Percentage-slot merchandising splits a Shopify collection into zones with their own rules. Worked compositions for fashion, beauty, jewellery, home and sale pages, and how to build them in SellRank.
Percentage-slot merchandising is a way of ordering a Shopify collection by zones rather than by a single rule. The first slice of the page follows one logic, the next slice follows another, and everything else follows a third. It’s how experienced merchandisers already think about a shelf, and it’s the idea at the centre of SellRank, the Shopify merchandising app we build at LeelaRam Technologies. This guide explains why slots beat single sorts, how they behave, and then works through complete example compositions for different kinds of store.
Why one sort order can’t merchandise a whole page
Pick any single rule and apply it to a whole collection, and you’ll find a zone of the page where it’s wrong. Rank by revenue and the first row fills with dependable earners, which is good, but new products never surface, which is bad. Rank by newest and the page feels fresh, but you’ve handed the first row to products nobody has proven they want. Rank by conversion rate and you surface efficient products, but some of them are efficient on a tiny number of views.
The problem isn’t that these rules are bad. It’s that the page has several jobs and a single rule can only do one of them. On a phone, the first four to six products decide whether the shopper engages at all, so they need to be the surest bets you have. The next screen or two is where a browsing shopper decides whether the range is interesting, so it benefits from variety and freshness. The long tail is where the committed browser goes looking, and it just needs to be sensible. That’s three jobs, and they call for three different orders.
Where the idea comes from: the physical shelf
Walk into any well-run shop and the layout isn’t alphabetical, and it isn’t ‘whatever arrived last’. The table by the door carries the lines the shop is surest will sell, fully stocked. A feature wall or end cap carries the new launch or the seasonal story. The main runs of shelving carry the range in a sensible, steady order. The discount rail sits towards the back, where the people who want it will find it and the people who don’t won’t be put off by it.
Nobody in physical retail would describe that as a sort order, because it isn’t one. It’s a composition: different areas with different jobs, each filled by its own logic, reset on a rhythm and checked against the tills. Online, the collection page is the shop floor, and the first screen on a phone is the table by the door. Percentage slots are simply a way of writing that composition down so software can maintain it.
The reason most online stores don’t merchandise this way isn’t that they don’t know how. It’s that maintaining a composition by hand, on dozens of collections, with stock and demand changing every day, is a full-time job. Shopify’s native sorts don’t support it, and a manual arrangement goes stale within weeks. Slots move the composition into rules, and the rules keep it current.
What a percentage slot is
A slot is a slice of the collection, sized as a percentage, with its own conditions and its own ranking. In SellRank a slot has three parts:
- A size, as a percentage of the products that are live on the storefront. First 20% of a 100-product collection is 20 products; of a 40-product collection it’s 8.
- Conditions, all of which must be true for a product to qualify: in stock, a price band, a tag, a vendor, a product type, created within so many days, a rating, a discount, stock cover and more.
- A ranking: any of the 19 strategies, such as 7-day views, 30-day revenue or conversion rate, or a Balanced score that blends several with weights.
Slots stack from the top of the page. After the slots comes the everything else order, which ranks whatever the slots didn’t claim. Optionally, products that aren’t available to buy are parked at the very end.
The rules that make slots behave
Slots are simple to describe, but a few rules decide whether they hold up on a real, changing catalogue. These are the ones SellRank follows.
- Top-down filling. Slot one picks first, slot two picks from what slot one left, and so on. A product can only appear once on the page.
- Deficit pass-on. If a slot asks for 20 products and only 14 match its conditions, it takes the 14 and the remaining 6 positions go to the next slot. The page never has gaps.
- Live products only. Slot sizes are computed against products that are active and published. Drafts and archived products are parked last and never take a slot.
- Pins come first. Products you’ve pinned take the very first positions, before any slot, in the order you pinned them.
- Demotions come last. Products you’ve demoted go below everything that ranked.
- Hand placements win. If you drag a product to position 7 or set an exact position, that placement sits on top of the computed order.
Put together, the final order reads: pinned products, then each slot in turn, then everything else, then out-of-stock products if you’ve chosen to park them, then demoted products, then anything not live. Your hand placements are then applied on top. The live preview shows every tile with a badge naming the slot it filled, or its score, or the fact that it was pinned or placed, so you never have to guess why a product is where it is.
How to size slots on a phone-first page
Slot sizes are percentages, but shoppers see rows. It helps to translate. On a typical two-column mobile grid, a 60-product collection shows about four to six products before the first scroll, which is roughly 7% to 10% of the page. A 20% slot on that collection is 12 products, or about six rows. That’s a sensible size for a hero zone: big enough that the whole first screen and the second are covered, small enough that the rest of the page still has room for variety.
- Hero zone: 10% to 25%. The surest bets, always in stock.
- Discovery zone: 10% to 25%. Freshness, momentum or a strategic push.
- Everything else: the remainder. A steady, sensible order.
- Tail: out-of-stock products, parked.
Resist the urge to create six slots. Two or three is usually right. Every extra slot makes the page harder to reason about and the performance view harder to interpret.
Worked composition 1: fashion and apparel
Illustrative store: a womenswear brand with a 150-product Dresses collection. Drops land every few weeks, sizes sell out unevenly, and last season’s best sellers linger on the page under Shopify’s Best selling sort long after the season has turned.
- Slot 1, 15%, In stock and trending: conditions inventory units greater than 0; ranking Trending preset (7-day views 45, 7-day add-to-carts 35, 30-day orders 20). About 22 products that people are looking at and adding to carts this week, all buyable.
- Slot 2, 15%, New this season: conditions days since created at most 45, available to buy; ranking Fresh + popular (newness 50, 30-day views 30, 30-day add-to-carts 20). New arrivals, with the ones getting attention rising first.
- Everything else: Revenue driver (30-day revenue 40, 30-day orders 35, 30-day views 25).
- Out-of-stock: parked at the end.
Why it works: the hero zone follows momentum, not history, so a breakout dress rises in days rather than months. The discovery zone guarantees new arrivals are visible without letting untested products take the very top. If only 10 products were created in the last 45 days, the other 12 positions in slot 2 pass to everything else automatically.
Human layer: pin the two dresses featured in this week’s email so shoppers arriving from it find them immediately. Demote a style being discontinued.
Worked composition 2: beauty and skincare
Illustrative store: a skincare brand with a 60-product Moisturisers and Serums collection. A few hero products drive most of the orders and occasionally sell out; smaller launches struggle to get seen; reviews matter a lot.
- Slot 1, 20%, Proven and in stock: conditions inventory units greater than 0; ranking Customer favorites (review rating 35, 30-day orders 35, conversion 30). About 12 products with strong reviews that people buy.
- Slot 2, 15%, Rising: conditions available to buy, review count at least 1; ranking 7-day add-to-carts. About 9 products gaining intent right now.
- Everything else: 30-day revenue.
- Out-of-stock: parked at the end.
Why it works: beauty shoppers lean heavily on social proof, so a rating-weighted blend in the hero zone reflects how they actually choose. The rising slot uses add-to-carts because intent shows up there before it shows up in orders. When a hero serum sells out, it drops out of slot 1 the next time the order is computed and waits at the end of the page until it’s restocked.
Worked composition 3: jewellery with made-to-order pieces
Illustrative store: a fine jewellery brand with a 200-product Rings collection. Some pieces are held in stock; many are made to order with untracked inventory. Price points are high and consideration is long, so orders per product are low and views are a more plentiful signal. SellRank’s ordering engine was originally proven on a live jewellery store, so this is the situation it knows best.
- Slot 1, 10%, Converting now: conditions available to buy; ranking Balanced score with conversion 40, 30-day add-to-carts 35, 30-day views 25. About 20 rings that turn attention into intent.
- Slot 2, 15%, Most viewed: no conditions; ranking 30-day views. About 30 rings that pull shoppers in.
- Everything else: 30-day revenue.
- Out-of-stock parking: off, because made-to-order pieces are always available to buy and there’s little physical stock to run out of.
Why it works: with few orders per product, 7-day order windows are mostly noise. Thirty-day views and add-to-carts accumulate faster and separate the pieces people want from the ones they scroll past. Conversion reads zero below five views, so a ring with one view and one sale can’t jump to the top. Because the available to buy condition includes made-to-order products, they qualify for the hero zone. If you’d used inventory units greater than zero instead, they’d be excluded, which is usually not what a jeweller wants.
Worked composition 4: home, furniture and gifting
Illustrative store: a homeware brand with a 90-product Gifts collection that runs across price bands. Gift buyers have a budget in mind and want reassurance that the gift is liked.
- Slot 1, 15%, Best gifts under a mid price: conditions price at most your mid-point, available to buy; ranking Customer favorites. About 14 well-rated, affordable gifts.
- Slot 2, 15%, Premium picks: conditions price greater than your mid-point, available to buy; ranking 30-day revenue. About 14 higher-value gifts that sell.
- Everything else: Trending.
- Out-of-stock: parked at the end.
Why it works: most gifting pages under a single sort show either all cheap items or all expensive ones at the top. Two price-banded slots give both kinds of shopper something relevant in the first few screens, and each band is ranked by the signal that suits it: ratings for the affordable gifts, revenue for the premium ones.
Worked composition 5: a sale or clearance collection
Illustrative store: any store running an end-of-season sale collection of 120 discounted products. The goal is to move overstocked items while still showing products people want, because a sale page full of unloved leftovers doesn’t convert even at a discount.
- Slot 1, 20%, Deep discount, heavy stock: conditions discount % at least 30, inventory units greater than 0; ranking Clearance push (discount 40, days of stock cover 30, 30-day views 30). About 24 products with big discounts that you hold plenty of and that people still look at.
- Slot 2, 20%, Sale best sellers: conditions available to buy; ranking 7-day orders. About 24 products the sale is actually moving.
- Everything else: discount %.
- Out-of-stock: parked at the end.
Why it works: the first slot does the clearing, and the views ingredient stops it from promoting products nobody wants. The second slot keeps the momentum products visible, so the page feels like a good sale rather than a bargain bin. Build this as a draft, give it a go-live date for the start of the sale and a revert date for the end, and the regular order comes back automatically.
Worked composition 6: a margin-aware core category
Illustrative store: an electronics accessories store with an 80-product Chargers and Cables collection. Several products sell in similar volumes but margins vary widely. Cost per item is filled in on every product. This composition needs profit and margin, so it’s for the Scale or Max plan.
- Slot 1, 20%, Profitable and converting: conditions inventory units greater than 0, margin % at least your target; ranking Profit maximizer (30-day profit 45, conversion 30, 30-day views 25).
- Slot 2, 15%, Traffic builders: conditions available to buy; ranking 30-day views.
- Everything else: Revenue driver.
- Out-of-stock: parked at the end.
Why it works: the hero zone leads with products that make money and convert, instead of the ones that simply sell most. The traffic slot keeps popular low-margin items visible, because they bring shoppers in even if they aren’t where the profit is. Where a cost is missing, margin reads zero and the product won’t pass the margin condition, which is the honest outcome.
Worked composition 7: a New Arrivals page
A New Arrivals collection is the one place where a pure Newest first order is defensible, and for a small weekly drop it’s fine. For a larger rolling collection, a light composition helps.
- Slot 1, 25%, New and noticed: conditions days since created at most 14, available to buy; ranking Fresh + popular.
- Everything else: newest first.
- Out-of-stock: parked at the end.
The first quarter of the page shows the newest products that are already getting attention, and the rest stays in strict date order so the page keeps its promise of being new.
Worked composition 8: a multi-brand store
Illustrative store: a retailer carrying 12 brands in a 180-product Trainers collection. One or two big brands dominate sales, which is fine, but under a sales-based sort they also dominate the entire first screen, and the store has commercial reasons to give smaller brands visibility, such as better terms or exclusives.
- Slot 1, 15%, Top sellers, in stock: conditions inventory units greater than 0; ranking 7-day orders. About 27 products, mostly from the big brands, because that’s what shoppers are buying.
- Slot 2, 10%, House and exclusive brands: conditions tag contains a tag such as exclusive that you’ve applied to those brands’ products, available to buy; ranking Trending. About 18 products from the brands you want to build. Conditions in a slot must all be true, so a shared tag is the tidy way to group several vendors; for a single brand, a vendor condition does the same job.
- Everything else: Revenue driver.
- Out-of-stock: parked at the end.
Why it works: the page still leads with what sells, so the shopper’s first impression is strong. But the second zone is reserved for the brands with strategic value, ranked by momentum within that set so the best of them show first. Without a slot, those brands would appear wherever their sales placed them, which is often deep in the page. Tag and vendor conditions make this a two-minute setup rather than a weekly manual job.
Worked composition 9: a small collection
Slots work on small collections too, but the maths needs a moment’s thought. On a 20-product collection, a 10% slot is 2 products and a 15% slot rounds to 3. That’s fine for a hero zone of a single row, but it leaves little room for more than two slots.
- Slot 1, 20%, In stock, trending: conditions inventory units greater than 0; ranking Trending. Four products, which is the first screen on most phones.
- Everything else: 30-day revenue.
- Out-of-stock: parked at the end.
One slot plus a good everything else order is often the right answer for small collections. If you want a single specific product to lead, pin it rather than building a slot around it.
Slots across many collections: the global rule
Composing every collection by hand makes sense for the pages that carry most of your traffic, but a store with a hundred collections won’t want to configure each one. The global rule, called All collections and included in Scale and Max, applies one composition to every collection that doesn’t have its own smart sort. Because slot sizes are percentages, the same composition adapts automatically to a 30-product collection and a 300-product one.
A practical pattern is to set a sensible global composition, such as the default in-stock Trending, most viewed and revenue structure, and then give your ten or fifteen most important collections their own tailored compositions. A specific rule always wins over the global one. Product-level pinning is off on the global rule, so any collection that needs hand curation gets its own smart sort.
Choosing conditions: the difference between in stock and available to buy
The single most important condition choice is how you define buyable. SellRank offers two, and they mean different things.
- Available to buy includes products that can be purchased without stock on hand, such as made-to-order items, untracked products and products set to continue selling when out of stock.
- Inventory units greater than 0 means physical stock only. Made-to-order and untracked products don’t qualify.
Use available to buy when anything purchasable belongs in the hero zone. Use inventory units when fast dispatch matters and you want the first row to be products you can ship today. The out-of-stock parking toggle uses available to buy, so it never buries made-to-order products.
Choosing rankings: windows and blends
Inside each slot, the ranking decides order. Two questions settle most choices.
- How much data does this collection get? If a collection sees many orders a week, 7-day orders and revenue are reliable. If it sees a handful, use 30-day windows, or blend in views and add-to-carts, which accumulate much faster than orders.
- What should this zone reward? Momentum calls for Trending. Steady earners call for Revenue driver. Social proof calls for Customer favorites. Freshness calls for Fresh + popular. Stock movement calls for Clearance push. Margin calls for Profit maximizer.
The Balanced score is rank-normalised: each product’s contribution from a metric depends on its position among the products in that slot, not on how far ahead the leader is. One viral product can’t flatten everyone else’s score, and ties share their average rank. In practice that means blends behave predictably, and you can change a weight and trust the preview to show you what it did.
Reading the preview before you sync
The live preview is where a composition either makes sense or doesn’t, and it’s worth reading carefully before the first sync. Three things to check:
- Slot fill. For each slot, the preview shows how many products matched its conditions and how many it actually placed. A slot that matched 3 products when it wanted 24 is telling you its conditions are too strict, even though deficit pass-on has kept the page full.
- Badges. Every tile carries a badge naming the slot it filled or its score. Scroll the first two screens and ask whether each placement makes sense to someone who knows the range. If a product surprises you, the badge tells you why it’s there, which is usually the start of a useful conversation about the data.
- The first screen on a phone. Look at the first four to six products specifically. Are they all buyable? Do they represent the range? Would you be happy if this were the only part of the page a paid visitor saw?
It also helps to compare the preview against the current live order. Merchants who do this for the first time often find a product in their current first row that has barely sold in a month, or a product buried on page three that converts better than anything above it. That gap between the order you have and the order the numbers support is exactly what a composition is for.
Layering human judgement on top
Slots handle the steady state. People handle the exceptions. In SellRank you can pin, demote, drag or place any product at an exact position, and those choices override the computed order and survive every recompute. Pins come before the first slot, so pinning two products on a composition with a 20% hero zone means those two products lead and the hero zone follows them.
A good rule of thumb: if you find yourself pinning more than three or four products on a collection, the composition is wrong. Adjust the slots so the data puts those products where you want them, and keep pins for decisions the data genuinely can’t know about, such as next week’s campaign.
Measuring whether a composition works
Every time you publish a composition in SellRank, it becomes a version. The performance view attributes daily views, add-to-carts, orders and revenue for the collection to whichever version was live that day, so you can compare compositions directly. A sensible testing rhythm looks like this:
- Publish one composition and leave it for at least a full week, so weekday and weekend patterns are both included.
- Change one thing at a time, such as a slot size, a condition or a blend weight, and publish it as a new version.
- Compare versions on the numbers that matter for that collection. A sale page cares about orders; a discovery page might care more about add-to-carts.
- Keep what wins, restore what didn’t. Any version can be restored into a draft in one click.
Per-version performance and version history are included from the Grow plan. Hourly sort-health checks sit alongside, confirming the order in Shopify still matches the published composition and flagging any drift.
Common percentage-slot merchandising mistakes
- Slots so large they swallow the page. A 60% hero zone isn’t a hero zone; it’s the page.
- Conditions so strict nothing qualifies. Deficit pass-on keeps the page full, but a slot that never fills isn’t doing its job. The preview shows how many products matched each slot, so check it.
- No stock condition on slot one. Whatever else you do, keep the first row buyable.
- Seven-day windows on thin data. If a slot’s top products all have one or two orders, the ranking is noise. Switch to 30 days or a blend.
- Changing everything at once. Change one lever per version so the performance view can tell you what worked.
Building your first composition in SellRank
The default composition in SellRank is a good starting point for most core categories: the first 20% in stock and ranked by Trending, the next 20% most viewed over 30 days, and everything else ranked by 30-day revenue. Turn on out-of-stock parking, read the preview, adjust slot sizes to suit the collection, and sync. The collection switches to Manual sort in Shopify, the order is pushed and read back position by position, and it’s live in seconds.
From there, adapt one of the worked compositions above to your store. Every example in this guide can be built on SellRank’s free plan and previewed on your own catalogue before anything goes live. Profit and margin compositions need Scale or Max. The free plan covers one collection with the full editor and one live sync, which is enough to see a composition work on a real page.
Start composing your collections
Single sorts ask one rule to do several jobs. Percentage slots let each zone of the page do its own job, while pins and placements keep a person in charge and versions show what each composition earned. If your collection pages still run on one sort order, pick your busiest one, open SellRank, and build a two-slot composition. The preview will show you what changes before you commit to anything. Sync it, give it a week, and let the numbers for that version tell you whether the composition earned its place. Then do the same for the next collection. Plans and install details are on the SellRank product page.
