"Shopify collections only give you six sort orders" is still one of the most repeated lines in Shopify how-to content. As of September 1, 2026, it's wrong.

The help center lists nine options, and the GraphQL Admin API's CollectionSortOrder enum matches at nine values. The difference is mostly one addition: Most relevant, shipped by Shopify on April 15, 2026 — and made the default sort for every collection created after that date. The default changed; the explanations didn't.

Two more facts quietly govern how your collections order themselves. Best selling counts all-time orders, not the trailing 30 or 90 days. And the free, Shopify-built Search & Discovery app's "push out-of-stock products to the end" setting applies only to search results and predictive search — it does not touch collection page ordering.

Below: what the native feature set can and can't do, sourced from Shopify's own documentation, and eight apps that push sold-out products to the bottom, compared on the same axes. Research date: September 1, 2026. All prices are USD as listed on the Shopify App Store.

The short version

By scenario:

• A few dozen products, and hiding sold-out items is acceptable → no app needed. Add the condition "Inventory stock is greater than 0" and set the source to Match all conditions. Done.
• You want to demote sold-out items, not hide them, at the lowest possible cost → MB Push Down Hide Out of Stock (free to $13.99/mo, real-time operation stated on every plan)
• You need a non-English admin UI → Nada • Hide Out of Stock ($9.99–$19.99/mo). One of the very few sort apps that lists a Japanese UI.
• You want inventory-aware sorting on a free plan → SortWise Collection Sort (hourly automated sorting starting on the free plan)
• Many collections, compound rules, low budget → Jedi ‑ Collection Sort Pro (from $2.99/mo, billed on collection count alone)
• You want to sort by GA4 pageviews or add-to-cart rate → ST: Product & Collection Sort (from $19/mo)
• Slow inventory turnover, and you want multivariate merchandising including return rate and margin → KX AI Collections (from $19/mo, but updates run 1–4 times per day)
• Your real problem isn't the sort order — it's the 60-condition cap or nested AND/OR → sort apps won't fix that. Smart Collection Pro (from $5/mo, Japanese UI listed) is the relevant tool.

And one thing every app in this comparison has in common: not a single listing states what happens to your product order after you uninstall — whether it persists or reverts to a native sort. That's the most important hidden cost here, and it's a question for each vendor's support team before you install.

There are nine sort orders

Here's how the admin labels map to the API enum values.

#	Admin label	API enum value
1	Most relevant (default for new collections)	MOST_RELEVANT
2	Best selling	BEST_SELLING
3	Product title A-Z	ALPHA_ASC
4	Product title Z-A	ALPHA_DESC
5	Highest price	PRICE_DESC
6	Lowest price	PRICE_ASC
7	Oldest	CREATED
8	Newest	CREATED_DESC
9	Manually	MANUAL

Most relevant is documented as "based on sales performance" and nothing more

Shopify's April 15, 2026 changelog announced Most relevant and noted only that existing collections' sort orders were left unchanged. The help center goes as far as "the most relevant products based on sales performance" and stops. The lookback window, the weighting, and whether inventory factors in are all unconfirmed — none of it is published.

That has two practical consequences. Any collection created on or after April 15, 2026 is ordered by an undisclosed algorithm unless someone deliberately changed it. And on any project where you have to justify merchandising decisions to a team or a client, leaving Most relevant as the default is hard to defend.

Best selling counts all-time orders

The help center defines Best selling as the all-time number of orders containing that product. Not the last 30 days, not the last 90. Products with no sales history fall back to newest-first.

This is the mechanism behind "I set it to best selling and a hit product from three years ago won't move off the top." There's no native option to set an aggregation window, so the moment you want "recent best sellers," you're evaluating apps.

Where native automated conditions run out

Item	Native behavior
Condition cap per collection	60 total
Behavior once the cap is reached	Not documented in current official docs — unconfirmed
Logical operator	Per source, either Match all conditions or Match any condition
Nesting	Not supported. You cannot build an AND/OR hierarchy.
Partial composition	One collection can hold multiple sources (Products / Variants / Collection / App), and sources effectively OR together. Exclusions then override every inclusion.
Fields usable in exclusions	Product category / Tag / Type / Vendor only
Comparison operators	>= and <= don't exist. Shopify's own documented workaround for "$100 or less" is "is less than $100.01" — shift by one cent.

So the most you can build natively is a two-level composition: OR across sources, with exclusions overriding. Genuine nesting — something like "(tag A or tag B) and price under a threshold" — cannot be expressed inside a single source.

Store-wide limits are documented too: 5,000 automated collections, 100 collections that include variants, 50 collections that include other collections, 5 collections that exclude other collections, and 128 metafield definitions each for products and variants usable in collection conditions. The Starter plan doesn't support collections at all.

You also can't assume the results are always correct. The help center's conditions page has a dedicated troubleshooting section acknowledging that "in rare cases" a collection includes products that don't match its conditions, or omits ones that do. The three documented workarounds are: duplicate a condition, save, then delete it; make a trivial change to the product, save, and revert it; or shift a price condition by $0.01, save, and revert it. All three amount to forcing a re-save so the collection refreshes — which means designing around a sale that must go live at an exact minute is risky.

"Just use the free official app" doesn't work

This is the most common misconception in this whole topic.

Shopify's Search & Discovery app has an "Out-of-stock products: show / hide / push to the end" setting. But the official documentation is explicit that this setting adjusts search results pages and predictive search. Collection page default sorting is out of scope. Product boosts are likewise limited to searches performed on the online store. What Search & Discovery gives collection pages is filtering, not ordering.

Shopify Flow is the same story. Flow's Send Admin API request action is documented as not supporting asynchronous mutations or mutations that return a Job type — and collectionReorderProducts, the reordering mutation, is exactly an async, Job-returning mutation. What Flow can do is unpublish products at zero inventory (the official template "Publish and unpublish products based on inventory level") and switch the sort method itself via collectionUpdate.

The only officially documented native solution is to add "Inventory stock is greater than 0" with Match all conditions and drop sold-out products out of the collection entirely — remove them, not demote them. If you need those product pages to stay live for SEO, back-in-stock notifications, or as ad landing pages, you're out of native options right there.

Manual adjustment isn't unlimited either. On the API side, collectionReorderProducts requires Collection.sortOrder to be MANUAL, and each request accepts a maximum of 250 moves. In the admin, Collection items paginates at 60 products per page, and beyond drag-and-drop you get Move, then To the beginning / To the end / To position. For small adjustments, Shopify also documents asking Sidekick to reorder products in natural language.

Language support is a separate question from a translated UI

Very few sort apps list a non-English admin UI. Across the apps researched here, only two list Japanese in the Languages field on apps.shopify.com: Nada • Hide Out of Stock and Smart Collection Pro. Sort'd, Jedi, ST, SortWise, MB, KX AI Collections, and Sortimize are English-only.

And a translated UI is not the same as support in that language. Each app's page on apps.shopify.com states whether the developer offers direct support in a given language. All nine apps researched here — including Nada and Smart Collection Pro, the two with a Japanese UI — are listed as not offering direct support in Japanese. The Languages field describes the interface, not the support desk.

Title sorting is not guaranteed to follow non-Latin-script collation rules. There is one Shopify Community thread, with a Shopify staff response, reporting that sorting a product list alphabetically in Dawn orders by the title in the store's default language rather than the currently active language. Whether locale-aware collation (for example, Japanese kana ordering) is applied is not addressed in official documentation, so this is unconfirmed. If your catalog uses non-Latin-script product titles, eyeball the actual output on your own data before you commit to Product title A-Z or Z-A.

Eight sort apps compared

Basics and cost first.

Item	Sort'd	Jedi	ST	SortWise	Nada	MB	KX AI Collections	Sortimize
Developer	Beefy Nachos Ltd (UK)	Jedi Apps (IN)	Green Apple Solutions (IN)	Mezereon Inc. (US)	Digismoothie (CZ)	merchbees (TR)	Kimonix (IL)	Podo Pals (US)
Built for Shopify	Yes	Yes	No	Yes	Yes	Yes	No	No
Rating / reviews	5.0 / 134	4.8 / 113	5.0 / ~236	5.0 / 22	4.8 / 23	4.8 / 140	5.0 / ~200	5.0 / 6
Launched	2019	2023	2021	2025	2021	2020	2020	2021
Free plan usable in production	Yes	Yes	Yes (conditional)	Yes	No	Yes	No	No
Paid entry point	$29/mo	$2.99/mo	$19/mo	$9/mo	$9.99/mo	$4.99/mo	From $19/mo	$9.99/mo
Top tier	From $350/mo, custom	$249/mo	$179/mo	$49/mo	$19.99/mo	$13.99/mo	From $299/mo	$64.99/mo
What drives the bill	Monthly sessions	Collection count	Sort runs plus order count	Products, collections, orders	Product count	Products, collections	Collection count plus average monthly orders	Collections plus products
Free trial	7 days	7 days	7 days	7 days	7 days	3 days	14 days	14 days
Japanese UI listed	No	No	No	No	Yes	No	No	No

Then run frequency and requested scopes.

Item	Sort'd	Jedi	ST	SortWise	Nada	MB	KX AI Collections	Sortimize
Run frequency as listed	Not stated	Not stated	Not stated	Hourly from the free plan	Real-time	Real-time on all plans	1–4x/day	1–4x/day; hourly on top tier only
Confidence in that frequency	Unconfirmed	Unconfirmed	Unconfirmed	Confirmed	Confirmed	Confirmed	Confirmed	Confirmed
Restores original position on restock	Unconfirmed	Stated	Unconfirmed	Unconfirmed	Handled as republishing	Stated	Unconfirmed	Unconfirmed
Metafield sorting	Yes (Premium)	Yes	Yes ($179 tier only)	Yes (paid)	No	No	Yes	No
A/B testing	Yes (from $70)	Yes	No	Yes (paid)	No	No	Yes	No
Sensitive customer data scopes	Behavioral data only	Name, email, phone, address	Not listed	Behavioral data only	None requested	None requested	Name, contact details, plus all-time order history, returns, shipping	None requested

What the tables actually tell you

The biggest hidden cost isn't the monthly fee — it's run frequency. KX AI Collections updates 1–4 times a day even on its $299/mo top tier, and Sortimize's entry plan runs once a day. On the other side, MB Push Down Hide Out of Stock states "All features work real-time" on every plan, SortWise runs hourly starting on the free plan, and Nada's listing says it reacts whenever inventory changes with no manual sync. If your stock moves minute to minute, daily batch apps don't meet the requirement; if turnover is slow — furniture, B2B — four runs a day does no real harm. Note also that KX AI Collections carries a "Real-time updates" feature tag while its pricing table says "Updates 1–4/day." Read the pricing table and the description, not the feature tags.

You can't compare monthly prices directly, because the metering axes differ fundamentally. A large catalog with moderate traffic favors Sort'd's session-based billing; high order volume across few collections reads better on Jedi. ST is the hardest to forecast: what counts as "1 Sort" can't be determined from the listing, so it's unconfirmed, and high-frequency operation is where that bites.

"Has a free plan" means eight different things. Nada's Free tier is limited to trial and development stores and can't be used in production. ST's Free is listed as limited to stores on the Shopify Basic or Professional plans with fewer than 250 monthly orders. MB, by contrast, is free in production up to 50 products and 100 collections, and SortWise gives you 10 collections, hourly sorting, and analytics for nothing.

Scope differences go straight to your security review. Nada, MB, and Sortimize request no order data and no customer data at all. KX AI Collections asks for all-time order history, returns, and shipping information, plus edit access to checkout pages and your theme. In organizations with strict personal data policies, permissions will narrow the field before anyone looks at features.

App notes

MB Push Down Hide Out of Stock

If "demote sold-out products" is the whole requirement, this is the best value here. You can also specify the sort method applied to in-stock products (best selling / price / new-old / manually), so "keep Best selling but push sold-out items to the bottom" is achievable in one app. The developer's changelog also documents extending the hiding behavior beyond the online store to multiple sales channels.

Two things in the official FAQ matter a lot. First, the app controls ordering by setting the collection's admin sort method to "Manually." Second, if you change that collection's sort method in the admin, the app automatically disables itself for that collection. If manual merchandising happens alongside this app in your store, build your operating rules around that behavior. Limits are 10,000 products and 300 collections; what happens on exceeding them is unconfirmed. The 3-day trial is also short.

Nada • Hide Out of Stock

Lists a Japanese UI and runs on a minimal scope that requests neither order nor customer data. It's the only app in this comparison advertising a 60-day money-back guarantee. Out-of-stock handling is the most granular here: a day threshold for "hide after N days out of stock," SEO-friendly redirects for hidden URLs, and hiding sold-out variants. There is no free plan for production stores. One 1-star review reports that automatic sorting is disabled every time a manual change is made outside the app — if your team reorders collections by hand often, test that conflict before you commit.

SortWise Collection Sort

The most generous free plan in this set: 10 collections, hourly automated sorting, sold-out demotion, analytics, and unlimited sort runs at no cost. Four billing axes are stated explicitly, so cost is forecastable. But the ceilings are 25,000 products and 10,000 monthly orders even on the top plan, and behavior on exceeding them is unconfirmed. Launched March 2025 with 22 reviews, so the track record is thin.

Jedi ‑ Collection Sort Pro

The cheapest paid entry point at $2.99/mo, billed solely on the number of managed collections. It lists restore-to-original-position on restock and nested sorting, and it's the only app here listing Claude / MCP / Sidekick integration. The caveat: it requests read scopes for customer name, email, phone, and address. Unit cost also climbs steeply with collection count ($30 at 250 collections, $80 at 1,000).

ST: Product & Collection Sort

Sorting by GA4 pageviews and add-to-cart rate is unique here, and it's effective for evaluating new products with thin sales history. It explicitly documents coexistence with Boost AI Search & Filter, Searchanise, and Search & Discovery, so it's less likely to conflict in stores that already run a search app. There's no Built for Shopify badge, and the free plan is listed as limited to stores on the Shopify Basic or Professional plans with fewer than 250 monthly orders.

Sort'd ‑ Prime Collection Sort

The elder statesman here, live since 2019, with an unlimited drag-and-drop merchandiser on the free plan. Because billing keys off monthly sessions alone, cost doesn't move as your catalog or collection count grows. The weakness is forecasting: the free plan's "Limited Sorting with Automations" doesn't state a concrete cap anywhere, so it's unconfirmed.

KX AI Collections

Formerly Kimonix. Its current listed name on the App Store is "KX AI Collections," so searching the old name mixes current and stale information. Feature breadth is the widest here — sorting on 100-plus parameters including return rate, margin, and size-level inventory, plus market-specific sorting strategies. But confirm three things before installing: updates run 1–4 times a day; the pricing table says "Plan STARTS at," so your real monthly cost isn't knowable in advance; and it requests the broadest scopes in this comparison.

Sortimize Collection Sort

The entry plan ($9.99/mo) updates once a day; hourly updates require the $64.99/mo Turbo tier. It's included here as the clearest illustration of run frequency as a hidden cost. Only 6 reviews, the most recent from December 2025 — not a base that supports conclusions.

Sidebar: when the 60-condition wall is the real problem

Smart Collection Pro (BazaarForge, from $5/mo) offers automated collection creation with unlimited conditions plus nested collections, addressing the 60-condition cap and the no-nested-AND/OR limitation head-on. It lists a Japanese UI, and its 21-day free trial is the longest in this research. But it has no sorting features. It complements a sort app; it doesn't replace one.

Before you install

• Can you live with sold-out products disappearing from the collection? If yes, native features are enough.
• How often does your inventory move? If it changes on a scale of minutes to an hour, daily batch apps are out.
• Of the billing axes — products, collections, orders, sessions, sort runs — which one grows fastest in your business?
• Have you confirmed the "free plan" is usable on a production store?
• Will the requested data access scopes pass your organization's privacy review?
• Have you specified UI language and support language as two separate requirements? All nine apps here are listed as not offering direct support in Japanese.
• Does anyone reorder collections manually in the admin? If so, have you tested that against the app's automatic sorting?
• Have you asked each vendor's support what happens to your product order after uninstall?
• If you plan to use title sorting on non-Latin-script product names, have you visually verified the output on real data?
• Do you have a test store or duplicated theme to validate in before this touches production?

Bottom line

Four decisions, in order.

First, update your assumptions. There are nine sort orders, and every collection created on or after April 15, 2026 defaults to Most relevant. Best selling counts all-time orders, so "recent best sellers" is not something native features can produce.

Second, check whether native is enough. If removing sold-out products from the collection is acceptable and your catalog is small, "Inventory stock is greater than 0" plus Match all conditions finishes the job. No app required.

Third, let go of the free official app. Search & Discovery's out-of-stock setting is search-only, and Shopify Flow cannot execute the reordering mutation. The moment you want to control collection ordering, it's a paid app or a custom build.

Fourth, filter on run frequency and billing axis, not monthly price. If your inventory moves fast, pick an app that explicitly states real-time or hourly operation. Pick one that doesn't bill on the metric that grows fastest for you, and your costs won't spike as you scale.

Every figure here reflects apps.shopify.com, help.shopify.com, and shopify.dev as of September 1, 2026. Pricing, review counts, and Built for Shopify badges all move — re-check the listing before you commit.