How to Build a Weekly Merchandising Review Workflow Around Product, Inventory, and Margin

Merchandising work gets messy because product decisions rarely sit in one system.

Sales are in the ecommerce platform or POS. Inventory is in an ERP, warehouse system, spreadsheet, or supplier file. Margin may sit with finance. Returns and service issues are somewhere else. Campaign performance is in marketing tools. Product notes live with the merchandising team. By the time the team has stitched the view together, the week has already moved on.

That is why a merchandising review should not be treated as a dashboard project. The dashboard is only useful if it supports a weekly operating rhythm: which products should we promote, replenish, discount, investigate, improve, or retire?

First, be clear about the job of the workflow

The job of a merchandising review workflow is to help a commercial team make better product decisions every week.

It should help the team answer:

• Which products are selling well, but at poor margin?
• Which products have healthy demand but too little stock?
• Which products have stock but weak sell-through?
• Which products are getting traffic but not converting?
• Which products are being returned, complained about, or creating service issues?
• Which categories are over-assorted, under-assorted, or drifting from the plan?
• Which products need an owner action before the next buying, campaign, pricing, or replenishment decision?

If the workflow only shows sales by SKU, it is too thin. Sales without inventory can lead to bad replenishment decisions. Sales without margin can hide profit leaks. Sales without returns can make a weak product look healthy. Sales without channel context can create the wrong promotion plan.

How merchandising review usually happens today

In many consumer and retail teams, the current process looks like this:

• The merchandising or commercial lead pulls SKU sales from the ecommerce platform, POS, marketplace, wholesale file, or BI dashboard.
• Someone adds stock on hand, inbound stock, weeks of cover, or warehouse availability from another system.
• Finance provides gross margin, contribution margin, discount impact, or landed-cost assumptions.
• Marketing adds campaign tags, channel performance, creative notes, or traffic data.
• Customer service adds returns, complaints, sizing issues, damaged-item notes, or recurring questions, if anyone remembers to ask.
• The team reviews the spreadsheet, deck, dashboard, or meeting notes and discusses what to do.
• Actions are written in email, chat, a project tool, or nowhere consistent.

The issue is not lack of effort. It is that the product decision requires several signals, and those signals are not connected into one workflow.

Where the workflow usually breaks

The first break is SKU identity. Product IDs, variant IDs, bundle IDs, marketplace SKUs, ERP item codes, and marketing names do not always line up. A merchandising workflow cannot be trusted until product identity is clean enough to connect sales, inventory, margin, returns, and campaign data.

The second break is timing. Sales may be daily, inventory may update hourly, margin may update after finance close, returns may lag, and campaign data may be tagged late. A weekly review should show data freshness so people know which numbers are final and which are directional.

The third break is metric definition. Sell-through, weeks of cover, margin, markdown rate, return rate, conversion rate, and stockout rate are simple words, but teams often calculate them differently. If the definition changes by spreadsheet, the meeting turns into a debate about the data.

The fourth break is missing context. A SKU can be down because demand softened, stock ran out, traffic dropped, price changed, product detail content is weak, returns rose, an ad stopped, a competitor discounted, or a supplier issue delayed replenishment. The workflow needs enough context to guide the next question.

The fifth break is action tracking. Teams talk about products, but the actions do not always become owned work. The same product appears in the review three weeks later with the same unresolved question.

What good looks like

A good merchandising review workflow has a simple weekly loop:

1. Collect: pull product, sales, inventory, margin, channel, campaign, returns, and service signals into a shared view.
2. Check: validate product mappings, missing data, freshness, abnormal movements, and definition changes.
3. Flag: identify SKUs, variants, categories, or channels that need attention.
4. Review: focus the meeting on exceptions and decisions, not every product line.
5. Act: assign owners for replenishment, promotion, discount, content fix, supplier follow-up, service investigation, or retirement.
6. Learn: track whether the action improved sell-through, margin, stock health, conversion, return rate, or service volume.

The first version does not need to cover every SKU. It should cover the products where decisions matter most: top sellers, high-margin products, low-stock winners, slow-moving stock, promotion candidates, high-return products, and new launches.

Start with the merchandising action map

A weekly review becomes much sharper when each flag points to a likely action. Here is a practical starting map:

This table is not meant to replace judgement. It gives the review a practical shape. The team can still decide that a product needs no action, but it should be a conscious decision.

Define the weekly metrics properly

The merchandising review should start with a small set of trusted metrics. For many teams, the core set is:

• Net sales: sales after returns, cancellations, and relevant adjustments.
• Units sold: product units sold by SKU, variant, channel, and period.
• Gross margin or contribution margin: margin after agreed cost and discount treatment.
• Sell-through: units sold compared with units available for sale in the period.
• Stock on hand: available inventory by SKU, variant, location, and channel.
• Weeks of cover: stock on hand divided by recent or forecast weekly demand.
• Markdown or discount rate: share of revenue sold below planned price.
• Return rate: returns as a share of orders, units, or revenue.
• Conversion rate: product page or channel conversion where relevant.
• Service issue rate: tickets, complaints, reviews, or contacts tied to the product.

Each metric needs an owner, source, calculation, refresh cadence, and caveat. For example, sell-through should say whether it uses beginning inventory, available-for-sale inventory, average inventory, or a period-specific stock denominator. Margin should say whether freight, marketplace fees, duties, returns, and discounts are included.

Design the review view around exceptions

Most merchandising meetings do not need a full tour of the catalog. They need an exception view.

Useful tabs or sections include:

• Winners with constraints: products with strong demand but low stock, low margin, fulfillment issues, or weak variant coverage.
• Slow movers: products with high stock and weak sell-through.
• Margin leaks: products with strong sales but poor contribution after discounts, returns, or channel costs.
• Campaign products: products currently promoted, planned for campaign, or affected by campaign traffic.
• Launch products: new products that need early read on traffic, conversion, availability, and feedback.
• Service and return flags: products creating friction after purchase.
• Decision log: products with open actions, owners, due dates, and previous-week follow-up.

The decision log is essential. Without it, the review becomes a meeting rather than a workflow.

The data you need underneath

The first useful merchandising workflow usually needs these data sources:

• Product catalog, SKU, variant, category, collection, brand, season, price, launch date, and status.
• Sales, units, orders, net revenue, discounts, cancellations, and channel split.
• Cost, gross margin, contribution margin, landed cost, marketplace fees, shipping cost, and return cost where available.
• Inventory on hand, inbound inventory, reserved stock, stockouts, locations, reorder points, and weeks of cover.
• Traffic, product page views, add-to-cart, checkout, conversion, and source/channel data.
• Campaign tags, promotion calendar, offer rules, creative notes, and paid media activity.
• Returns, refund reasons, service tickets, reviews, product defects, sizing notes, and fulfillment issues.
• Actions, owners, due dates, status, decisions, and comments from prior reviews.

Start with the data that changes decisions. If margin data is not ready, start with sales, inventory, sell-through, and return flags. If traffic data is messy, start with weekly channel-level views before product-page detail. The first version should improve the meeting, not solve the entire data estate.

The systems usually involved

Consumer and retail merchandising data usually lives across:

• Ecommerce platform or POS: orders, units, product performance, discounts, and channel data.
• ERP, inventory, warehouse, or order-management system: stock, inbound, reserved, fulfilled, and location data.
• PIM or product catalog: product hierarchy, attributes, variants, status, descriptions, and launch notes.
• Finance or accounting system: cost, margin, freight, duties, marketplace fees, and return cost.
• Analytics and BI tools: dashboards, conversion, traffic, channel mix, and category reporting.
• Marketing platforms: campaign tags, offers, creative, email, ads, and audience data.
• Customer service and returns tools: tickets, complaints, return reasons, and product issue themes.
• Workflow tools: decision log, owner assignment, follow-up, and review cadence.

The integration does not need to be perfect on day one. The useful question is: what does the merchandiser, planner, ecommerce lead, or commercial owner need to see before making the decision?

Where AI can help

AI can help in merchandising review when the workflow already has clean enough product, inventory, margin, and action data. It should assist the decision, not invent one.

Useful AI support includes:

• Performance summaries: explain why a SKU, category, or channel changed compared with last week, plan, or prior period.
• Exception explanations: draft likely reasons for low sell-through, high returns, margin drops, or stockout risk using source-linked signals.
• Product grouping: classify products into useful review groups based on category, behavior, margin, stock, or return pattern.
• Review questions: suggest questions the team should ask before promoting, discounting, replenishing, or retiring a product.
• Commentary drafts: create first-pass notes for weekly merchandising packs or leadership updates.
• Service-theme linking: connect tickets, reviews, or return reasons back to products and categories.
• Action follow-up: identify products where last week's action was not completed or did not appear to change the metric.

The most useful AI output is usually a better starting point for the meeting: "Here are the products that need attention, here is why, here is the evidence, and here is the likely next question."

Where human review still matters

Merchandising decisions affect revenue, margin, customer experience, supplier relationships, cash tied up in inventory, and brand positioning. Human review still matters when:

• A product may be discounted, delisted, reordered, or heavily promoted.
• Margin movement depends on cost assumptions, returns, freight, or channel fees that may be incomplete.
• A product is new and the data is still thin.
• A return or service signal may indicate quality, safety, sizing, fulfillment, or supplier issues.
• A decision affects key customers, wholesale partners, retail stores, or marketplace commitments.
• AI-generated commentary will be used in a leadership, supplier, investor, or board update.

The workflow should make human review better by showing the facts, definitions, caveats, and prior actions in one place.

What to fix first

Do not start by building the perfect merchandising command center. Pick one weekly review where decisions are already being made manually.

Good starting points include:

• Top 100 SKUs by revenue or margin.
• A high-inventory category.
• A category under active promotion.
• New product launches from the last 8 to 12 weeks.
• Products with high returns or service issues.
• Products with low stock but strong sell-through.

Then build the first version around six decisions:

1. What products are in scope? top SKUs, category, launch cohort, seasonal range, channel, or promotion set.
2. What data is needed? sales, inventory, margin, sell-through, returns, traffic, campaign, and service signals.
3. What flags matter? low stock, high stock, weak margin, high returns, low conversion, stockout risk, or campaign mismatch.
4. Who owns the action? merchandising, planning, ecommerce, marketing, finance, operations, supplier, or service.
5. How is the decision captured? promote, replenish, discount, investigate, improve, pause, or retire.
6. How do we know it worked? next-week change in sell-through, margin, conversion, stock health, return rate, or action completion.

A practical 30/60/90 day path

The first project should make one weekly merchandising meeting materially better.

First 30 days: map the review and build the first view

Pick the product scope, map how data is gathered today, define the core metrics, clean the SKU mapping, and build the first exception view.

The output should be concrete:

• A product scope and category map.
• A metric dictionary.
• A SKU/variant/source-system mapping.
• A first weekly exception view.
• A merchandising action map.
• A decision log with owners and due dates.

Next 30 days: add margin, inventory, and service signals

Month two should make the review more complete. Add margin and cost assumptions, stock coverage, inbound inventory, returns, service issues, and campaign tags. Improve the exception rules based on how the first reviews performed.

The team should be able to answer:

• Which products are worth promoting now?
• Which products need replenishment before marketing pushes them?
• Which products need markdown or exit planning?
• Which products have margin or return problems?
• Which prior actions are still open?

Days 60 to 90: add AI support and close the loop

Once the workflow is stable, add AI support for performance summaries, exception explanations, review questions, service-theme linking, and action follow-up.

Start measuring the workflow:

• Products reviewed by exception rather than manually scanned.
• Actions created per weekly review.
• Actions completed before the next review.
• Stockout risk reduced for high-performing products.
• Slow-moving inventory identified earlier.
• Margin issues flagged before promotion.
• Return or service issues tied back to product decisions.

The goal is not a prettier product dashboard. It is a weekly operating loop that makes product, inventory, margin, and customer signals easier to act on.

Common mistakes

Merchandising workflow projects usually fail in familiar ways.

Mistake 1: ranking products only by revenue

Revenue can hide stock constraints, margin leakage, returns, and service problems. Use a multi-signal view.

Mistake 2: ignoring variant-level problems

A product may look healthy at parent level while key sizes, colors, packs, or locations are broken underneath.

Mistake 3: reviewing too many products

A weekly review should focus attention. If every SKU is discussed, the team will miss the products that need decisions.

Mistake 4: separating inventory from campaign planning

Promoting a product without stock, margin, or fulfillment visibility creates avoidable customer and margin problems.

Mistake 5: adding AI before definitions are clear

AI can explain and summarize, but it needs trusted SKU mapping, metric definitions, and source data.

Mistake 6: not tracking decisions

If actions do not have owners, due dates, and follow-up, the review will keep rediscovering the same issues.

How Ubisar would approach it

Ubisar would start with one merchandising review that already matters to the business: a category, a SKU set, a launch cohort, a promotion group, or a high-inventory range. We would map the current review, clean the product and source mapping, define the metrics, build the exception view, add an action tracker, and then layer in AI where it reduces analysis and commentary work.

The work usually touches all three layers:

• Data: SKUs, variants, categories, sales, margin, inventory, returns, traffic, campaigns, service issues, action status, and source freshness.
• Tech: ecommerce platform, POS, ERP, inventory system, PIM, finance system, BI, campaign tools, service tools, and workflow apps.
• AI: performance summaries, exception explanations, product grouping, review prompts, service-theme linking, commentary drafts, and action follow-up.

This connects directly to consumer and retail workflow implementation. It also fits the AI, Data & Tech Implementation Retainer, because merchandising gets better through repeated cycles: review the products, take action, see what changed, improve the data, and expand the workflow into adjacent decisions.

A checklist for your next merchandising review

Before building another dashboard, choose one product scope and answer these questions:

• Which products are in scope for the weekly review?
• Which IDs connect product, inventory, sales, margin, returns, and campaign data?
• Which metric definitions are currently unclear?
• Which products need decisions, not just reporting?
• Which products are winners with constraints?
• Which products are tying up stock?
• Which products are margin leaks?
• Which returns or service issues should be visible?
• Who owns the action after the meeting?
• What should change before next week's review?

If you can answer those questions, you can start improving merchandising decisions before rebuilding the whole retail data stack.

Sources and useful references

For retail metric context, Shopify's guides to inventory reporting, retail reports and KPIs, and sell-through rate are useful references for the kinds of sales and stock signals teams often start with. For assortment and category decision patterns, NielsenIQ's assortment optimization material shows why product mix, customer demand, and substitutability matter. For markdown decisions, McKinsey's article on markdown management is a useful reminder that analytics should inform merchandising decisions rather than replace them.

The practical next step is not to produce more product reports. It is to make one weekly product review easier to run, easier to trust, and easier to act on.