Shopify Merchandising: The Definitive Guide

Why does manual curation break down?

Manual curation breaks on two rocks: time and staleness. Dragging products into order works the day you do it, and the grid starts rotting the day after.

The time cost compounds with the catalog. Say your store runs 24 collections and you re-sort each when stock turns, about ten minutes of dragging tiles per collection. That is four hours of sorting, and stock turns more than once a month. The work lands on whoever is nearest, usually you, usually at night.

Staleness is the quieter failure. A bestseller sells out Friday evening and leads the grid all weekend, a front table piled with something nobody can buy. The fresh drop you photographed and paid to promote enters at the bottom, under products from two seasons ago. The sale ends Sunday, and the sale sort hangs around until somebody remembers Wednesday.

And manual order cannot see what shoppers do. Your store knows which products get clicked, carted, and bought this week. A hand-built sort knows what you believed the afternoon you built it. Those drift apart fast: the product you love is not always the product shoppers vote for, and the gap is a front table earning less than it could.

None of this means hand-picking is wrong. Pinning a hero product, leading a campaign collection with the campaign: these are the good part of curation. The trap is making every placement a human call, because then every placement waits for a human. The rest of this guide is about keeping the calls and automating the upkeep.

Can you merchandise a Shopify store without an app?

Partly, yes. Shopify ships collection sort orders and a free Search & Discovery app, and both do honest work. The limits show up when you want to pin a few products and automate the rest, schedule a sort, rank by your store's own metrics, or control collections and search from one place.

Collection sort orders. Every Shopify collection has a sort: best selling, price low to high or high to low, alphabetical, newest, or manual. Set it in the admin, done, free. The limits are structural. One sort per collection, no blending: best sellers first OR newest first, never "new arrivals on top, best sellers after." Manual is all or nothing; the moment you drag one tile, you own the position of every tile. Best selling runs on Shopify's own logic, with no window you can tune and no way to weight it. Nothing schedules: the sale sort goes up by hand and comes down by memory. And sort orders govern collections only; search results live elsewhere.

Shopify Search & Discovery. Shopify's free first-party app covers storefront search and a bit beyond: filters for search and collection pages, synonym groups, product boosts that float chosen products higher in search results, and related or complementary products on product pages. For a small catalog with stable bestsellers, it may be enough.

Its limits, fairly stated: boosts are hand-picked product lists, so the staleness problem from chapter two applies in full. There are no conditions, nothing like "boost what's in stock and selling this month." No scheduling, no preview before it ships, and no keyword scoping, so a boost lifts a product across all of search rather than for the query you care about. Collection grids are untouched. And nothing reports what a boost earned, so "did that help" stays a feeling.

The honest summary: native tools place products by hand and hold them there. If your catalog is small and changes rarely, start there, free is free. The moment placement should respond to stock, sales, seasons, or searches, hand-placed stops scaling, and rules take over.

How does rule-based merchandising work?

A merchandising rule is a placement decision written down once and enforced continuously. Instead of dragging products into order, you describe the order: pin these, boost what is selling, demote what is low in stock, hide what should not show. The rule keeps the grid arranged as products come and go.

The eight strategies in Sledge Merchandising read like verbs because they are:

• Pin holds chosen products at the top. The hero stays the hero.
• Boost lifts products up the ranking without locking the whole grid.
• Demote sinks what you would rather not lead with: last season, low margin, low stock.
• Hide removes a product from view without touching the collection itself, so the catalog stays intact while the grid stays clean.
• Filter narrows the grid to only the products that meet your conditions.
• Inject product drops products from another collection into a position you pick.
• Bundle, Discount & Upsell places a discounted bundle or an upsell at a chosen position, set per surface.
• Banner places a promotional banner at a chosen position, set per surface.

Rules run on collection pages and on search, the two places a shopper reads a list instead of a single product. One rule can carry several strategies at once.

The conditions are where rules earn their keep. Build them from product type, vendor, tag, price, inventory, and new arrivals, then combine them with your store's own metrics: best sellers, most viewed, click-through, add-to-cart rate, each measured over a window you choose. "New arrivals first" is one condition. "In-stock best sellers under $50 lead the grid" is three. The rule keeps working as stock moves and seasons turn, the upkeep manual sorting could not survive.

Two finishing touches close the old gaps. Rules take start and end dates, so the holiday sort arrives on schedule and the everyday order returns on its own, no Wednesday surprises. And live preview shows the rule before it publishes, the collection grid on desktop and the same grid on a phone, so it does not ship sight unseen.

The hand-versus-rule choice is not either-or. Pin the hero by hand, let a rule sort everything beneath it. Boost by hand. Or by rule.

How do product badges and labels work?

Badges are the loud part of merchandising: small labels on product cards that say "New," "Almost gone," or "Bestseller" before a shopper reads a word of the title. A good badge moves the eye to the card you want clicked; a lazy badge papers the grid until shoppers stop seeing badges at all.

In Sledge, badges run as campaigns, not stickers. Pick a preset label or write your own, set the colors, shapes, and positions, then aim the campaign at chosen products or a whole collection. A "New season" campaign can cover the collection you built for the drop, and an "Almost gone" label follows the product's own stock so it clears itself when it stops being true.

Scheduling keeps badges honest. Campaigns take start and end dates, so "Holiday pick" leaves with the holidays instead of haunting February. A product badge is a claim, and claims expire; campaigns make the expiry automatic.

Scarcity is the rule worth tattooing somewhere: a badge on everything is a badge on nothing. If half the grid shouts, the grid is just loud. Aim campaigns narrowly, at the products with a real claim to make, and let the rest stay quiet so the claim can land.

And a badge campaign can be A/B tested, so "Almost gone" runs against "New" on live traffic and the winner comes back as a number instead of a preference. The badge that does not move anything is decoration, and decoration goes in the next cleanup.

Can you merchandise search results, not just collections?

Yes, and the stores that do usually find search was the better half of the job. A shopper browsing a collection is wandering; a shopper typing into the search bar is asking for something in their own words. Merchandising that answers the words is placement at the exact moment of intent.

Sledge rules scope to search keywords the same way they scope to collections. Search for "gifts," and a keyword-scoped rule leads with your gift sets while the rest of search stays untouched. "Sale" can lead with the deepest cuts. A brand term can lead with that vendor's bestsellers. Each rule answers one query pattern, and the rest of the store never notices.

The same placement runs in both places shoppers search: the search results page and instant search, the overlay that appears as a shopper types. Instant search is the highest-rent shelf in the store, five to eight slots in a dropdown, so what ranks there is a merchandising decision whether you make it or not. Keyword-scoped rules make it yours.

This is also where merchandising and search quality meet. Ranking the right products only pays if search finds them in the first place: typo tolerance, synonyms, and intent-aware matching come first, placement second. Sledge Search handles that half, and the search optimization guide covers it end to end.

Native tools stop short here. Collection sort orders never touch search, and hand-picked boosts lift a product across all queries or none. Keyword scoping, place this product for this search, is the piece that turns search results into shelf space you control.

How do you measure whether merchandising worked?

The same way you measure any offer: in dollars, against the version you replaced. A reordered grid that does not change what shoppers click and buy is a tidy grid, and tidy is not the goal. Revenue is.

The cleanest answer is an A/B test. Run the new order against the old on live traffic, and let shoppers vote with their carts. Sledge A/B tests merchandising rules like any other entity, with revenue tracked per variant, so "did new-arrivals-first beat best-sellers-first" comes back as a number, not a debate. Test everything. Keep what wins.

Badges measure the same way. Put two labels head to head in an A/B test and the results tell you which words work on your shoppers. Keep the label that pulls, retire the one that decorates.

Then the question nobody has time to ask weekly: which merchandising move should come next? Growth Intelligence asks it for you. It reads your store and flags the moves worth making, fresh stock getting buried under old bestsellers, for example, each with a dollar estimate attached. Click it and the rule arrives pre-filled, ready to review and save. Once it's live, Growth Intelligence measures what the rule earned in real sales and shows the estimate next to the result. Every suggestion is checked against real sales, so you always know what's working. And why.

For the wide view, Analytics ties revenue back to features and pages, so collections and search each report what they earn, and the journey view shows whether shoppers stall while browsing or later. A grid is a hypothesis about what shoppers want. Measurement is how it stops being a guess.