Best Merchandising for Shopify Plus Fashion Brands

Key Takeaways

Shopify Plus gives fashion brands 9 fixed sort options with no access to margin, sell-through rate, or custom formulas. At 5,000+ SKUs, that is not merchandising. It is alphabetical with extra steps.
The three capabilities that separate real fashion merchandising from basic sorting: margin-aware ranking formulas, scheduled seasonal rules, and visual discovery (camera search, shop-the-look, visual similarity).
The best merchandising platform is the one your merchandising team can operate without filing engineering tickets.

Why do fashion brands outgrow Shopify's native merchandising?

Shopify Plus ships nine sort options for collection pages, and none of them include margin, sell-through rate, inventory velocity, or return rate.

For fashion brands running seasonal drops with 5,000+ SKUs, those missing signals mean best-selling sort buries new arrivals, manual sort burns thousands of hours per year, and high-margin styles get outranked by markdown volume. Here is the complete list from the CollectionSortOrder enum:

Alphabetical A–Z
Alphabetical Z–A
Best Selling
Date Created (oldest first)
Date Created (newest first)
Manual
Most Relevant
Price (low to high)
Price (high to low)

That is it. Nine options, all fixed. No margin, no sell-through rate, no inventory days, no return rate. No newness velocity, no seasonal weighting, no brand diversity controls. For a detailed comparison of when Shopify's built-in tools stop being enough, see our Shopify search threshold guide.

The "best-selling" trap. Best-selling sort creates a self-reinforcing loop. Baymard Institute has documented this across hundreds of large ecommerce sites: products at the top of a collection get disproportionately more impressions, clicks, and purchases, and those purchases push them higher still. New arrivals, high-margin styles, and seasonal launches get buried under their own lack of history.

For a basics-heavy store selling t-shirts year-round, that loop is fine. For a fashion brand dropping new styles every two weeks, it buries the inventory you most need to move.

Worse, best-selling does not distinguish between full-price sales and markdown sales, which means a product that moved 200 units at 60% off during last month's clearance event outranks a product that moved 50 units at full margin this week. The sort rewards volume, not profitability.

The deeper problem: most fashion teams know best-selling sort is broken and use it anyway because manual curation costs more time than they have.

So the inventory you worked hardest to source, the new arrivals and high-margin pieces, sits below the fold behind last season's markdown volume.

The brands that break this cycle are not the ones that hired more merchandisers or added another manual review day to the calendar. They replaced the sort logic entirely, trading drag-and-drop for formulas that blend margin, velocity, and newness into a single weighted score.

The manual workaround. Without custom ranking, the alternative is Manual sort: drag-and-drop positioning for every product in every collection, every week. At 200 collections with 50+ products each, that is a full-time job for a human who could be doing actual merchandising strategy.

"Layers was the first time we were able to create the kind of sort orders we were used to having in Salesforce. We saw a pretty immediate impact on conversion rate."

David Cost, VP of eCommerce, Rainbow Shops

David's team ran into the same wall most fashion operators hit: they had the merchandising instincts from enterprise retail, the kind of multi-signal sort logic they built in Salesforce for years, and Shopify gave them nine fixed levers where they needed dozens.

This article is an evaluation framework for fashion merchandisers who have outgrown those nine sort options and need what a dedicated merchandising platform provides. If you need a system that handles newness velocity, margin contribution, sell-through rate, and return rate, here is how to evaluate your options.

What makes fashion merchandising different from general ecommerce merchandising?

Fashion is not electronics or supplements. The merchandising job is fundamentally different in five ways. Seasonal velocity, visual-first discovery, outfit merchandising, return rates as ranking signals, and brand diversity all shape how a collection page needs to behave. General ecommerce tools built for stable catalogs with two-year product lifecycles break under the speed and visual complexity of fashion.

1. Seasonal velocity. Fashion products expire. A puffer jacket that lands in October needs to be at the top of its collection within the first two weeks of sell-through data, not after three months of accumulating "best-selling" rank. Newness is a ranking signal, not a nice-to-have.

2. Visual-first discovery. Shoppers cannot describe what they want in words. NNGroup's research on product page UX confirms that categories driven by aesthetics require image-heavy browse experiences. "Blue floral midi dress with a square neckline" returns wildly different results depending on the search engine. They need to show what they want, whether that is a screenshot from Instagram, a photo from a magazine, or a picture of something they saw on the street.

3. Outfit and look merchandising. Fashion shoppers buy outfits, not isolated items. A collection page that shows 12 black dresses in a row fails to cross-sell. The page needs to surface complementary pieces (the jacket that goes with the dress, the bag that finishes the look) alongside the hero product.

4. Return rate as a ranking signal. Fashion has 20-30%+ return rates, higher than almost any other vertical. A dress that converts at 8% but gets returned 40% of the time is worse for your business than a dress that converts at 5% with a 10% return rate. If your merchandising platform cannot factor in returns, you are optimizing for the wrong number.

5. Brand and style diversity. A multi-brand retailer showing 12 items from the same vendor at the top of a collection page has failed at the most basic job of assortment representation, and the shopper who came to browse the full catalog will bounce before scrolling past the first fold. Trust erodes fast. The page needs to represent the breadth of your assortment, not the depth of one supplier's catalog.

General ecommerce can get away with ignoring most of these. Fashion cannot, because every week that passes without the right products in the right positions is a week of margin lost to a catalog that moves faster than your merchandising tools can keep up.

If your merchandising tool was designed for electronics catalogs where products have a two-year shelf life and text search covers 95% of discovery, it will break under the speed and visual complexity of fashion.

What are the eight capabilities fashion merchandisers should evaluate?

Not every merchandising platform is built for fashion. Here are the eight capabilities that determine whether your team can operate at the speed fashion demands: custom ranking formulas, computed attributes, scheduled rules, visual search, shop-the-look, diversity expressions, live preview, and real-time collaboration. Each one maps to a specific failure mode in Shopify's native merchandising.

1. Custom ranking formulas with business metrics

The most important capability is the one Shopify does not offer natively: the ability to build sort orders that blend multiple business signals into a single weighted score.

You should be able to combine signals like:

Gross margin or contribution margin
Sell-through rate over a rolling window
Units sold per day
Inventory days remaining
Full-price sell-through versus markdown sell-through

We support this through weighted attribute groups and priority rules. You assign relative weights to each signal, and promoted products always appear at the top as hard overrides. No code. No engineering tickets. The merchandising team controls the formula.

Rainbow Shops uses this approach to build what they call "What's Hot" sort orders: rankings that blend velocity and margin data, updated continuously, across hundreds of collections.

The sort adapts as new sales data flows in, so a product that starts trending on Tuesday is already climbing by Wednesday morning.

The difference from Shopify's native "Best Selling" sort: you define what "best" means for your business. It might be gross margin weighted at 40%, sell-through rate at 35%, and recency at 25%. You set the formula. The ranking executes it.

2. Computed attributes for derived metrics

Raw product data rarely contains the exact signals you need, which is why most merchandising teams end up exporting to spreadsheets, calculating discount depth or days-since-arrival by hand, and re-importing the numbers into a system that should have derived them automatically.

Discount percentage, days since arrival, margin per unit, inventory coverage ratio: these are all derived from existing fields.

We ship five built-in computed attributes out of the box:

SKU Coverage: variant availability ratio
Days Available: duration since product publication
Has Image: featured image presence check
Age of Newest Variant: most recent variant creation date
Price Varies: multi-variant pricing flag

Beyond the defaults, you can build custom computed attributes using JavaScript functions or the visual Derive Builder. Map product tags to seasonal classifications, calculate discount depth from compare-at price versus current price, or derive an inventory health score.

Use them anywhere. These computed values become sortable, filterable, and available in merchandising rules the moment you save them.

A derived metric like "discount depth greater than 30% AND inventory coverage below 40%" can trigger an automatic demotion rule without you ever opening a spreadsheet.

3. Scheduled merchandising with time windows

Fashion merchandising is seasonal by nature. Your "New Arrivals" sort should behave differently during a launch week than during a clearance period. Your homepage collection should rotate automatically for a flash sale.

We support scheduled merchandising rules that activate and deactivate on a time window you define. Set a rule to boost a new collection starting Monday at 9 AM and revert to your standard ranking on Friday at midnight. No manual intervention, no forgotten reverts, no engineer needed at midnight.

4. Visual search and camera discovery

This is where fashion merchandising diverges from every other vertical, because your shoppers think in images, not keywords, and no amount of synonym tuning will close the gap between what they see in their head and what they can type.

We offer two APIs that power visual discovery:

Image Search API: shoppers upload a photo (or take one with their camera), and we return visually similar products from your catalog. The API accepts image uploads or base64-encoded payloads and supports full filtering, faceting, and personalization on top of visual similarity.
Similar Products API: given any product in your catalog, we return the most visually and semantically similar items. Results are precomputed and cached daily for speed, with automatic real-time fallback for newly added products.

These APIs power "shop the look" experiences, "more like this" carousels, and camera search on mobile. For a fashion brand, visual discovery is not a nice-to-have; it is a category requirement, because the shopper who screenshots a look from Instagram and cannot search your catalog with that image will find a store that lets them.

If you want to go deeper on the technology, implementation, and conversion data behind visual search, we wrote a full guide: AI Visual Search in Ecommerce: How It Works and Why It Converts.

If visual discovery is the capability gap you are trying to close, book a walkthrough and we will show you how camera search and similar-products work on your actual catalog.

5. Shop-the-look and product grouping

Fashion shoppers buy outfits. A single product page for a blazer should surface the matching trousers, the shirt underneath, and the belt that completes the look.

The Similar Products API makes this possible by combining visual and semantic signals. You can filter by product type (show only accessories when the hero product is a dress) and apply your merchandising rules on top. The result is an outfit grid, not a random "you might also like" carousel.

6. Diversity expressions

A collection page showing six near-identical black dresses in a row is a page failure. Shoppers need variety to stay engaged.

We handle this with diversity expressions that cap how many products from the same group appear in the top window of results. You configure up to five diversity axes (product family, vendor, product type, or any attribute), set a maximum per group, and define the window size.

For example: "In the top 20 positions of this collection, show no more than 2 items from the same brand and ensure at least 10 different styles are represented." The sort order applies normally beyond that window. Pins and priority rules still take precedence within it.

7. Live preview on your actual storefront

Merchandising changes are high stakes, because a bad sort order on your homepage collection during a sale weekend, the kind of mistake that takes two minutes to make and eight hours to notice, can cost tens of thousands in revenue.

We provide preview functionality that lets you see your merchandising rules on your live storefront before publishing. Sort effect annotations show which rule affected each product's position, so you can debug before anything goes live. Draft mode lets you save work-in-progress configurations without affecting the storefront.

8. Real-time collaboration with live cursors and field locking

Fashion merchandising is a team sport. The buyer, the visual merchandiser, and the ecommerce manager all have opinions about collection order. They should not have to take turns.

We support real-time collaboration with:

Live cursors: see exactly where your teammates are working on the page
Field locking: automatic locks prevent two people from editing the same rule simultaneously, with colored borders showing who has the lock
Drag locking: products lock during reordering so two people cannot rearrange the same collection at the same time
Threaded comments: leave comments on any merchandising rule or sort order, with threaded replies and notification badges

This is not a nice-to-have. When your team is scrambling to set up 50 collections for a seasonal launch, the ability to work simultaneously without overwriting each other is the difference between launching on time and launching late.

How does automation change the economics of fashion merchandising?

Here is the math on manual merchandising: a mid-size fashion brand manages 200 collections, each needing a weekly sort review to account for new arrivals, sell-through changes, and inventory shifts. Each review takes 15–30 minutes of drag-and-drop work.

That is 200 collections multiplied by 20 minutes per review, multiplied by 52 weeks per year, which comes to roughly 3,500 hours annually. More than 1.5 full-time employees doing nothing but dragging products up and down a list.

Automated sort formulas eliminate the drag-and-drop entirely. Rainbow Shops replaced their manual process with weighted sort orders that blend sell-through velocity and margin data. The sort updates continuously as new data comes in, which means a product that starts trending at 10 AM on a Tuesday is already climbing the collection by noon without anyone opening the admin panel or dragging a single product card. No weekly ritual. No bottleneck.

Scheduled rules eliminate the midnight deploy. Flash sale starts at midnight? The boosting rule activates automatically. Sale ends Sunday? The rule deactivates and normal ranking resumes. No one needs to be online.

Diversity constraints eliminate page failures. Instead of manually checking that your "Dresses" collection does not show six items from the same vendor in a row, a diversity expression enforces the constraint automatically. Every page load, every time.

The ROI goes beyond time saved. Data-driven sorts outperform manual reordering because they respond to signals faster than a human can. By the time a merchandiser notices a product is trending, the formula has already lifted it.

There is also the cost of errors. A manual sort that forgets to demote an out-of-season product, or that accidentally buries a new arrival behind 40 carryover items, directly costs you revenue.

Automated formulas do not forget or get distracted; they process every signal for every product on every page load.

What questions should a fashion merchandiser ask during a platform evaluation?

Seven questions separate a fashion-ready merchandising platform from one that was designed for general ecommerce and adapted for fashion after the fact. For each question below, you will see what a strong answer looks like and what a weak answer reveals about the platform's architecture.

1. Can I sort by margin, sell-through, and return rate? Yes looks like: you build a weighted formula from imported business metrics, no code required. No reveals: the platform sorts by clicks and conversions only, and your high-return, high-discount products will dominate every collection.

2. Can I schedule merchandising rules with start and end times? Yes looks like: a calendar interface where you set activation windows, and rules revert automatically. No reveals: someone on your team is setting a phone alarm to manually unpublish a sale-weekend boost at midnight on Sunday.

3. Does the platform support visual search and camera discovery? Yes looks like: an image search API that returns products ranked by visual similarity, with filtering and personalization. No reveals: your shoppers are stuck typing "blue floral midi dress" and hoping for the best.

4. Can I create computed attributes from existing product data? Yes looks like: a function editor or visual mapper that derives discount percentage, days since arrival, or margin from raw fields, and makes the result sortable. No reveals: you are exporting to a spreadsheet, calculating the metric, and re-importing it. Weekly.

5. Does the platform enforce diversity across a collection page? Yes looks like: diversity expressions that cap repeats per brand, per product family, or per attribute within a defined window. No reveals: you are manually scanning every collection page to check that one vendor is not dominating the top 20.

6. Can I preview changes on my live storefront before publishing? Yes looks like: a preview mode that renders your actual store with the proposed rules applied, with annotations showing which rule moved which product. No reveals: you publish to production and hope for the best.

7. Can multiple team members edit merchandising rules simultaneously? Yes looks like: live cursors, field locking, and threaded comments so two merchandisers can work on the same sort order without overwriting each other. No reveals: your team takes turns, or worse, someone's 30 minutes of work gets silently overwritten.

Does the platform find the strategy, or just execute it?

The hardest problem in fashion merchandising is not executing a sort order you already defined. It is figuring out what the right sort order should be in the first place.

There is a difference. A platform that only moves products around is a tool. A platform that shows you which products should move, why, and whether the change worked, is an operating system for merchandising decisions.

"Layers doesn't just merchandise, it also finds the strategy and provides unparalleled control, all while saving us hours we didn't know we were losing. Not to mention, we've seen a massive lift in key KPIs."

Brittany Csik, eCommerce Manager, Negative Underwear

Brittany's distinction is the one that matters most. Basic merchandising tools execute a strategy you already have. You decide the sort order, the platform applies it. That works when you know exactly what your collections should look like, but fashion moves too fast for any merchandiser to hold the optimal sort order for 200 collections in their head simultaneously, especially when sell-through data, return rates, and inventory levels shift daily.

The harder problem is figuring out what the right strategy is in the first place.

That requires analytics. LayersQL, our domain-specific query language for merchandising analytics, lets you query sales, views, and conversions across any time window, segment by geography or marketing channel, compare periods, and visualize the results.

The same metrics that power your sort orders power your analysis of whether those sort orders are working. If you want to benchmark your current search and autocomplete performance, start there before evaluating platforms.

The loop closes when your merchandising decisions and your merchandising measurement live in the same system. You build a sort formula, measure its impact in LayersQL, adjust the weights, and measure again.

Detailed performance analytics track impressions, click-through rate, add-to-cart rate, purchase rate, revenue, and revenue per impression across every merchandising rule. Trend visualization over 7, 14, and 30 days shows whether a rule is gaining or losing effectiveness.

That is the difference. One moves products. The other tells you which ones should move.

Why is fashion merchandising a revenue strategy, not a grid layout?

The brands that win in fashion ecommerce are not the ones with the best photography or the biggest ad budgets. They are the ones whose merchandising adapts faster than their competitors', so when a style starts trending on social media on Tuesday morning, the sort formula has already lifted it by Wednesday without anyone opening a spreadsheet.

Seasonal velocity and margin-aware ranking. Visual discovery and diversity constraints. Automated scheduling, real-time collaboration, and analytics that close the loop.

These are not features on a checklist; they are the operating system for fashion merchandising at scale.

If your current setup involves dragging products in a grid, exporting to spreadsheets, and hoping your best-selling sort is not burying your newest, highest-margin inventory, it is time to evaluate what a purpose-built merchandising platform paired with AI search can do.

Start with the eight capabilities above. Ask the seven questions. And test whether the platform finds the strategy, not just executes it.

For a broader evaluation of AI search capabilities at $10M+, see our AI search evaluation guide. If you sell across multiple Shopify Markets, our multi-market discovery guide covers the additional criteria for international expansion.