside it. The Trend Intelligence Scout matches social, search and influencer signals to your product attributes and proposes early demand scores for new-to-market styles, by region; it writes nothing directly, and its scores feed the range and launch proposals. The New Product Catalyst watches early sales of new styles against their launch plan and proposes buy sizes for each launch wave, test groups, and when to chase or cut; the buyer approves its proposals. By default, the person who owns the decision approves each change before anything is written back. Teams can also keep OODARIS read-only, or let it act on its own inside limits the decision’s owner sets; [how it works](https://oodaris.ai/en/how-it-works) describes the four levels. Assortment is one of the five capabilities available to new clients today. ### Questions Q: What is the difference between assortment planning and merchandise financial planning? A: Merchandise financial planning sets the money: sales, margin, inventory and receipts by department and period. Assortment planning decides which products fill that money in each store cluster. Q: What is the difference between breadth and depth in an assortment? A: Breadth is the number of different options a category carries. Depth is how many units of each option are carried, including the spread across sizes. With fixed space, more breadth means less depth per option. Q: What is store clustering in assortment planning? A: Store clustering groups stores with similar sales volume, climate, space or customers so that each group can get a range suited to it. The cluster, not the single store, is usually the level at which ranges are set. Q: How is assortment planning different from allocation? A: Assortment planning decides which products each cluster carries. Allocation decides how many units of each product, in which sizes, each store receives and when. Q: How does the Assortment Curator decide what to drop? A: It ranks each option by local demand, space, margin and exit risk for every store cluster and proposes adds, keeps and drops for each cluster, with the reason for each. The assortment planner and the buyer approve before the range is written back to your assortment plan. ## Retail allocation and the allocation agent (https://oodaris.ai/en/agents/allocation) Allocation decides how many units of each product, in which sizes, go to each store, and when. This guide covers how allocators work today, how size curves and a need-based allocation are calculated, and how OODARIS proposes transfers and pack changes before a size run breaks. ### What allocation is (https://oodaris.ai/en/agents/allocation#definition) In retail, allocation is the process of distributing inventory from distribution centers, and sometimes straight from vendors, to stores and channels. Where the [assortment plan](https://oodaris.ai/en/agents/assortment-planning) says which products each store cluster carries, allocation says how many units of each product each store receives, in which sizes, and when. It happens in three moments: - Initial allocation: the first delivery of a new product, before or at launch, usually with part of the buy held back in the distribution center. - In-season allocation: follow-up deliveries from the held-back stock, sent to the stores where the product is actually selling. - Transfers: moving stock between stores, or back to a warehouse, to fix imbalances that allocation could not prevent. In fashion and footwear, allocation is mostly about sizes. A store can hold plenty of a style in total and still lose sales because its core sizes are gone. A size run is broken when key sizes are missing. Customers who need those sizes walk away, and the sizes that remain usually end up marked down. ### How allocators work today (https://oodaris.ai/en/agents/allocation#today) An allocator typically looks after a set of categories across a few hundred stores. For a new product they set store grades (A, B and C stores by sales volume, for example), decide how much to send up front and how much to hold back, choose size curves and packs, and release the allocation. In season they watch sell-through and stock by store, send follow-up allocations and arrange transfers. The tools are the allocation module of the merchandising system plus spreadsheets for the exceptions. Size curves are often set at category level from historical sales, and prepacks (cartons with a fixed mix of sizes) are bought from vendors in one or two configurations. Packs keep handling costs down, but they force the same size mix on stores whose customers differ. Reporting adds delay. Store-level stock by size is usually reviewed in a weekly report, so a broken size run is often found after the best stores have already sold through their core sizes. By then the sale is lost, and the fix is a transfer that arrives late or a markdown on the sizes left behind. ### Initial allocation: how much to send and how much to hold back (https://oodaris.ai/en/agents/allocation#hold-back) The first decision for a new product is how much of the buy to send at launch. Send too much and stock sits in stores where the product does not sell, to be transferred or marked down later. Send too little and the best stores run out in the first weeks, before the warehouse can react. Most allocators send an opening quantity and hold the rest back in the distribution center, then release it as sales come in. The opening quantity usually covers the presentation minimum (enough to show a full size run on the floor) plus a few weeks of expected sales, with more for stores graded A and less for C stores. The hold-back is larger for new and untested products and smaller for proven repeats, where the forecast is more reliable. A product with a short selling window, such as a holiday style, may be sent almost in full, because there is no time to react. Store grades are a blunt tool. Two stores with the same volume can sell very different sizes, colors and price points. Allocating by store and by size, from each store’s own sales and the product’s early results, gets closer to what each store will sell, at the cost of more calculation than a person can do by hand across hundreds of stores. ### Size curves and packs: a worked example (https://oodaris.ai/en/agents/allocation#size-curves) A size curve is the share of a style’s sales that each size takes. It is built from sales history, ideally for the same type of product in the same cluster, and corrected for periods when a size was out of stock (otherwise a size that sold out early looks less popular than it is). A women’s boot size curve for one cluster, and the closest 12-pair prepack (illustrative numbers): | Size | 6 | 6.5 | 7 | 7.5 | 8 | 8.5 | 9 | | --- | --- | --- | --- | --- | --- | --- | --- | | Share of sales | 8% | 12% | 18% | 20% | 18% | 14% | 10% | | Pairs in a 12-pair pack | 1 | 2 | 2 | 2 | 2 | 2 | 1 | Figure: The size curve against the 12-pair prepack: each size’s share of sales beside its share of the pack (illustrative numbers). Paired bars for sizes 6 to 9. Size 7.5 takes 20% of sales but gets 2 pairs in 12, about 17%, so it breaks first. Size 6.5 takes 12% of sales and also gets 2 pairs, so it sits. Twelve pairs cannot match the curve exactly. Size 7.5 takes 20% of sales but gets 2 pairs in 12 (17%), while size 6.5 takes 12% and also gets 2 pairs. A store that receives only this pack will break size 7.5 first and sit on 6.5. Across a few hundred stores with different customers, those small mismatches add up to the broken runs and end-of-season markdowns allocators spend their week chasing. ### Need-based allocation: a worked example (https://oodaris.ai/en/agents/allocation#need) Most in-season allocation is need-based: each store gets what it needs to reach a target level of cover, given what it already holds and what is on the way. ``` Target stock = forecast weekly sales × target weeks of cover Store need = target stock − (on hand + in transit) ``` Three stores, one style, a four-week cover target and 12-pair packs, rounded to the nearest whole pack (illustrative numbers): | Store | Forecast weekly sales | Target stock | On hand + in transit | Need | Packs sent | | --- | --- | --- | --- | --- | --- | | A | 10 | 40 | 24 | 16 | 1 | | B | 4 | 16 | 5 | 11 | 1 | | C | 2 | 8 | 12 | −4 | 0 | Store A needs 16 pairs but packs come in 12s, so it gets one pack and ends just under its target. Store B needs 11 and gets 12. Store C holds 4 pairs more than it needs, so it gets nothing, and if store A is short in its core sizes, a transfer of those sizes from store C may beat waiting for the next delivery. Real allocation adds constraints to this arithmetic: the stock available in the warehouse, minimum presentation quantities, store capacity and the size curve of each store. ### Measures allocators watch (https://oodaris.ai/en/agents/allocation#measures) - Weeks of cover: stock divided by the forecast weekly sales, by store and size. - Sell-through: units sold as a share of units received, by store. - Size integrity: the share of stores that still hold every core size of a style. - Stock-to-sales: the store’s share of stock against its share of sales, to spot over- and under-allocated stores. Definitions of these and related terms are in the [retail planning glossary](https://oodaris.ai/en/glossary). Allocation also works closely with [replenishment](https://oodaris.ai/en/agents/replenishment), which keeps continuing products in stock by reordering them as they sell. ### How OODARIS runs allocation (https://oodaris.ai/en/agents/allocation#oodaris) In OODARIS, allocation is one of the ten capabilities the Agentic OS covers: packs, size curves and transfers, store by store. It is part of how OODARIS applies [agentic AI to retail planning](https://oodaris.ai/en/agentic-ai-for-retail). On the agent roster, the allocation agent is the Inventory Optimizer, which works alongside the Replenishment Director. - Watches: Stock, sell-through and weeks of cover by store, warehouse and size. - Proposes: Transfers and safety-stock changes before a size run breaks. - Approved by: Allocation. The allocator approves each transfer and order. - Writes back: Approved transfer orders, to your ERP or warehouse system. It plans across stores and warehouses at SKU and size level, sets safety stock from demand variance, and respects vendor lead times and minimum order quantities. It works with ERP and order management, warehouse management and supplier EDI feeds. With OODARIS, breaks are caught door by door, and each one arrives with a pack change, a transfer plan and its commercial impact. What is written back is transfer orders and pack changes, to your ERP. The inputs come from the other decisions: the approved range by cluster, which allocation plans from, and the Demand Intelligence Agent’s approved forecast, which the plan, buy and allocation agents use. By default, the person who owns the decision approves each change before anything is written back. Teams can also keep OODARIS read-only, or let it act on its own inside limits the decision’s owner sets; [how it works](https://oodaris.ai/en/how-it-works) describes the four levels. Allocation is one of the five capabilities available to new clients today. ### Questions Q: What is the difference between allocation and replenishment? A: Allocation distributes a limited quantity of a product, often a seasonal one, across stores and sizes. Replenishment reorders continuing products as they sell, to keep them in stock. Q: What is a size curve? A: A size curve is the share of a style’s sales that each size takes, built from sales history for similar products and stores. Allocators use it to decide how many units of each size to send and which packs to buy. Q: What is a broken size run? A: A size run is broken when a store is missing key sizes of a style while it still holds others. Customers who need the missing sizes leave without buying, and the remaining sizes are often marked down at the end of the season. Q: What is the difference between push and pull allocation? A: Push allocation sends stock to stores from a central plan, usually at launch, based on store grades and forecasts. Pull allocation sends stock in response to what each store actually sells, usually in season from held-back warehouse stock. Q: Does the allocation agent move stock on its own? A: Not by default. The Inventory Optimizer proposes transfers and safety-stock changes, and the allocator approves each transfer and order before the approved transfer orders are written back to your ERP or warehouse system. ## Retail replenishment and the Replenishment Director agent (https://oodaris.ai/en/agents/replenishment) Replenishment keeps continuing products in stock by reordering them as they sell. This guide covers the reorder point, safety stock and order-up-to calculations with a worked example, how replenishment teams work today, and how the OODARIS Replenishment Director agent drafts the daily order. ### What replenishment is (https://oodaris.ai/en/agents/replenishment#definition) Replenishment is the process of reordering products a retailer sells continuously, so that stock is there when customers want it without holding more than the business needs. It applies to basics and continuity lines, such as the core sizes of a classic loafer or everyday accessories, rather than to seasonal fashion, which is usually bought once and then [allocated](https://oodaris.ai/en/agents/allocation). There are two flows. Store replenishment moves stock from a distribution center to stores. Distribution center replenishment buys stock from vendors through purchase orders. Both answer the same two questions for every item and location: when to order, and how much. Replenishment runs on a few parameters per item and location: the forecast rate of sale, the lead time from order to arrival, how often the item is reviewed, the safety stock held against uncertainty, and the ordering rules (pack size, minimum order quantity and vendor schedules). Get the parameters right and most orders can be generated by a calculation. Get them wrong and the team spends its week correcting suggested orders by hand. ### How replenishment teams work today (https://oodaris.ai/en/agents/replenishment#today) Replenishment analysts set parameters for each item and location, such as min and max levels, reorder points, safety stock and review cycles. A replenishment system or the ERP generates suggested orders from those parameters each day or week. Analysts review the exceptions, adjust for promotions and events, and release the orders to vendors or to the warehouse. The same problems come up again and again: - Parameters are set once and rarely revisited, so they drift away from current demand. - Safety stock is a flat number of weeks for a whole category, not a figure based on how variable each item’s demand is. - Vendor lead times are taken from the vendor’s promise rather than from how the vendor actually delivers. - Promotions and events cause spikes that the standard calculation does not expect. - Pack sizes and minimum orders push order quantities up, and nobody tracks the stock that rounding adds. The result is a team that spends most of its time writing and checking orders instead of improving the parameters that produce them. ### Reorder point, safety stock and order quantity (https://oodaris.ai/en/agents/replenishment#formulas) A periodic-review system, the most common setup in retail, uses three calculations: ``` Reorder point = forecast demand during the lead time + safety stock Order-up-to level = forecast demand over (lead time + review period) + safety stock Order quantity = order-up-to level − (on hand + on order), rounded to the pack size and at least the minimum order ``` At each review, if the inventory position (on hand plus on order) is at or below the reorder point, an order is placed to bring it back up to the order-up-to level. Min/max systems work the same way, with the reorder point as the min and the order-up-to level as the max. Safety stock covers the uncertainty in demand during the lead time. A common way to set it uses the standard deviation of demand: ``` Safety stock = z × σ × √(lead time) ``` Here σ is the standard deviation of demand per period, the lead time is in the same periods, and z is the service-level factor: about 1.28 for a 90% cycle service level, 1.65 for 95% and 2.33 for 99%. The formula assumes a steady lead time. When lead times vary too, safety stock has to cover that variation as well, which is why unreliable vendors cost inventory. ### A worked replenishment example (https://oodaris.ai/en/agents/replenishment#example) A core sneaker at one distribution center (illustrative numbers): | Input or result | Value | | --- | --- | | Forecast demand | 120 pairs a week | | Standard deviation of weekly demand | 30 pairs | | Lead time | 2 weeks | | Review period | 1 week | | Service level | 95% (z = 1.65) | | Pack size and minimum order | Packs of 24, at least 96 pairs | | Safety stock | 1.65 × 30 × √2 = 70 pairs | | Reorder point | 120 × 2 + 70 = 310 pairs | | Order-up-to level | 120 × 3 + 70 = 430 pairs | | Inventory position | 220 on hand + 60 on order = 280 pairs | | Order | 430 − 280 = 150 pairs, rounded up to 7 packs = 168 pairs | The inventory position of 280 is below the reorder point of 310, so an order goes out. The raw quantity of 150 pairs becomes 168 once it is rounded up to whole packs, which also clears the 96-pair minimum. Figure: The core sneaker over five weeks, reviewed weekly: the first order from the table, then the same rules each week (pairs, illustrative numbers). A line chart of the inventory position. At the first review it is 280, below the reorder point of 310, so an order of 168 pairs brings it to 448. It falls by 120 a week. At 328 there is no order; at 208 an order of 240 pairs brings it back to 448. The dashed lines are the order-up-to level of 430 and the reorder point of 310, and safety stock is 70. Now suppose the vendor’s lead time slips from two weeks to three. Safety stock becomes 1.65 × 30 × √3 = 86 pairs, the reorder point becomes 120 × 3 + 86 = 446 and the order-up-to level 120 × 4 + 86 = 566. The same inventory position of 280 now calls for 286 pairs, or 12 packs (288 pairs). One extra week of lead time raised the order by more than two-thirds. Lead time matters as much as the sales forecast, which is why replenishment teams track how vendors actually deliver. ### Choosing service levels and review cycles (https://oodaris.ai/en/agents/replenishment#service-levels) Not every item deserves the same service level. Each step up costs more safety stock than the last: in the formula above, moving from 95% to 99% raises z from 1.65 to 2.33, or about 41% more safety stock for four points of service. Teams usually set higher service levels for items whose stockouts hurt most, such as core sizes of best-selling continuity styles and items customers expect to find every time, and lower ones for slow movers and items with close substitutes. The review cycle is a trade-off too. Reviewing daily lets orders follow demand closely and keeps the order-up-to level low, but it means more, smaller orders, which pack sizes, minimum orders and delivery schedules may not allow. Weekly review is common for vendor orders, daily review for store orders from a nearby distribution center. Finally, parameters need upkeep. Forecasts, demand variability and vendor lead times change through the year, and a reorder point set in spring can be wrong by autumn. Reviewing the parameters on a schedule, and whenever a vendor’s delivery record changes, keeps the calculation honest. ### Measures replenishment teams watch (https://oodaris.ai/en/agents/replenishment#measures) - In-stock rate: the share of item-location combinations with stock on hand. - Fill rate: the share of demand, or of ordered units, met from stock. - Weeks of supply: stock divided by the forecast weekly sales. - Forecast error: how far actual sales landed from the forecast, which drives the safety stock needed. - Vendor lead time and fill rate: how long vendors actually take, and how much of each order they ship. Short definitions are in the [retail planning glossary](https://oodaris.ai/en/glossary). Replenishment orders also draw on the buying budget, so they show up in [open to buy](https://oodaris.ai/en/agents/open-to-buy) as stock on order. ### How the OODARIS Replenishment Director agent works (https://oodaris.ai/en/agents/replenishment#oodaris) In OODARIS, the Replenishment Director works in the allocation crew, alongside the Inventory Optimizer. It is part of how OODARIS applies [agentic AI to retail planning](https://oodaris.ai/en/agentic-ai-for-retail). On the agents page, it is where OODARIS suggests starting when order writing eats the week. - Watches: Daily sales and stock on hand and on order, against reorder points. - Proposes: The daily replenishment order, within pack, minimum-order and vendor rules. - Approved by: Replenishment or allocation. - Writes back: Approved orders, to your ERP purchasing module. It drafts daily orders that respect pack sizes and minimums, scores suppliers and feeds the scores back to the other agents, and leaves your system of record as the authority. It works with ERP purchasing, supplier portals and transportation management. In the open-to-buy crew, the Procurement Navigator also scores suppliers on delivery, quality and risk. OODARIS reads from the systems you run today and writes changes back to the system that owns them. Each writeback uses the target system’s own APIs or event paths and is logged with the decision that caused it. By default, the person who owns the decision approves each change before anything is written back. Teams can also keep OODARIS read-only, or let it act on its own inside limits the decision’s owner sets, with every action logged with its reason and reversible; [how it works](https://oodaris.ai/en/how-it-works) describes the four levels. ### Questions Q: How do you calculate a reorder point? A: Reorder point = forecast demand during the lead time + safety stock. With 120 units a week of demand, a two-week lead time and 70 units of safety stock, the reorder point is 310 units. Q: How much safety stock should a retailer hold? A: Enough to cover demand uncertainty during the lead time at the service level you want. A common formula is z × σ × √(lead time), where σ is the standard deviation of demand per period and z is about 1.65 for a 95% service level. More variable demand or longer, less reliable lead times need more safety stock. Q: What is the difference between min/max and order-up-to replenishment? A: They are closely related. In a min/max system, an order is placed when stock falls to the min and brings it back up to the max. In an order-up-to system, the reorder point plays the role of the min and the order-up-to level the role of the max, both calculated from the forecast, lead time and safety stock. Q: What is the difference between replenishment and allocation? A: Replenishment reorders continuing products as they sell. Allocation distributes a limited quantity of a product, often seasonal, across stores and sizes. Q: Does the Replenishment Director send orders to suppliers? A: It drafts the daily replenishment order within pack, minimum-order and vendor rules. By default, replenishment or allocation approves it, and approved orders are written back to your ERP purchasing module. Your system of record stays the authority. ## Markdown optimization and the Pricing Strategist agent (https://oodaris.ai/en/agents/markdown-optimization) Markdown optimization decides how deep to cut a product’s price, and when, so it sells through by the end of its season with as much margin left as possible. This guide gives the formulas with a worked example, shows how planners manage markdowns today, and explains how the OODARIS Pricing Strategist agent proposes ladders for each cluster. ### What markdown optimization means (https://oodaris.ai/en/agents/markdown-optimization#definition) Markdown optimization is choosing how deep to cut a product’s price, and when, so it sells through by its exit date, the week it is due to leave the range, with as much gross margin left as possible. The result is a markdown ladder for each style and store cluster: a planned sequence of price cuts, each with a depth and a start date. A markdown is a permanent or temporary price cut, usually to clear stock that is selling slower than planned. The [merchandise financial plan](https://oodaris.ai/en/agents/merchandise-financial-planning) sets a markdown budget for each department and month, and markdown optimization decides how it is spent, style by style. It balances four things: - Depth: too shallow and the stock does not move; too deep and margin goes on units that would have sold anyway. - Timing: an early first cut can be shallow, because there is time left to sell; a late one usually has to be deep. - Sell-through target: the share of the stock to sell by the exit date. - Margin: what is left after the discounts, including the value of unsold stock. Price elasticity links depth to units: it says how many more units a given cut will sell. Clearance is the ladder’s last step, when what remains is marked down deeply to sell before the season ends or to make room for new receipts. ### The markdown formulas (https://oodaris.ai/en/agents/markdown-optimization#formula) The first check is whether the stock will sell by its exit date at today’s price, using weeks of supply (also called weeks of cover): ``` Weeks of supply = units on hand ÷ forecast weekly units at the current price ``` When weeks of supply is longer than the weeks left, the product will not sell through at its current price, and the excess is what a markdown has to move. Two more formulas size the cut: ``` Discount depth = (original price − marked-down price) ÷ original price Price elasticity = % change in units sold ÷ % change in price ``` Elasticity is negative, since units rise as the price falls; the larger its size, the more units each point of discount brings. It is estimated from past price changes on similar products and differs by category, cluster and point in the season. Ladders are then compared on the margin they leave, not on sell-through alone: ``` Gross margin = revenue at the prices achieved − unit cost × units sold ``` Units still unsold at the exit date count too: they go to an outlet or an off-price buyer, often below cost, or are carried into another season. Teams also differ in when they record a markdown: a permanent cut is often booked on all the stock when the price changes, a temporary one at the till as each unit sells. The [retail planning glossary](https://oodaris.ai/en/glossary) defines each term. ### A worked markdown optimization example (https://oodaris.ai/en/agents/markdown-optimization#example) Take one boot style in one store cluster with 1,000 pairs on hand, a $100 original price, a $40 unit cost and eight weeks to its exit date. At full price the forecast is 50 pairs a week: 20 weeks of supply, against eight weeks left. The target is 85% sell-through, and the planner compares two ladders. Two ladders for the style, and the pairs forecast at each step (illustrative numbers): | Ladder | Weeks | Discount | Price | Pairs forecast | | --- | --- | --- | --- | --- | | A | 1 to 4 | 40% | $60 | 480 | | A | 5 to 8 | 60% | $40 | 440 | | B | 1 to 2 | 25% | $75 | 200 | | B | 3 to 4 | 40% | $60 | 240 | | B | 5 to 8 | 60% | $40 | 450 | The forecasts come from the style’s elasticity and allow for pairs a deep first cut sells early that would have sold later anyway. Ladder A sells 920 pairs, a 92% sell-through, for revenue of 480 × $60 + 440 × $40 = $46,400 and a gross margin of $46,400 − 920 × $40 = $9,600. Ladder B sells 890 pairs, an 89% sell-through, for revenue of 200 × $75 + 240 × $60 + 450 × $40 = $47,400 and a gross margin of $47,400 − 890 × $40 = $11,800. Both ladders clear the target. Ladder A sells 30 more pairs but earns $2,200 less, because its deeper first cut gives money away on pairs that would have sold at $75. Writing off the 80 and 110 pairs left over would not change the answer. Now suppose ladder B’s first two weeks sell 120 pairs, not 200. If the rest of the forecast holds, 1,000 − 120 − 240 − 450 = 190 pairs are left at the exit date, an 81% sell-through, short of the target. The planner can bring the 40% cut forward, deepen the last cut or plan a bigger clearance, and checks each option on margin. Figure: The example’s two ladders and what each earns (illustrative numbers). Two step lines over eight weeks from a $100 price. Ladder A drops to $60, then to $40 in week 5: 920 pairs and $9,600 of gross margin. Ladder B drops to $75, then to $60 in week 3 and $40 in week 5: 890 pairs and $11,800. The example shows why depth is a margin decision, not only a sell-through one, and why a ladder has to move with each week’s sales: left alone after its slow start, ladder B would miss its target. ### How planners manage markdowns today (https://oodaris.ai/en/agents/markdown-optimization#today) Markdowns are usually proposed by merchandise planning or a pricing team, agreed with the buyer, and kept within a margin limit that merchandise finance sets. Each week the planner reads sell-through and weeks of supply by style and decides which styles take a first or next cut. In stores each change means new tickets or shelf labels, so markdowns usually follow a fixed weekly cycle. Many ladders follow rules written seasons ago: a first cut after a set number of weeks, at a set depth, for every style in a category. The first cut is often deeper than the stock needs, because a planner reviewing hundreds of styles cannot weigh each one. Depths come from experience rather than a measured elasticity, and the same ladder often goes to every cluster, though a style can sell in one and sit in another. Practices that keep markdowns working: - Set an exit date and a sell-through target for every style before the season. - Start with a shallow cut and hold it long enough to read the response. - Plan by cluster or channel where prices can differ, not one ladder for every store. - Check whether a transfer to a cluster where the style sells would beat a markdown. - Review each markdown against its forecast afterwards, so next season’s depths start from evidence. ### Where markdowns sit in the planning cycle (https://oodaris.ai/en/agents/markdown-optimization#markdown-and-otb) Planned markdowns are one of the terms in the [open to buy](https://oodaris.ai/en/agents/open-to-buy) formula, because a price cut lowers the retail value of the stock just as a sale does. When markdowns run ahead of plan, gross margin falls short. When they run behind, unsold stock stays on the books at full value, and next month’s budget for receipts shrinks. They also feed the next season: what sold only at a deep discount tells the line review what to buy less of, and stock that outstays its exit date takes space and budget the new range needs. ### How the OODARIS Pricing Strategist agent runs it (https://oodaris.ai/en/agents/markdown-optimization#oodaris) In OODARIS, markdown optimization is part of Pricing, markdown & promotions, one of the ten capabilities the Agentic OS covers. It is part of how OODARIS applies [agentic AI to retail planning](https://oodaris.ai/en/agentic-ai-for-retail). The Pricing Strategist is the specialist agent for it. - Watches: Sell-through, weeks of cover, price elasticity and competitor prices, by cluster. - Proposes: Price ladders, and markdown depth and timing for each cluster, with the expected margin. - Approved by: Merchandise planning. Merchandise finance clears the margin limit. - Writes back: Approved price conditions, to your ERP or pricing system. It models price elasticity by product hierarchy and region, sequences markdowns to clear stock with shallower cuts, and answers competitor price moves with the evidence attached. It works with competitor price feeds, promotion planning tools, and web and shelf price systems. With OODARIS, price ladders follow elasticity, inside your MAP, MSRP and margin floors, and markdown depth and timing are proposed to hit sell-through. In the illustrative product view on the [Sell page](https://oodaris.ai/en/decisions/sell), a store cluster’s current ladder is compared with a proposed one with a shallower first cut, on end-of-season gross margin, average discount and sell-through, both simulated against the cluster’s own sell-through history. What is written back is the approved price conditions and the cluster’s new ladder; the approval goes on the record, and the writeback has a rollback path. The Promotions Director works in the same crew. It watches promotion calendars, uplift and cannibalization by segment and channel, and proposes offer depth, timing and audience for each campaign, with the forecast uplift. Marketing approves its offers, with merchandising, and the approved offers are written to your promotion and pricing tools. By default, the person who owns the decision approves each change before anything is written back. Teams can also keep OODARIS read-only, or let it act on its own inside limits the decision’s owner sets; [how it works](https://oodaris.ai/en/how-it-works) describes the four levels. The pricing, markdown and promotions capability is one of the five available to new clients today. ### Questions Q: How do you calculate a markdown percentage? A: Divide the cut by the original price: (original price − marked-down price) ÷ original price, so $100 to $75 is 25% off. In a merchandise plan, markdown percentage usually means the period’s markdown dollars as a share of its net sales. Q: What is a markdown ladder? A: A planned sequence of price cuts for a product: a shallow first cut, then deeper ones at set intervals, ending in clearance. Ladder B in the worked example goes from 25% off to 40% and then 60%. Q: When should a retailer take the first markdown? A: When weeks of supply at the current price is longer than the weeks left to the exit date. An early first cut can be shallow; a late one usually has to be deep. Q: How does price elasticity affect markdown depth? A: Elasticity says how many extra units a cut will sell. For an inelastic product, a deeper cut adds few units and mostly gives margin away, so a transfer or a later clearance may cost less. Q: What is the difference between a markdown and a promotion? A: A markdown cuts the price to clear stock. A promotion is a temporary offer meant to lift traffic or sales before the price returns, so its cost is judged against the lift it brings. Q: Does the Pricing Strategist change prices on its own? A: Not by default. It proposes price ladders, and markdown depth and timing for each cluster, with the expected margin. Merchandise planning approves, and merchandise finance clears the margin limit, before the approved price conditions are written back to your ERP or pricing system. ## Retail planning glossary (https://oodaris.ai/en/glossary) Short definitions of the terms planners, buyers, allocators and pricing teams use every week, across the ten planning capabilities from the demand forecast to order fulfillment. The main terms link to the pages that cover them in depth. ### Plan: Demand forecasting - [Demand forecasting](https://oodaris.ai/en/glossary#demand-forecasting): Estimating how many units of each product customers will buy, by store, channel and week, from sales history, seasonality, prices, promotions and events. The forecast feeds the merchandise plan, the assortment, allocation, replenishment and markdowns, so its errors travel down the whole planning chain. - [Baseline demand](https://oodaris.ai/en/glossary#baseline-demand): The demand a product would see at its regular price, without promotions, markdowns or one-off events. Forecasters estimate it first and add the effect of each planned price change or campaign on top, so a plan can tell normal selling from a lift that was paid for. - [Unconstrained demand](https://oodaris.ai/en/glossary#unconstrained-demand): What customers would have bought if a product had been in stock in every size and store. Sales history understates it whenever items sold out, so forecasters estimate the lost sales before using history to plan the next season. - [Seasonality](https://oodaris.ai/en/glossary#seasonality): The pattern by which demand rises and falls with the calendar: holidays, school terms, weather and the fashion seasons. A forecast separates seasonality from trend, so a strong spring is not mistaken for lasting growth, and aligns fiscal weeks that shift from year to year. - [Promotional lift](https://oodaris.ai/en/glossary#promotional-lift): The extra units a promotion sells over baseline demand while it runs. A useful estimate also counts what the promotion takes away: sales pulled forward from the weeks after it, and sales moved from other products in the range. - [Forecast accuracy](https://oodaris.ai/en/glossary#forecast-accuracy): How close a forecast came to what actually sold, measured at the level where decisions are made, such as style by store by week. Accuracy at a high level can hide large errors by store and size, which is where allocation and replenishment act. - [Forecast bias](https://oodaris.ai/en/glossary#forecast-bias): A forecast’s tendency to run consistently above or below actual demand. Bias does more damage than random error because it compounds: a forecast that always runs high buys too much, allocates too much and ends in markdowns. ### Plan: MFP & OTB - [Merchandise financial planning](https://oodaris.ai/en/glossary#mfp): Setting a retailer’s financial targets for its merchandise (sales, gross margin, markdowns, inventory and receipts by department and period) and managing the business against them. The plan is built before the season, reconciled top-down and bottom-up, and re-forecast in season. - [Open to buy](https://oodaris.ai/en/glossary#otb): The amount a buyer can still spend for a period without going over the inventory plan: the purchases the merchandise plan needs, less what is already on order. It is set by department, channel and month, and it moves whenever sales, stock or orders change. - [Plan version](https://oodaris.ai/en/glossary#plan-version): One of the parallel sets of numbers a merchandise plan keeps: the original budget approved before the season, the working plan planners change, the latest re-forecast, and last year’s actuals for comparison. Keeping versions apart shows how far the season has moved from its commitments. - [Weekly sales, stock and intake](https://oodaris.ai/en/glossary#wssi): The week-by-week view of the merchandise plan, common in UK and fashion retail. Each week shows opening stock, sales, markdowns, intake and closing stock as plan, actual and re-forecast, so a gap between plan and actuals shows while there is still time to act. - [BOP and EOP inventory](https://oodaris.ai/en/glossary#bop-eop-inventory): Beginning-of-period and end-of-period inventory: the stock a department or category holds at the start and at the end of a week or month, usually valued at retail and at cost. They are the anchors of the merchandise plan and the open-to-buy calculation. - [Receipts](https://oodaris.ai/en/glossary#receipts): Merchandise arriving into a retailer’s stock in a period. Planned receipts are what the merchandise financial plan needs to arrive for its sales and ending inventory targets to hold. - [On order](https://oodaris.ai/en/glossary#on-order): Merchandise a retailer has committed to buy but not yet received, counted by the period it is due to arrive. It is subtracted from planned purchases to give open to buy. - [Gross margin](https://oodaris.ai/en/glossary#gross-margin): Sales revenue less the cost of the goods sold, stated in currency or as a share of sales. Merchandise plans set gross margin targets by department and period, and markdowns, shrink and the cost of receipts all move it. - [Initial markup](https://oodaris.ai/en/glossary#initial-markup): The difference between a product’s first retail price and its cost, as a share of that price. Buyers plan it when they set prices and negotiate costs; it is the margin a season starts with, before markdowns and shrink take their share. - [Maintained markup](https://oodaris.ai/en/glossary#maintained-markup): The markup actually kept once markdowns, discounts and shrink have reduced the initial markup. Comparing the two shows how much margin a season gave back, and where. - [Average unit retail](https://oodaris.ai/en/glossary#average-unit-retail): The average price at which units sold: sales revenue divided by units sold. Planners read it alongside units, because a falling average with steady units usually means more markdowns or a shift toward cheaper products. - [Stock-to-sales ratio](https://oodaris.ai/en/glossary#stock-to-sales-ratio): The inventory at the start of a period set against the sales planned for it. Planners set it month by month to decide how much stock each period needs to support its sales, which in turn sets how much has to be received. - [Shrink](https://oodaris.ai/en/glossary#shrink): Inventory lost to theft, damage, administrative error or supplier fraud: the gap between the stock the books expect and the stock counted. Merchandise plans carry a shrink allowance, because every unit lost lowers both inventory and margin. - [Inventory turnover](https://oodaris.ai/en/glossary#inventory-turnover): How many times a business sells through its average inventory in a period, usually a year. Higher turn means stock sells quickly rather than sitting idle; turn that is too high can mean the business is running short. - [GMROI](https://oodaris.ai/en/glossary#gmroi): Gross margin return on inventory investment: the gross margin a product, category or store earns for each dollar of inventory held, measured at cost. It shows whether stock is earning its place, and it is used to compare categories with different margins and turn. - [Comparable sales](https://oodaris.ai/en/glossary#comparable-sales): Sales growth measured only on stores open long enough to compare with the same period last year, leaving out openings, closures and major refits. Also called like-for-like sales, it shows how existing stores are trading apart from a changing store count. ### Plan: Assortment planning & optimization - [Assortment planning](https://oodaris.ai/en/glossary#assortment-planning): Choosing which products a retailer carries in a season, in which stores and channels, and in what breadth and depth. Ranges are usually set by store cluster and agreed in line reviews, inside the budget the merchandise financial plan sets. - [Line review](https://oodaris.ai/en/glossary#line-review): The meeting where buyers, planners and merchants agree the range for a season: which options to keep, add or drop, their prices and the clusters they go to. Its decisions become the assortment plan and the starting point for purchase orders. - [Option](https://oodaris.ai/en/glossary#option): One style in one color, the unit most assortment plans count. The number of options in a range sets its breadth; each option then gets its own depth, size curve and allocation. - [Stock keeping unit](https://oodaris.ai/en/glossary#sku): The most detailed product identifier a retailer tracks: one style in one color and one size. Plans are built at higher levels, but allocation, replenishment and stock counts work at the stock keeping unit. - [Breadth and depth](https://oodaris.ai/en/glossary#breadth-and-depth): Breadth is the number of different options a category carries; depth is how many units of each option are carried, including the spread across sizes. With fixed space and budget, adding breadth reduces depth. - [Store cluster](https://oodaris.ai/en/glossary#store-cluster): A group of stores with similar sales volume, climate, space or customers that receive the same range. Clusters let a retailer localize its assortment without planning every store separately. - [Assortment capacity](https://oodaris.ai/en/glossary#assortment-capacity): How many options a store or cluster can present well, set by its space and fixtures. Capacity caps the breadth an assortment plan can send to each cluster, so smaller stores get a tighter edit of the range. - [Product lifecycle](https://oodaris.ai/en/glossary#product-lifecycle): The stages a product moves through from launch to exit: introduction, peak selling, decline and clearance. Assortments mix core products that run for years with seasonal and fashion products that live for weeks, and each is planned, bought and marked down differently. - [Sell-through](https://oodaris.ai/en/glossary#sell-through): The share of the units a store or channel received that sold in a period, often measured at full price. It is the main measure of how well a product, an option or a range performed, and it guides keep, add and drop decisions and markdown timing. ### Buy: Purchase orders & procurement - [Procurement](https://oodaris.ai/en/glossary#procurement): Buying the merchandise the plan calls for: choosing suppliers, negotiating cost and terms, and placing and tracking purchase orders. It turns open to buy into commitments, and decides what happens when a supplier ships late, short or not at all. - [Purchase order](https://oodaris.ai/en/glossary#purchase-order): A buyer’s commitment to a supplier for specific products, quantities, costs and delivery dates. Once placed, its value counts as on order and reduces open to buy, and its delivery window decides when the stock can reach stores. - [Lead time](https://oodaris.ai/en/glossary#lead-time): The time between placing an order and the stock being ready to sell. Long lead times force earlier, less certain commitments, and replenishment and safety stock both depend on how long and how variable lead time is. - [Minimum order quantity](https://oodaris.ai/en/glossary#minimum-order-quantity): The smallest order a supplier will accept for a product, a color or a whole purchase order. Minimums can force a buyer to commit more than the plan needs, so they shape which options are worth carrying in a smaller range. - [Landed cost](https://oodaris.ai/en/glossary#landed-cost): The full cost of a product once it reaches the retailer: the supplier’s price plus freight, duties, insurance and handling. Margin plans built on the supplier’s price alone overstate profit, so an estimated landed cost is used until the actual one is known. ### Buy: Sizing & prepacking - [Size curve](https://oodaris.ai/en/glossary#size-curve): The share of a style’s sales that each size takes, built from sales history for similar products and stores. Allocators use it to decide how many units of each size to send and which prepacks to buy. A size run is broken when key sizes are missing. - [Size run](https://oodaris.ai/en/glossary#size-run): The sizes a style is offered in, and the stock held in each. A size run is broken when key sizes sell out while others remain, which leaves inventory that looks healthy in total but cannot meet demand. - [Prepack](https://oodaris.ai/en/glossary#prepack): A carton holding a fixed mix of sizes, or of colors, that ships as one unit from the supplier to the store. Prepacks cut handling costs, but a pack that does not match a store’s size curve sends it sizes it will not sell. ### Flow: Allocation planning & optimization - [Allocation](https://oodaris.ai/en/glossary#allocation): Distributing inventory from distribution centers or vendors to stores and channels: how many units of each product, in which sizes, each store receives, and when. It covers the first delivery at launch, follow-up deliveries in season and transfers between stores. - [Initial allocation](https://oodaris.ai/en/glossary#initial-allocation): The first distribution of a new product to stores at launch, before any sales history for it exists. It rests on forecasts, store grades and presentation needs, and holds stock back in reserve to follow up where the product sells. - [Store grade](https://oodaris.ai/en/glossary#store-grade): A ranking of stores, usually by sales volume or capacity, that sets how much of a product each receives and which options it carries. Grades are simpler than store clusters and are often used with them. - [Presentation minimum](https://oodaris.ai/en/glossary#presentation-minimum): The fewest units a store needs to display a product properly, such as one of each size on the shelf. Allocation and replenishment do not take a store below it, even when its sales alone would justify less. - [Reserve stock](https://oodaris.ai/en/glossary#reserve-stock): Inventory held back at the distribution center instead of sent to stores at launch, so it can follow demand once early sales show where a product is selling. Too little reserve forces transfers later; too much leaves stores short. - [Weeks of supply](https://oodaris.ai/en/glossary#weeks-of-supply): How many weeks current stock would last at the expected rate of sale. Also called weeks of cover or forward cover, it is used to spot overstocked and understocked products, stores and sizes, and to set allocation and replenishment targets. ### Flow: Replenishment - [Replenishment](https://oodaris.ai/en/glossary#replenishment): Reordering products a retailer sells continuously, so they stay in stock without excess. It applies to basics and continuity lines, from a distribution center to stores and from vendors to the distribution center, within pack sizes, minimum orders and vendor schedules. - [Reorder point](https://oodaris.ai/en/glossary#reorder-point): The stock level at which a new replenishment order is placed: enough to cover expected demand until the order arrives, plus safety stock. It is set for each item and location from the forecast and the vendor’s lead time. - [Safety stock](https://oodaris.ai/en/glossary#safety-stock): Extra stock held to cover uncertainty in demand and in supplier lead times, so an item stays available until the next delivery. More variable demand, longer or less reliable lead times and higher service targets all call for more of it. - [Service level](https://oodaris.ai/en/glossary#service-level): The share of demand a retailer aims to meet from stock without running out, set for each product and location. A higher service level needs more safety stock, so it is set by weighing the cost of a lost sale against the cost of holding inventory. - [Stockout](https://oodaris.ai/en/glossary#stockout): A product or size that is out of stock where a customer wanted it. A stockout loses the sale and often the visit, and it hides true demand in the sales history the next forecast learns from. ### Flow: Transfers - [Store-to-store transfer](https://oodaris.ai/en/glossary#store-transfer): Moving stock from a store where a product is selling slowly to one where it is selling out, instead of waiting for new receipts. Transfers repair broken size runs and save markdowns, but every move has a cost, so they are planned where the margin gained exceeds it. - [Inventory rebalancing](https://oodaris.ai/en/glossary#inventory-rebalancing): Redistributing stock across stores and the distribution center to match where demand has turned out to be, usually in season. It combines transfers, releases of reserve stock and reallocation, and is judged by how much full-price selling it adds. ### Sell: Pricing, markdown & promotions - [Markdown](https://oodaris.ai/en/glossary#markdown): A permanent or temporary reduction in a product’s selling price, usually to clear stock that is selling slower than planned or reaching the end of its season. Markdowns reduce margin, so retailers plan their depth and timing, often as a ladder of steps by store cluster. - [Markdown ladder](https://oodaris.ai/en/glossary#markdown-ladder): The planned sequence of price reductions for a product: a shallow first cut, then deeper ones at set intervals until the stock is gone. A ladder chosen for each product and cluster usually sells through with less discount than one deep cut everywhere. - [Promotion](https://oodaris.ai/en/glossary#promotion): A temporary offer, such as a discount for a weekend or a multi-buy, meant to lift traffic or sales for a set period before the price returns. Unlike a markdown it is not meant to clear stock, so its cost is judged against the lift it brings. - [Discount depth](https://oodaris.ai/en/glossary#discount-depth): How far a price is cut below the original retail price, as a share of it. Planners track the average depth of markdowns and promotions, because a deeper average that sells no extra units is margin given away. - [Price elasticity](https://oodaris.ai/en/glossary#price-elasticity): How much demand for a product changes when its price changes. An elastic product gains many units from a small discount; an inelastic one sells much the same at any reasonable price, so discounting it gives margin away. - [Clearance](https://oodaris.ai/en/glossary#clearance): The last stage of a product’s life, when the remaining stock is marked down deeply to sell through before the season ends or to make room for new receipts. Planning it early, along the markdown ladder, usually recovers more margin than a late clearance. ### Sell: Order fulfillment - [Order fulfillment](https://oodaris.ai/en/glossary#order-fulfillment): Getting each customer order from available stock to the customer: choosing the location that ships it, then picking, packing and delivery or pickup. For a retailer with stores, it decides whether store stock sells online or waits for a customer to walk in. - [Order routing](https://oodaris.ai/en/glossary#order-routing): Choosing which location fulfills each online order, weighing its stock, the shipping cost, the delivery time and the risk of leaving a store short. Good routing ships from stock that would otherwise end in a markdown. - [Ship from store](https://oodaris.ai/en/glossary#ship-from-store): Fulfilling online orders from a store’s stock instead of the distribution center. It sells inventory that would otherwise wait for a walk-in customer and shortens delivery, but every unit shipped leaves the store’s shelf, so it has to respect presentation minimums. - [Click and collect](https://oodaris.ai/en/glossary#click-and-collect): An online order the customer picks up at a store, also called buy online, pick up in store. It needs store stock that is accurate in real time, and it draws on the same units the store sells to customers who walk in. - [Split shipment](https://oodaris.ai/en/glossary#split-shipment): An order sent in more than one package because no single location holds every item in it. Each extra package adds shipping cost, so routing tries to fill an order from one place when it can. ### Across every decision - [Agentic AI](https://oodaris.ai/en/glossary#agentic-ai): Software built from AI agents that sense what is happening, decide what to do and act on it, inside limits people set. In retail planning, specialist agents watch decisions such as the merchandise plan, open to buy, allocation and markdowns, propose changes with their evidence and, once approved, write them back to the systems that run the business. ## Contact (https://oodaris.ai/en/contact) Tell us which decision your team argues about every week. We'll build the first call around it. Our pricing is scoped with you in that conversation. - Book a demo: https://oodaris.ai/en/contact?topic=retail-decision-briefing - Plan a pilot: https://oodaris.ai/en/contact?topic=governed-pilot-assessment - Request an architecture review: https://oodaris.ai/en/contact?topic=technical-architecture-review - Email: contact@oodaris.ai. A reply within 24 hours: A person on the OODARIS team reads your note and writes back to set up a first call. World Trade Centre 413 rue Saint-Jacques Bureau 700 - 7097 Office Montreal, Quebec H2Y 1N9 Canada