Product Discovery for Multi-Market Shopify Plus

Key Takeaways

Most search platforms require a separate index per market or locale. Adding a region multiplies your search infrastructure, merchandising rules, and operational overhead.
Single-index multilingual search, per-market pricing, and availability filtering are the three architectural requirements that determine whether international expansion is a configuration change or a rebuild.
Shopify Markets handles catalog localization. Your search platform must match that localization in results, ranking, and merchandising.

You expanded to Canada, the UK, and Australia on Shopify Markets. Your catalog shows local prices and availability. Your search results do not.

A shopper in Sydney types a query and sees USD prices in autocomplete. A shopper in London gets results for products you do not even ship to the UK.

The storefront looks right. The discovery experience behind it is still running on your home market.

That gap is not a bug in your theme or your currency settings, and it is not something a developer can patch with a few lines of Liquid, because it is a structural gap between what Shopify Markets localizes, the catalog, and what your search platform ignores, everything else.

On the onboarding calls we run with Plus brands, the pattern repeats. The currency and the translated content landed fine. Then someone searches the store from a non-US IP and the whole discovery layer acts like there is only one market.

Autocomplete shows USD, results include products the shopper cannot buy, and sort order is whatever works in the US, applied everywhere.

Here is what most multi-market expansion plans miss: the conversion gap between your home market and new markets is not always a market maturity problem.

On our onboarding calls, we consistently see a significant portion of that gap trace back to the discovery layer, not the market itself. A UK shopper who encounters USD prices in autocomplete does not file a support ticket. She leaves.

That cost shows up as "the UK market is harder" in the quarterly review, when the actual cause is a search platform that ignores her market context.

The evaluation criteria for multi-market product discovery are fundamentally different from single-market. Most vendors treat international as an afterthought, because their architecture was never built for it.

What follows is the evaluation framework. Can your discovery platform keep up with your expansion, or does every new market mean another rebuild?

What does Shopify Markets actually give you, and where does it stop?

Shopify Markets handles the catalog side of international selling: per-market pricing, product availability, localized content, and tax-inclusive pricing. It does not handle the discovery side. Search results, collection rankings, autocomplete, facets, and recommendations all ignore which market the shopper is in. Your catalog is localized, but your search experience is not.

What Markets resolves per market:

Per-market pricing. Each variant carries its own price per market and country, in that country's native currency.
Currency conversion. Fixed-amount values convert between currencies at checkout.
Product availability. You control which products are published to which markets.
Localized content. Translated themes, localized domains, market-specific languages.
Tax-inclusive pricing. Market-specific tax configuration.

All of that is real and important. It is also where Markets stops. And the line between "what Markets handles" and "what Markets does not" is where most international expansion projects quietly stall.

What Markets does not touch:

Search results. A query returns the same ranked products regardless of which market the shopper is in.
Collection page ranking. Sort order on collection pages does not change per market.
Autocomplete suggestions. Predicted results show the same products, in the same currency, for every shopper.
Facet values. Price facets, availability facets, and attribute filters are not market-scoped.
Recommendation logic. Product recommendations do not adjust by market context.

Shopify's native search has expanded its language coverage with semantic search, which helps shoppers search in natural language across multiple languages. But native search still has no per-market merchandising, no market-scoped ranking, and no conditional rules by region.

The consequence is straightforward. Your catalog is localized, but your discovery experience is not.

A shopper in the UK sees a localized storefront and then encounters a search experience that behaves like they are shopping in the US. NNGroup puts the point bluntly: to win in an international market, sites "must go beyond translation and localization."

Stripe makes a sharper version of the same point about pricing: showing amounts in a foreign currency "immediately introduces problems," because shoppers have to mentally convert and start worrying about hidden fees.

If that friction exists in your search results before a shopper even reaches checkout, you are losing them at the top of the funnel.

This is the gap your search platform has to close. Markets gave you the catalog, the localized prices, the translated content, and the per-market availability rules. Discovery, the part that decides which products a shopper actually sees when they search, browse, or receive a recommendation, is the part you still need to solve.

Why does the single-index vs. multi-index architecture matter?

The architecture your search platform uses to handle multiple markets determines how much operational work every new region costs you. Two approaches exist: one index per market, or a single index with conditional market resolution. The difference compounds fast because every merchandising rule, sort order, and synonym configuration either multiplies with each new market or stays at one.

Three mechanisms make the architecture decision matter more than any individual feature, and each one compounds quietly until the cost of adding a new market becomes the cost of rebuilding your entire discovery stack.

The market multiplication problem. A multi-index architecture requires a separate search index per market or locale. Three markets means three indexes. Each needs its own merchandising rules, sort orders, and synonym configuration. If you run 50 collections, three markets means 150 collection configurations to build and maintain. A single-index architecture with conditional overrides keeps it at 50, because the same configuration adapts per visitor context.
Operational cost compounds quarterly. Merchandising is not set-and-forget. Sort orders shift with seasons, promotions rotate, new products launch. At two hours per collection per quarter for merchandising updates, the multi-index approach costs roughly 300 hours per quarter across three markets. The single-index approach costs roughly 100, because the work happens once and the conditions handle the rest.
Drift between indexes is invisible until it hurts. When each market has its own index, there is no guarantee they stay consistent. A merchandising rule updated in the US index might not get replicated to the UK index for weeks. A product added to one index might be missing from another. The bugs are silent; shoppers just see worse results, and nobody files a ticket because they do not know what they are missing.

We built our Shopify Markets integration on a single unified index that dynamically resolves market pricing and availability per request, so every query, whether from a shopper in Tokyo or Toronto, hits the same index and returns prices in the correct local currency automatically. Your catalog lives in one place.

When a shopper's country resolves to a market, we apply that market's pricing and filter to that market's available products, all from the same index. One set of merchandising rules. One set of sort orders.

Market-specific behavior comes from conditions on those rules, not from duplicating the entire configuration.

The practical difference: adding a fourth market is a configuration change that takes minutes, not a migration that takes weeks of engineering time and merchandiser effort to replicate rules across indexes.

The market resolution itself is lightweight. The shopper's country, carried as context.geo.country in ISO 3166-1 alpha-2 format, resolves against the synced region table to find the right market. Our JavaScript SDK and storefront pixel populate this automatically via geolocation.

If the country is unspecified or unmappable, the system falls back to the primary market with default pricing and full product availability, so no shopper ever encounters an error page or an empty result set because of a geolocation edge case. No errors, no empty results.

What are the five capabilities a multi-market discovery platform needs?

Five capabilities separate a search platform that genuinely works across borders from one that works in your home market and pretends elsewhere: per-market pricing in results, market-aware availability filtering, multilingual single-index search, conditional merchandising rules, and unified cross-market analytics. Miss any one and the discovery experience degrades for every market beyond your first.

Per-market pricing in search results and autocomplete. Every product result, every autocomplete suggestion, every recommendation must show the price in the shopper's local currency at the market-specific amount. Not a converted estimate. The actual price Shopify Markets assigned to that variant for that market and country.

We resolve this automatically through our Shopify Markets sync: the response carries each variant's market-resolved price along with currency_code and currency_decimal_places in _meta.pricing, so the storefront formats the number correctly without any client-side conversion logic.
Market-aware product availability filtering. A shopper should never see a product they cannot buy in their market. In Strict mode, we filter results to products published to the shopper's resolved market. Products not available in that market are excluded from search, browse, and autocomplete entirely.

We also expose available_market_ids as a filterable catalog attribute, so merchandising rules can reference which markets a product belongs to. The alternative, showing unavailable products and hoping the shopper does not notice, erodes trust the moment someone tries to add to cart and discovers they cannot.
Multilingual search from a single index. Your German shoppers search in German. Your Japanese shoppers search in Japanese. Your UK shoppers search for "trainers" while your US shoppers search for "sneakers." A multi-market search platform needs to handle all of this from one index, not a separate index per language.

We support 110 languages with automatic language detection, so the engine identifies what language the shopper is searching in and returns relevant results without separate language-specific configurations. Our user intent processing handles regional variants automatically: a UK shopper searching "trainers" sees athletic shoes, while the same query from a US shopper expands to include "sneakers."

If you want to test how your current search handles multilingual queries across your markets, our autocomplete and search results audit walks through the exact test you can run on your store in 15 minutes.
Per-market merchandising without duplicating rules. You should not need to rebuild your merchandising for every market. Segmented sorting lets one sort order behave differently by region. A conditional sort expression activates when geo.country matches a specific value, with an unconditional fallback for everyone else.

Segmented metrics go further: they automatically surface region-specific performance data per visitor, so a product that converts well in the UK ranks higher for UK shoppers without you creating a separate UK sort order. The smoothing factor controls how much weight goes to local performance versus global, so you can tune the balance as your regional data matures.

David Cost, VP of eCommerce at Rainbow Shops, described the shift away from maintaining separate configurations:

"Layers was the first time we were able to create the kind of sort orders we were used to having in Salesforce. We saw a pretty immediate impact on conversion rate."

For a retailer operating across markets, that impact multiplied because the sort orders adapted by region automatically, without duplicating the configuration for each one.

Unified analytics across markets. You need to see how search performs per market without running separate analytics per index. LayersQL provides cross-market analytics through geographic segmentation, so you can break down search performance, conversion, and revenue by market, country, or region from a single query interface. One dashboard. One query language. No stitching together reports from separate tools.

How do you evaluate a vendor's multi-market readiness?

Eight questions cut through the pitch deck and reveal whether a vendor's multi-market support is native architecture or a marketing checkbox. Ask each one during a live demo, not a slide deck.

A strong answer demonstrates the capability on real data with a non-primary market context applied. A weak answer defers to documentation, a future roadmap, or a caveat about how "we're working on that for Q4" which tells you everything you need to know about whether the architecture was built for multi-market or bolted on after the fact.

"How many indexes do I need for five markets?" A strong answer is one. A weak answer involves the word "replica" or "per-locale."
"If I add a sixth market next quarter, what work does that create?" Strong: enable the market in Shopify, confirm sync, done. Weak: provision a new index, migrate merchandising rules, rebuild sort orders.
"Show me a search result with market-specific pricing in a non-primary currency." Strong: the response carries the exact market-resolved price and currency metadata. Weak: the vendor converts client-side, or worse, shows the primary market price.
"How do you handle a shopper searching in a language my catalog is not translated into?" Strong: automatic language detection across 100+ languages from a single index, with cross-lingual retrieval. Weak: requires translated catalog content per language, or supports fewer than 20 languages.
"Can I set a different sort order for the UK without duplicating my US sort order?" Strong: conditional expressions on the same sort order, scoped by geo.country or market handle. Weak: create a separate sort order, apply it to a separate collection, maintain both.
"How do you prevent out-of-market products from appearing in results?" Strong: availability filtering at the engine level, configurable per mode. Weak: relies on frontend filtering or requires manual exclusion lists.
"Where do I see search performance broken down by market?" Strong: one analytics interface with geographic segmentation built in. Weak: export data per index, build your own reports.
"What happens when a shopper's country does not map to any configured market?" Strong: graceful fallback to the primary market with default pricing. Weak: errors, empty results, or undefined behavior.

If a vendor cannot answer questions 1 through 3 with confidence, the rest of the conversation is academic. Those three reveal whether the architecture was designed for multi-market or retrofitted for it. For the full set of search evaluation criteria beyond multi-market, see our AI search guide for $10M+ Shopify Plus brands.

One pattern worth watching for: a vendor that answers "yes" to everything but cannot show you a live demo with a non-US country context applied. Claims are easy. Proof is not.

The proof is the API response carrying the right currency code and the right price for a shopper in Berlin, with the correct VAT-inclusive amount, the correct number of decimal places, and products filtered to what is actually available in the German market.

Ask to see it live, on the call, with a non-US country context applied to your own catalog data. If the answer is "we can set that up," you have learned something about how native the capability actually is.

Baymard Institute has documented more than 700 search usability issues across multi-million-dollar sites. A recurring theme: query language a store cannot parse drives shoppers to leave rather than retry.

In a multi-market context, that problem multiplies because regional terms, language variants, and currency mismatches compound across every market you serve, turning a single search UX gap into a conversion leak that scales linearly with your international footprint. The evaluation questions above are designed to surface those issues before they reach your shoppers.

What does multi-market discovery look like in practice?

Consider a Shopify Plus fashion brand selling in the US, UK, Canada, and Australia, four markets with four currencies, overlapping but not identical product catalogs, and shoppers who expect every touchpoint from autocomplete to checkout to reflect their local market.

This is a common setup, and it is where multi-market discovery, from text search to visual discovery, either works or visibly breaks.

For a shopper in London: She searches "jumper" on the UK storefront, and the engine detects the query language, understands "jumper" as the UK term for a pullover sweater, and returns relevant knitwear with every result showing GBP pricing at the UK market rate. Products not published to the UK market are excluded entirely. She never sees a product she cannot buy, and she never sees a USD price.

For the merchandiser: She manages one set of collections and one set of sort orders, which means a single change to the knitwear collection's sort logic automatically adapts for every market without her touching a second configuration. A conditional expression on the knitwear collection sorts by UK conversion rate for UK shoppers and by US conversion rate for US shoppers. She did not build a separate collection for each market. When the brand expands to Germany next quarter, she enables the market in Shopify, confirms sync, and watches the existing sort orders adapt automatically because the segmented metrics already handle geographic relevance, meaning a product that converts well in Germany starts ranking higher for German shoppers without a single rule change. No new configuration. No duplicated sort orders.

For the analyst: She opens LayersQL and queries search conversion by market for the last 30 days. One query, four markets, one table. She spots that "jumper" converts at 4.2% in the UK but "sweater" converts at 3.1% in the US for the same product category, and uses that insight to inform a merchandising decision.

This is what we built. The Shopify Markets integration handles the market resolution and pricing, so the merchandiser never touches a currency conversion and the shopper never sees a price that does not belong to their market. Segmented sorting handles per-market merchandising. 110 languages handle the multilingual search. One index. The same architecture that powers AI search for single-market stores.

The underlying mechanics stay out of the way. Market data syncs through webhooks (markets/create, markets/update, markets/delete, plus product updates), so when you add or modify a market in Shopify, we pick up the change without a manual refresh.

The three modes, Off, Pricing Only, and Strict, let you control how aggressively market context is applied. From no filtering to full availability enforcement, you choose the level that fits your expansion stage.

If you are already running multi-market on Shopify Plus and want the tactical setup, we wrote a full walkthrough in Multi-Market Merchandising on Shopify Plus.

It covers market modes, conditional sort expressions, regional language handling, and a per-market QA grid you can run on your store this week to verify that every market sees the right prices, the right products, and the right language handling before you send paid traffic to it.

What should you do next?

Adding a market should be a growth decision, not an infrastructure project, and if your current search platform makes every new region feel like a migration where you re-provision indexes, duplicate rules, and rebuild sort orders from scratch, the architecture is the problem.

One vendor call with the evaluation questions above will tell you whether you are on a single-index platform or paying the multi-index tax on every region you add.

Run them against your current vendor. If the answers involve separate indexes, manual rule duplication, or frontend currency conversion, you know where the ceiling is.

The brands we work with did not choose us because of a feature checklist; they chose us because adding market five felt exactly like adding market two. Enable it in Shopify, confirm sync, move on. That is the bar.

If fashion merchandising across markets is your specific challenge, our fashion merchandising guide covers visual discovery, margin-aware ranking, and seasonal velocity.

We built our search and merchandising for exactly this problem: one index, one set of rules, automatic market resolution, 110 languages. If you want to see what that looks like on your catalog, the fastest path is a conversation.