Merchandising and e-commerce: showing the right products at the right time on the website, with rules that follow stock and sales.
The e-commerce merchandiser decides what the customer sees: the order of products in categories, product highlighting, collections, campaign pages, sales and strikethrough prices. Much of their time goes on re-sorting hundreds of products by hand, removing those that have run out of sizes and checking prices before a campaign. Automation applies their rules continuously; they keep their eye on the shop window.
The role
Between product, e-commerce, marketing, pricing and replenishment.
The e-commerce merchandiser is the website’s salesperson: they stage the offer, choose the products to push, build collections and campaign pages, and adjust the presentation according to sales and stock. They prepare sales and promotions with pricing, pass replenishment needs to the replenishment team and share with product what sells and what does not. In a mid-sized company, they often also hold the website’s commercial calendar.
Day to day
A typical day.
- MorningRead sales, views and add-to-baskets by product and by category, and spot what is falling behind.
- MorningRe-sort categories: bring new arrivals and best sellers up, push products out of sizes down.
- MiddayPrepare homepage and campaign page highlights for the coming days.
- AfternoonBuild collections and selections: new arrivals, themes, gift ideas, products for a campaign.
- AfternoonPrepare sales and promotions with pricing: products involved, strikethrough prices, pages and banners.
- End of weekReview sales by category and request replenishment of fast-selling products.
The objectives
What the role is measured on.
- Conversion rate by category
- Sales of highlighted products
- Collection sell-through
- Share of visible products out of stock
- Campaign revenue
- Margin after markdown
Indicators are named, never quantified: levels depend on each company, and the diagnostic measures them.
Day to day
The problems that come up most.
Category sorting done by hand
Dragging and dropping products one by one, category by category, country by country: the sort is out of date as soon as it is finished.
Out-of-stock products at the top of the list
A product with only extreme sizes left stays on the first row and takes sales away from products that are available.
Sales prepared in a rush
Product lists, discount rates, reference prices, pages and banners are prepared in spreadsheets, with a risk of error on strikethrough prices.
Collections rebuilt every season
New arrivals, themes, selections: each collection is built by hand and empties as stockouts occur.
Little visibility on what customers search for
Searches with no results and the filters used are not read, although they show what is missing or badly named.
Replenishment requested too late
A product selling fast on the website is spotted at the end of the week, when some sizes have already run out.
By sector
One role, several realities.
- Fashion and luxury
Editorial presentation and staging of pieces, product order designed like a shop window, few discounts and reserved private sales, scarce stock not to be shown out of stock.
- Retail
Collections renewed every season, sizes and colours managed product by product, sales and strikethrough prices set on the same day online and in store, many short campaigns.
- Cosmetics and beauty
Launches and new products to push, seasonal gift sets, routines and complementary products to pair, limited editions to remove as soon as they sell out.
- B2B distribution and wholesale
Deep categories to sort by sales, margin and stock, many supplier promotions, on-site search and filters decisive for finding the right product.
What changes
Automate the task, keep the judgement.
What gets automated
- Category sorting by weighted rules: new arrivals, sales, margin, available stock by size, recalculated every day by country.
- Automatic demotion of products out of sizes and removal of highlights with no stock.
- Preparation of sales and strikethrough prices from a validated list, with a reference price check before publication.
- Dynamic collections built from PIM attributes and stock, which fill and empty on their own.
- Reading of zero-result searches and product performance, with suggestions drafted by an AI agent.
What stays human
- Staging choices, hero products and editorial tone.
- Sorting rules and their weighting, and the exceptions the brand wants.
- Products and discounts for sales and campaigns, with pricing and sales management.
Use cases
How to automate, case by case.
Each case starts from a question leadership teams ask. The starting point, what gets automated, the tools involved, and what stays in the teams’ hands.
How do you automate website merchandising?
- Today
- The merchandiser orders the products of each category by hand in the back office, country by country. The sort is done on Monday, and by Wednesday sold-out or rarely viewed products occupy the first rows.
- What gets automated
- A sort score is calculated every night for each product and each category, from weighted rules: newness, recent sales, product page conversion rate, margin, share of sizes available. The resulting order is pushed to the e-commerce platform by country, respecting the products pinned by the merchandiser. A report shows what moved and why.
- What stays human
- The merchandiser sets the weightings, pins hero products and corrects sorts that do not match the intended image.
- E-commerce platform
- PIM
- ERP
- Web analytics tool
How do you automate the demotion of products that are out of sizes?
- Today
- A product sold out in the core sizes stays at the top of the category with only extreme sizes left; the customer clicks, does not find their size and leaves. The merchandiser only notices when browsing the categories again.
- What gets automated
- For each product, the share of sizes or variants available to sell is calculated continuously, weighted by the usual demand for each size. Below a threshold set by the team, the product moves down the categories and out of highlights; it moves back up as soon as replenishment arrives.
- What stays human
- The team sets the threshold and decides to keep an iconic product visible even when incomplete.
- E-commerce platform
- OMS
- ERP
How do you automate setting up sales and strikethrough prices on the website?
- Today
- Before the sales, the merchandiser and pricing prepare in a spreadsheet the list of products and their discounts, load it into the platform the night before, then check product by product. A wrongly chosen reference price or a forgotten discount is discovered on opening morning.
- What gets automated
- The validated list of products and discounts is loaded once and checked automatically: product online, stock available, reference price compliant with the legal rule calculated from the price history, consistency with the store price. Strikethrough prices, badges and sales pages are published at the scheduled time, and original prices come back on their own at the end.
- What stays human
- The choice of products, discounts and successive markdowns stays with pricing and sales management.
- E-commerce platform
- ERP
- Store tills
- Spreadsheet
How do you automate the creation of collections and category pages?
- Today
- Each season, the merchandiser builds collections by hand (new arrivals, themes, gift ideas), adding products one by one. Collections empty as stockouts occur and new products do not get into them.
- What gets automated
- Collections are defined by criteria: PIM attributes (material, colour, use, line), arrival date, price range, available stock. They fill and empty on their own, and an alert warns when a published collection falls below a number of products set by the team. An AI agent suggests the page’s introductory text.
- What stays human
- The merchandiser defines the criteria, reviews the copy and adds the products that attributes alone cannot select.
- PIM
- E-commerce platform
- ERP
How do you automate the website’s highlighting calendar?
- Today
- Homepage banners, product blocks and campaign pages are scheduled by hand, the day before or the same morning. A highlighted product is sometimes out of stock, or a banner stays online after the campaign has ended.
- What gets automated
- Each highlight is described in a shared calendar: slot, dates, products, visuals, countries. It goes live and comes down on its own at the scheduled times, and a check the day before verifies the stock of the linked products, the presence of visuals and prices. Highlights whose products sell out along the way are flagged.
- What stays human
- The merchandiser and marketing choose the slots, the messages and the products highlighted.
- E-commerce platform
- DAM
- Project management tool
How do you automate conversion analysis by product and by category?
- Today
- Sales by product are known, but not the journey that precedes them: views, add-to-baskets, returns. The merchandiser does not know whether a product sells little because it is rarely seen or because it converts poorly.
- What gets automated
- For each product and each category, a dashboard brings together every day product page views, add-to-baskets, sales, return rate and stock. Highly viewed products that convert poorly, and those that convert well but get little exposure, are listed separately. An AI agent links low-conversion products to customer reviews and return reasons to suggest a cause.
- What stays human
- The merchandiser decides to revise a product’s visuals, page, price or exposure, with the catalogue team and pricing.
- Web analytics tool
- E-commerce platform
- OMS
- Reporting tool (BI)
How do you automate complementary product recommendations?
- Today
- “You may also like” or “complete the look” blocks are filled by hand for a few hero products, or left to default rules that suggest sold-out or unrelated products.
- What gets automated
- Pairings are calculated from products bought together, PIM attributes and the brand’s rules (same line, compatible colours, skincare routine). Suggested products must be in stock in the common sizes, and the merchandiser can force or forbid pairings. Recommendations are measured by the additional sales they generate.
- What stays human
- The merchandiser validates the pairing rules and corrects those that do not respect the brand’s world.
- E-commerce platform
- PIM
- OMS
- Web analytics tool
How do you automate the detection of products to sell through on the website?
- Today
- At the end of the season, the merchandiser spots slow products by eye, cross-checking sales and stock exports. Markdown or exposure decisions come late, when the remaining stock is hard to sell.
- What gets automated
- Every week, a calculation compares for each product the remaining stock, the recent rate of sale and the time left before the end of the season. Products whose cover exceeds the selling horizon are listed with their views, conversion and current exposure, to distinguish those lacking visibility from those needing a discount.
- What stays human
- The choice between more exposure, a discount, a transfer to stores or a move to another channel stays with the merchandiser and pricing.
- ERP
- E-commerce platform
- Web analytics tool
- Reporting tool (BI)
How do you automate tracking of zero-result searches on the website?
- Today
- The on-site search engine records what customers type, but nobody reads the data. Products exist under another name, common typos find nothing and repeated requests go unnoticed.
- What gets automated
- Every week, searches with no results and those that lead to no click are grouped by intent. An AI agent suggests for each one a synonym to add, an existing product to link or a gap in the assortment, and validated corrections are pushed to the search engine.
- What stays human
- The merchandiser validates the synonyms and passes to product the requests that reveal a gap in the offer.
- Website search engine
- Web analytics tool
- PIM
How do you automate website replenishment requests to the replenishment team?
- Today
- A product is selling fast on the website; the merchandiser notices at the end of the week and e-mails the replenishment team with a screenshot. The core sizes have already sold out when the request is handled.
- What gets automated
- The online rate of sale of each product is compared every day with its available-to-sell stock, size by size. When cover falls below the set threshold, a replenishment request goes to the replenishment team with recent sales, the sizes concerned and the store stock that could be transferred.
- What stays human
- The replenishment team decides whether to replenish, transfer store stock or let the product sell out.
- ERP
- OMS
- E-commerce platform
- Messaging
Questions from the teams
What teams look to improve.
- How do you automate product sorting in the categories of an e-commerce website?
- By writing down the rules the merchandiser already applies: weight of new arrivals, recent sales, margin and available stock by size. The sort is recalculated every day and by country, and the merchandiser pins by hand the products they want to keep at the top.
- How do you avoid showing out-of-stock products at the top of categories?
- By calculating for each product the share of sizes or variants still available, and moving it down below a threshold set by the team. The product stays online but no longer takes the place of a sellable product.
- How do you automate strikethrough prices during sales?
- By starting from a validated list of products and discounts, then calculating the reference price according to the legal rule (in principle the lowest price charged over the thirty days preceding the reduction) from the price history. A check blocks the display of a strikethrough price that does not comply with the rule.
- What is a dynamic collection on an e-commerce website?
- A selection of products defined by criteria rather than by a list: PIM attributes, arrival date, stock, price. It updates itself when a product enters or leaves the criteria, which avoids rebuilding collections every season.
- Can AI do a website’s merchandising?
- It can sort, recommend complementary products and spot underperforming products. Brand image, hero products and editorial stay the merchandiser’s choices.
- How do you know which products to highlight?
- By combining sales, product page conversion, available stock, margin and newness. A dashboard ranks the products that deserve more visibility and those that get too much for what they sell.
- How do you make use of on-site searches?
- By reading every week the searches with no results and those that lead to no click, then adding synonyms, correcting labels or flagging a gap in the assortment.
- Do you need a dedicated merchandising tool?
- Not always. Many platforms accept a sort order supplied by API or file; automation calculates this order and pushes it. A dedicated tool is justified when personalisation by customer becomes central.
Related roles
02 — Getting started
Describe what’s holding your teams back.
A first conversation to qualify your need. Then a scoped diagnostic: the scope, the solution, the estimate.
Discuss your projectReply within 48 business hours.