Shopify Collection Sorting for Fashion Stores: Size Breaks, Seasons, Drops and Returns

Shopify collection sorting for fashion stores has to handle broken size runs, season changeovers, colour variants, drops, sell-through and returns. A practical playbook, collection by collection.

Shopify collection sorting for fashion stores is harder than for almost any other category, and most generic advice ignores why. A fashion product is really a grid of sizes and colours, and it can look healthy at the product level while being unsellable to most visitors. Demand swings with the season, the weather and whatever was worn on someone's feed last week. Launches arrive as drops rather than a steady trickle. And a meaningful share of what sells comes back. A sort that works for a candle shop falls apart on a clothing catalogue within a fortnight.

This guide is written for apparel, footwear and accessories merchants who have outgrown 'Best selling' and a hand-dragged manual order. It covers the six problems that make fashion merchandising on Shopify different, size breaks, seasonality, colour variants, drops, sell-through and returns, and then gives collection-by-collection recipes you can copy. Along the way we say exactly what SellRank, our Shopify collection sort app, does about each problem, and just as plainly where it needs your judgement instead. Worked numbers are illustrative, not data from a specific store.

Why fashion breaks generic Shopify product sorting

Most sorting logic assumes a product is one thing with one stock level and one demand curve. In fashion, none of those hold. A dress in five sizes and three colours is fifteen variants, each with its own stock, and the shopper only cares about one of them. Its demand curve has a launch spike, a full-price plateau that lasts weeks rather than months, and a markdown tail. Its sales figure includes items that will be returned. And it sits in a catalogue that is replaced, in large part, every season.

The consequence is that the numbers a generic tool ranks by, lifetime sales, total stock, raw orders, give a misleading picture of what should lead a fashion collection. Good fashion merchandising on Shopify is about using the right signal for each job, and having a human override for the things no signal captures.

Problem 1: Size breaks, the silent conversion killer

A size break, or broken size run, is when a style still has stock but has sold out of its core sizes. The jeans are 'in stock' because two pairs of size 24 and one pair of size 34 remain. To almost every visitor who taps them, they are sold out. Product-level stock logic cannot see this: the product is available, so it keeps its place in the first row, and it converts like a sold-out product.

Size breaks are also self-reinforcing on sales-based sorts. The style that sold best is the one whose middle sizes went first, so it is the one most likely to be broken, so the sort keeps it at the top because of the sales it made before it broke.

How to handle size breaks in your collection order

There are three layers of defence, and the first one is automatic if you rank on the right metric.

• Rank the top of the page on conversion rate, not orders. When core sizes sell out, views keep coming but orders stop, so conversion falls and the product slides down on its own. This is the single most useful change a fashion store can make, because it responds to size breaks without needing to know about sizes at all.
• Tag broken styles and exclude the tag from your top slots. If your inventory team, a Shopify Flow or a stock report can flag styles with core sizes gone, a tag such as 'broken-sizes' lets any slot condition keep them out of the first rows while leaving them findable further down.
• Demote by hand for the styles you know about. When a buyer says the core sizes of a hero style will not be back for three weeks, demoting it takes a second and survives every re-sort until you remove it.

To be clear about what SellRank does and does not do here: it ranks products, not individual sizes, and it does not compute a size-break score by itself. What it gives you is the tools that make size breaks manageable: a Conversion % strategy powered by its own first-party pixel, so views are counted even for products that stop selling; tag conditions on every slot; and demotions that are remembered on every re-sync. In practice, conversion-led ranking catches most size breaks within days, and tags or demotions catch the rest.

Problem 2: Seasonality and the changeover weeks

Fashion demand is seasonal in two ways. There is the calendar season, when coats sell in autumn and linen in spring, and there is the transition, the few weeks when weather and wardrobes change faster than your data. The transitional weeks are where most collection pages go wrong, because every performance-based ranking is looking backwards.

Illustrative scenario: it is the last week of September for a northern-hemisphere store. The past thirty days of revenue are dominated by summer dresses, still selling on late-summer weekends. The first cold snap arrives on Thursday. A revenue ranking over thirty days keeps summer dresses at the top of 'Dresses' for another two weeks, while shoppers arriving from an autumn campaign scroll past sundresses looking for long sleeves.

Merchandising the season change without guessing

• Plan the changeover as a scheduled draft. Build the autumn order days in advance, with a slot for the new season's styles near the top, and give it a go-live date that matches your campaign. It publishes itself at the right moment.
• Use conditions, not just rankings. A slot that says 'first 25%: in-stock products tagged autumn, ranked by conversion' puts the season forward immediately while still ranking within it by evidence.
• Lean on engagement metrics in transition weeks. Views and add-to-carts move before orders do. A Trending-style Balanced score that weights recent engagement will pick up the shift sooner than revenue alone.
• Let data take over once the season settles. After two or three weeks the new season has its own sales history, and the forced slot can shrink.

SellRank supports each of these: drafts with scheduled go-live and automatic revert, tag conditions on slots, and the Balanced score with weighted blends such as Trending. Drafts and scheduling are on Grow and above. Because the previous order is kept as a version, reversing a changeover that came too early is a one-click restore, not an evening of dragging.

Problem 3: Colour variants and the first-row monoculture

Fashion stores structure colours in one of two ways. Either each style is one Shopify product with colour as a variant, or each colourway is its own product, often grouped visually with swatches. Both have merchandising consequences.

When colour is a variant, the product earns its position on the combined performance of all colours. The navy may be doing all the work while the tile shows the mustard, and the shopper judges the style by the colour on the card. When each colour is a separate product, each earns its own position, which is fairer, but a successful style can then occupy half the first row in four colours. The page looks like it only sells one thing, and the shopper who does not like that shape has nothing else to look at above the fold.

Keeping colour variety in the first row

• If colours are separate products, decide how many of one style you want above the fold. Two colourways of a hero style is merchandising; five is a monoculture.
• Use demotions to thin out duplicates in the first row. Keep the best-converting colourway at the top, demote the others a few rows, and let them earn their way back.
• Use slots to force variety of product type. A first slot limited to a product type, say tops, followed by a slot limited to dresses, guarantees a mixed first screen on a broad 'New In' collection.
• If colours are variants, make sure the featured image on the card is the best-selling colour. No sorting tool fixes a tile that shows the wrong colour.

SellRank does not automatically de-duplicate colourways of the same style, and we would be wary of any tool that claims to guess which products are 'the same' reliably across every catalogue structure. What it does give you is product type, vendor, tag and price conditions on each slot, plus pins, demotions and exact placement in the preview, so you can see the first screen exactly as the shopper will and fix a monoculture before it goes live.

Problem 4: Drops, launches and the first 72 hours

Many fashion brands sell in drops: a capsule goes live at a set time, the audience has been primed for days, and most of the sales happen in the first few days. A drop breaks performance-based ranking completely at launch, because the new products have no history, and it breaks newness-based ranking a week later, when the drop is half sold out and still leading every page.

A drop-day sorting plan that runs itself

• Before the drop: build the launch order as a draft. Pin the hero pieces of the drop to the top of 'New In' and of their category collections, in the order the campaign features them.
• At the drop: schedule the draft to go live at the exact launch time, so the collection pages match the email and social posts the moment they land. Nobody has to be awake at midnight.
• In the first 48 to 72 hours: watch conversion and add-to-cart rates, not totals. The drop pieces with strong rates deserve to stay up; the ones with weak rates should come down early.
• After the first wave: remove the pins and let the drop compete. A freshness slot keeps the drop visible without letting it squat on the first row, and conversion-led ranking decides which pieces keep their places.
• At the end: schedule the revert. If the drop was a limited event, the previous order comes back automatically on the date you set.

Every step here maps to something SellRank does: pins that survive re-syncs, drafts with scheduled publishes and automatic revert, Newness and Conversion % strategies for the freshness slot and the main ranking, and per-version performance so the drop order can be compared with the order that preceded it. On a drop, the difference between the page matching the campaign at 9:00 and matching it at 9:40 is real money, and it is the kind of job that should never depend on someone remembering.

Problem 5: Sell-through, the number fashion actually runs on

Fashion buyers live by sell-through: the share of what you bought that you sold, usually tracked at full price over a season. It is the metric that decides reorders, markdowns and next season's buy. Oddly, it is almost never used to order collection pages, even though the collection page is the main lever a store has on it.

Sell-through matters for merchandising in both directions. A style with high sell-through and plenty of stock left is a winner you should be pushing harder. A style with low sell-through deep into the season is heading for markdown, and the cheapest intervention is exposure: moving it up the page at full price while there is still time for it to sell.

Using sell-through and days of stock cover to order fashion collections

Two related metrics do most of the work. Sell-through compares units sold against units in stock. Days of stock cover estimates how long current stock will last at the recent rate of sale. Together they tell you which styles are genuinely at risk, a high cover and low sell-through, and which are about to run out, a low cover, and should stop being pushed hard before they break on sizes.

• On a category page mid-season, reserve a modest slot, around 10 to 15% of the page, for in-stock styles with the lowest sell-through. It gives them a full-price chance before markdown.
• Keep the first row on conversion. Pushing slow sellers into the very first positions costs more in lost conversion than it saves in markdown.
• Watch stock cover on your winners. A best-seller with a week of cover and no reorder coming should start moving down before it breaks, not after.
• Hand your end-of-season clearance decisions to the data: styles still at low sell-through after a full-price exposure test are the ones that genuinely need a discount.

In SellRank, Sell-through and days of stock cover are ranking strategies on every plan, including Free. On Scale and Max, profit-aware ranking adds Margin % and estimated Profit, so a clearance slot can favour the slow movers where a full-price sale is worth the most. We go into the full clearance sequence in our post on sell-through, days of stock cover and clearance sorting.

Problem 6: Returns and the bestseller that is not

Fashion has high return rates compared with most e-commerce categories, and returns are uneven. A style with a fit problem can sell brilliantly and come back at several times the rate of its neighbours. On an orders or revenue ranking, it looks like a hero. In your warehouse and your margin, it is one of your worst products.

Illustrative example: two dresses each sell 100 units in a month at $80. Dress A has a return rate in line with the category; Dress B runs small and a large share of it comes back. Gross revenue ranks them level. Kept revenue, after returns, is not close, and Dress B also carries the cost of return shipping, handling and restocking items that may no longer be sellable at full price.

Keeping high-return styles out of the first row

• Review return rates by style from your Shopify reports or returns platform at least monthly, and flag styles that are well above the category average.
• Tag the chronic returners, for example 'fit-issue', and exclude that tag from the top slots of your collections. They stay findable; they stop being promoted.
• Use ratings as a partial signal. Fit complaints tend to show up in reviews, and a Customer favourites blend that weights review rating will lean away from styles that disappoint in hand.
• Fix the product page too. A size note or a better size guide often lowers returns more than any ranking change, and a fixed style can have its tag removed.

Be clear on the boundary here: SellRank ranks on views, add-to-carts, orders, revenue, conversion, profit, sell-through, stock, newness, discount and ratings. It does not have a return-rate strategy, so returns are handled through tags, the Ratings signal, which SellRank reads from the standard review metafields that the main review apps write, and demotions. That is honest and it works; the return data lives with your returns process, and a tag is the clean bridge between it and your collection order.

Shopify collection sorting for fashion: recipes, collection by collection

Here are slot recipes for the collections most fashion stores run. Treat them as starting points to preview against your own catalogue, not as rules. Percentages scale with the collection size.

Recipe: 'New In'

• Slot 1, first 30%: in-stock products ranked by Newness. The page is about recency, so recency leads.
• Slot 2, next 30%: in-stock products ranked by a Trending blend, so recent arrivals that are already converting stay visible as newer ones push in.
• Everything else: Newness, with out-of-stock products parked at the end.
• Human layer: pin the current drop's hero pieces for their launch week, then remove the pins.

Recipe: core category pages such as 'Dresses' or 'Knitwear'

• Slot 1, first 20%: in-stock products, excluding the 'broken-sizes' and 'fit-issue' tags, ranked by Conversion %.
• Slot 2, next 15%: in-stock products ranked by Newness, the freshness slot.
• Slot 3, next 10%: in-stock products ranked by lowest sell-through, the full-price exposure slot for slow movers.
• Everything else: Revenue, with out-of-stock products parked at the end.

Recipe: 'Sale' or the Shopify clearance collection

• Slot 1, first 25%: in-stock products ranked by Conversion %, so the first screen sells.
• Slot 2, next 25%: in-stock products ranked by days of stock cover, highest first, so the deepest overstock gets strong placement.
• Everything else: Discount, deepest first, which is what sale shoppers expect when they scroll.
• Schedule the whole order to publish at the start of the sale and revert at the end.

Recipe: 'Bestsellers'

• One strategy is enough here: Revenue or Orders over recent activity, in-stock only in the first 40%.
• Add a Conversion % blend if your traffic is uneven across products, so a product that got a burst of paid traffic does not dominate on volume alone.
• Demote anything with a known size break or fit issue. A bestsellers page is where broken hero styles do the most damage.

Footwear, accessories and basics behave differently

Not every part of a fashion catalogue behaves like seasonal ready-to-wear, and the differences change how you should sort.

• Footwear has the most severe size breaks of any category, because half sizes and widths multiply the grid. Conversion-led ranking matters even more here, and a 'broken-sizes' tag is worth automating early.
• Accessories, such as bags, belts, scarves and jewellery, are usually one-size, so product-level stock is accurate and in-stock conditions work perfectly. They are also often higher margin, which makes them strong candidates for a profit-weighted slot on Scale and up.
• Basics and core lines, the plain tee, the everyday sock, sell steadily all year. They are the backbone of revenue but make a boring first row. A common pattern is to rank them by revenue in the body of the page and keep the first screen for newer, more visual pieces.
• Occasionwear, such as wedding guest dresses or party season edits, has sharp, predictable peaks. It is the clearest case for scheduled drafts: build the occasion order weeks ahead and let it publish itself.

A worked example: three weeks of a 'Dresses' page

Here is an illustrative walk-through of how the core category recipe above behaves over time on a 120-product 'Dresses' collection. The numbers are made up to show the mechanics, not measured.

Week one. The first 20%, 24 products, is in-stock dresses ranked by conversion, excluding the broken-sizes and fit-issue tags. A floral midi that sold 60 units last month but has lost its sizes 8 to 12 is outside the first slot: its views are still high, its orders have almost stopped, and its conversion has fallen accordingly. Its place is taken by a slip dress that gets a quarter of the views but converts well. The next 18 products are the freshness slot, so the eight dresses uploaded that week are all above the second scroll. The slow-mover slot surfaces twelve styles that had been sitting below position 80.

Week two. The floral midi is restocked in its core sizes. The stock team removes the tag, conversion recovers as the sizes return, and the dress climbs back into the first slot within days without anyone touching the collection. Two of the eight new dresses have strong conversion and are now ranking inside the first slot on merit. The others have rolled into the body of the page as newer arrivals took their freshness positions.

Week three. The performance view shows the version running since week one against the revenue-sorted version that preceded it, over the same number of days. The merchandiser compares conversion from view to order and revenue per view, rather than total revenue, because a campaign ran during week two. The comparison either justifies the new order or says clearly that it made no difference, and both answers are worth having. In either case, nobody spent an evening re-dragging 120 dresses.

Fashion sorting mistakes worth avoiding

• Leading the whole store with newness. New In is a collection, not a strategy for every page.
• Letting last season's best seller squat on the first row because its lifetime numbers are big.
• Pushing slow movers into the first row to clear them. Give them a slot lower down; the first row has to convert.
• Changing the order during a drop because the first hour was quiet. Drops often build through the day as emails and posts land.
• Forgetting that a pin is permanent until you remove it. Put a reminder on every launch pin, or pin only in a scheduled draft that reverts.
• Sorting sale pages by discount alone, so the first screen is the deepest cuts on the least desirable stock.

What the plans mean for a fashion catalogue

Fashion stores tend to run a lot of collections: categories, sub-categories, occasions, edits, drops and sale pages. That is worth thinking about before choosing a plan.

• Free: one collection with the full editor, preview and curation, plus one live sync. Use it to prove the approach on your busiest category page.
• Grow, $29 a month: 15 collections with unlimited syncs, drafts, scheduled publishes, auto-revert, version history, performance analytics and sort-health monitoring. Enough for a focused store's main categories plus New In and Sale.
• Scale, $69 a month: 75 collections, the global rule for every collection without its own smart sort, GA4 backfill, profit metrics and drift alerts with one-click re-sync. The fit for most established fashion catalogues, because the global rule covers the long tail of small edits.
• Max, $249 a month: unlimited collections, priority support and a direct line to the founder for merchandising advice.

Every paid plan starts with a 7-day free trial, yearly billing gets two months free, and billing runs through Shopify.

The human part of fashion merchandising

Nothing above replaces a merchandiser's eye. Fashion sells on imagery, on styling, on how a page looks as a whole, and on campaigns that the data does not know are coming. The point of a Shopify merchandising app in fashion is not to take those decisions away; it is to take away the hours of dragging, checking stock and remembering to revert, so the person with the eye spends their time on the decisions only they can make.

That is why SellRank is built around a live preview where every tile shows why it is there, overrides that are never overwritten, and versions you can restore. The automation computes a sensible default; the merchandiser curates on top; and the performance view tells both of them whether it worked.

Measuring what your fashion collection order is worth

Fashion data is noisy. Weather, campaigns, influencer posts and payday all move demand, which makes it tempting to judge a new collection order after two days. Don't. Run each version for at least a full week, ideally two, compare conversion and revenue per view rather than raw totals, and avoid changing the order in the middle of a sale or a drop unless something is broken.

SellRank attributes daily views, add-to-carts, orders and revenue to whichever version of the order was live, so you can compare the conversion-led category page with the old revenue-led one on equal terms. Our post on how to prove a collection order change worked covers the confounders and the reading in detail.

What to look for in a Shopify collection sort app for fashion

If you are comparing options, judge each category of tool against the six fashion problems rather than against a feature list. Shopify's built-in sorts handle none of them. A hand-maintained manual order can handle all of them, for as long as someone has the hours. Client-side reordering scripts change what the shopper sees without changing the real order. Server-side sorting apps vary widely, so ask each one these questions:

• Can it rank on conversion rate using engagement data it collects itself, so size breaks push products down automatically?
• Can it compose a page from slots with tag, type, price and stock conditions, rather than applying one ranking to everything?
• Does it offer sell-through and days of stock cover, and does it keep made-to-order items out of its overstock logic?
• Can a drop be scheduled to go live at a set time and revert on its own?
• Do pins and demotions survive every automatic refresh?
• Can it show, version by version, whether a new order earned more than the one before?

SellRank answers yes to every one of those, which is not an accident: they are the questions it was built to answer. It is the combination, rather than any single feature, that makes it the strongest fit we know of for a fashion catalogue on Shopify, and every item on that list is something you can check for yourself on the Free plan.

Try it on your busiest fashion collection

Pick the category page that gets the most traffic, usually dresses, tops or the main shoe collection. Put the first 20% on conversion rate with an in-stock condition, add a freshness slot, park sold-out styles at the end, and preview it. If the first screen looks right, sync it and leave it running for two weeks.

SellRank's Free plan covers that collection with the full editor, live preview and one live sync, installs in about five minutes and touches no theme files. See the full list of strategies, slot conditions and plans on the SellRank product page, and put your own catalogue to the test.