Shopify replaced its old smart-and-custom collections split with a source-based model in the 2026-07 Admin API. A single collection can now combine automated rules, hand-picked products, exclusions, other collections, and even app-supplied products, and it can target individual variants instead of whole products.
- For nearly two decades a Shopify collection was either custom (manual) or smart (rule-based). The new model lets one collection do both at once.
- Collections can now exclude products, reference other collections as building blocks, and target specific variants (only the red ones, only XS and XXL).
- Apps can feed shareable sources into your native collections, so a reviews or subscriptions app's intelligence lands inside a collection you still control.
- Existing collections keep working and there is no deadline. The one thing to check: the apps that read your collections should be on Shopify's 2026-07 API.
Shopify just rebuilt how collections work, and most merchants have not heard yet. For nearly two decades, a Shopify collection came in exactly two shapes. A custom collection, where you picked every product by hand. Or a smart collection, where you set rules and Shopify filled it automatically. You chose one. In Shopify's 2026-07 release, that either-or is gone.
A collection is now built from sources. A source is one ingredient: an automated rule, a hand-picked list, an exclusion, another collection, even a set of products an app supplies. You stack the ones you need. So a single collection can run on rules and let you override them by hand, at the same time, which is the thing merchants have wanted for years. Shopify laid this out in its partner announcement and developer changelog in June 2026.
I have merchandised on Shopify from every seat: as an early employee when this catalog model was being built, as an operator at WIN Brands Group running collections across a portfolio of brands, and as an advisor since. So this is the operator's read, not the changelog. Here is what actually changed, what it unlocks for how you merchandise, and the one small thing worth checking. Good news up front: nothing is on fire, and there is no deadline.
The smart-or-custom
choice is
over.
The headline is simple. Collections are no longer locked into the old smart-versus-custom split. Under the hood, every collection is now made of one or more sources, and a source can be automated, manual, or supplied by an app. Your existing collections still work exactly as they did. They are just the simplest version of a wider model.
That wider model does four things the old one could not. A single collection can now mix automation with hand-curation, and exclude products rather than only include them. It can also target individual variants instead of whole products, treat another collection as a building block, and pull in products an app supplies. The table lays the old model against the new one.
| Question | Old (smart or custom) | New (source-based) |
|---|---|---|
How you build it |
Pick one: all manual, or all rules |
Stack sources: rules plus manual plus more, together |
Automate and curate at once? |
No, one or the other |
Yes, in the same collection |
Exclude specific products? |
No native exclusions |
Yes, exclusion conditions |
Target a single variant? |
No, whole products only |
Yes, target variants (e.g. only red) |
Reference another collection? |
No |
Yes, collections as building blocks |
Let an app supply products? |
Not into a native collection |
Yes, shareable app sources |
None of this forces a migration. Shopify built the change to be non-breaking: your current smart and custom collections keep running, the old fields stay available, and you adopt the new pieces when you have a reason to. There is no switch-off date, which is the opposite of how the Scripts deprecation felt earlier this year. If you lived through that scramble, you can read what the end of Shopify Scripts actually meant for the contrast: this change is all upside and no clock.
Why the old model
made you
pick a lane.
To see why this matters, remember what the old model cost you. A smart collection was great at staying current. Set a rule like "tag is summer" and it updated itself as you tagged products. But the moment you wanted to hand-place one hero product at the top, or pull one item that technically matched but did not belong, you were stuck. Smart collections did not take manual overrides.
Custom collections had the opposite problem. Total control, zero automation. You curated every product by hand, then re-curated every time your catalog changed. For a small drop, fine. For a "Sale" collection across a few thousand SKUs, it was a part-time job, and it drifted out of date the day someone changed a price.
So merchants did what merchants do: they worked around it. They ran duplicate collections, rewrote product lists on a schedule, leaned on a merchandising app to take over the whole collection, or built a collection-like page outside Shopify entirely. Every workaround added drift and maintenance. Shopify's own framing of the change puts it plainly: good merchandising is rarely fully automated or fully manual.
"For nearly twenty years, a Shopify collection made you choose: automate it, or curate it. The new model is Shopify finally admitting the best merchandising is both."
That is the frustration this release retires. The change is subtractive: it takes away a constraint that shaped how every Shopify merchant built a catalog page for two decades. Once you internalize that a collection is now a stack rather than a single mode, a lot of daily merchandising friction disappears.
This is the kind of platform call I help DTC operators make. If it is landing, the form takes two minutes.
A collection is
now a stack
of sources.
The whole model fits in one sentence: a collection is a set of sources, and each source adds products or variants using its own logic. You can have one source or several. Shopify evaluates them together and keeps the result current as your catalog changes, so you are not re-listing products by hand. The pieces you can stack are below.
| Source type | What it does | Merchant example |
|---|---|---|
Conditions (rules) |
Auto-includes products that match, like the old smart collections. |
Most common Tag is "summer", or price is under $50. |
Manual selections |
Hand-picked products you place yourself. |
Curation Feature three hero items at the top. |
Exclusions |
Removes products that match, even if a rule included them. |
New Everything on sale except clearance. |
Sub-collections |
Pulls products from other collections you reference. |
New A "Shop All" built from your category collections. |
App sources |
Products an app supplies into your collection. |
New Top-rated items pulled from your reviews app. |
Two of these deserve a flag because they are genuinely new. Exclusions mean you can finally say "everything on sale except clearance" in one collection, instead of building around the items you do not want. And sub-collections mean a parent collection can be assembled from other collections and stay in sync automatically: change a child, and the parent updates. If you have ever maintained a "Shop All" that lagged behind your category pages, that problem just went away.
- 01 Start with a rule: include every product with a sale price. The collection fills itself and stays current as prices change.
- 02 Add a hand-picked source: drop in three hero products you want featured whether or not they are discounted.
- 03 Add an exclusion: leave out the Clearance category, which is technically on sale but off-brand for this campaign.
- 04 One collection, three sources, no duplicate lists to maintain. That is the thing the old model simply could not do.
The campaigns
you couldn't
build before.
The payoff is concrete. The real value shows up in the merchandising moments that used to need a workaround, and a few of them come up constantly in the brands I have run and advised.
The seasonal campaign that curates and automates. A Black Friday page or a spring drop wants both: a rule so it captures every eligible product automatically, and a human hand to feature the pieces you are betting on and cut the ones that clash. That is now one collection instead of a rule-built collection plus a manually-reordered duplicate someone has to babysit.
The "Shop All" that never lags. Build it from your category collections as sub-collections, and it updates itself whenever a category does. No more flagship browse page quietly missing last week's launch because nobody re-added it by hand.
The exclusion you always wanted. "New arrivals, minus final-sale." "Best sellers, excluding the three SKUs we are sunsetting." "Gifts under $50, but not the items low on stock." These were painful before and are trivial now, and they are exactly the collections that protect margin and brand.
None of these are exotic. They are the everyday merchandising a growing brand does weekly, at every growth stage, and the old model taxed every one of them with maintenance. Removing that tax is quietly one of the more useful platform changes of the year for operators, even if it arrived dressed as an API note. If you have been paying a third-party merchandising app to do this kind of combining, it is worth re-reading what Shopify now does natively before your next renewal.
Merchandise the
variant, not the
whole product.
The second genuinely new capability is variant targeting, and it fixes a problem every apparel and multi-variant brand knows too well. Until now, a collection held whole products. If you built a "Valentine's Reds" collection, you could only pull in entire products, so a shirt that comes in eight colors showed up with all eight, wrong swatches and all. The new model lets a collection target specific variants.
That sounds small. On a campaign page it is not. When a collection includes the whole product for the sake of one variant, you get the messy result Shopify itself calls out: wrong color swatches, misleading price ranges, inaccurate filters, and too many choices on a page meant to feel tight and intentional. Variant targeting cleans all of that up.
The uses are obvious once you have them. A Valentine's collection that holds only the red variants. An extended-size promotion that pulls only the XS and XXL. A clearance collection built from specific clearance variants, not whole products that are half full-price. A point-of-sale collection that surfaces only the variants your store staff need to find fast. Each of these used to require a workaround or a separate duplicate product. Now it is a targeting choice inside one collection.
One collection,
used across the
whole business.
Collections were never only a storefront tool. You use product groups in discounts, shipping rules, tax overrides, point-of-sale, feeds, and catalog operations. The problem was that each of those often needed its own separate list, so you maintained the same logic in five places and they drifted apart. The new model makes a collection a shared building block the whole business can point at.
Build one collection around a set of variants, and it can drive a storefront campaign and a discount at once. Build a collection from rules, and every downstream workflow that references it stays aligned as products change. Keep a collection unpublished, and you can use it purely as an internal operations group buyers never see. The result is less duplication and less drift, with one definition to maintain instead of five.
What app-owned sources mean for you
The piece with the biggest long-term implications is shareable app sources. Your apps have always known things your collections wanted. A reviews app knows which products are top-rated. A subscriptions app knows which are subscribe-eligible. A search app knows what is trending. Before, they had no clean way to put that intelligence inside a native Shopify collection. They took over the whole collection, or rendered their own thing off to the side.
Now an app can publish a source that you add to a collection you control, and combine with your own rules, manual picks, and exclusions. So "top-rated items, plus my hand-picked heroes, minus clearance" can be one native collection, with the ratings supplied by your reviews app and the rest by you. You stay in control of the final collection; the app just contributes what it knows. For brands, that means less of your merchandising trapped inside individual apps, which is a healthier place to be when you reassess what actually belongs in your Shopify stack at each stage.
What you should
actually do
now.
The calm version first, because this is not a fire drill. There is no deadline, nothing switches off, and your current collections keep working untouched. This is a capability you adopt when a real merchandising need calls for it, not a migration you have to schedule. Three moves, in order.
Questions I keep
getting asked.
If you are rethinking how your store is merchandised, or which apps still earn their place now that Shopify does more of this natively, that is exactly the kind of call I help operators make. The DTC brand practice is where we work it through. The form takes two minutes: start the conversation.
Scaling a consumer brand?
I work with a deliberately small number of DTC operators. I've run brands at this scale myself, from $5M past $100M. Not theory. If you're in that range, the form takes two minutes.
Start a conversation More about Taylor →