Running Shopify in 2026: The Operator's Guide

Running Shopify well in 2026 comes down to a dozen decisions, and most merchants get them in the wrong order. The short version: pick the plan for checkout, B2B and team size, not for AI access, because every plan can sell in AI chats. Keep a theme unless you have an application and the engineers to own it. Run one app per job. Customize checkout at the edges. Use Markets before a second store, and Flow before another app. Then spend the time you saved on the product data that ChatGPT, Copilot, Google's AI Mode and Microsoft's Brand Agents all read.

This guide takes each decision in turn, the way a merchant would ask it, gives the answer we'd give, and links the article that goes deeper. Platform facts come from Shopify's and Microsoft's own documentation, listed under Sources. Where a number is our heuristic rather than a documented fact, we say so and show how we use it.

Which Shopify plan do you actually need in 2026?

Most brands between $100K and $10M a year need Grow or Advanced. The differences that matter are staff accounts, the fee Shopify charges when you use a third-party payment provider, and checkout customization. Agentic selling is not one of them: Shopify's pricing page marks "Sell in AI chats" as included on Basic, Grow, Advanced and Plus alike, and the same goes for Markets, local currencies, translation and local domains.

Read the table by your constraint, not your revenue. If you need more than five people in the admin, Grow is out. If you settle payments through a third-party provider, the fee row can outweigh the subscription difference, so run it on your own volume. If you need to change the information, shipping or payment steps of checkout, or you sell wholesale across more than three catalogs, that's Plus. If none of those bind, the cheaper plan is the right plan, and the money is better spent on catalog data.

The longer argument, including what Plus does and doesn't change for AI channels, is in do you need Shopify Plus for agentic commerce.

What is Shopify's Agentic plan, and who is it for?

Agentic plan

A Shopify plan for brands that sell on another ecommerce platform. It syncs their products into Shopify Catalog so they can sell through AI channels without moving their online store to Shopify.

The Agentic plan is not a tier above Plus. It is a way in for merchants who are not on Shopify at all. Shopify's plan page lists ChatGPT, Google AI Mode and Gemini, Copilot, Perplexity and the Shop app as channels, says no Shopify online store is required, and describes the model as paying when products sell rather than a monthly subscription. It doesn't publish the per-sale rate on that page, and it invites merchants to get updates on availability, so read the current terms before you model margin on it.

Shopify's help center adds the operational detail: the Agentic plan lets you set up agentic storefronts while keeping your existing online store, setup differs from a standard plan, and direct checkout in AI channels is off by default. Shopify's Hydrogen documentation also says Oxygen hosting isn't available on Agentic plans, so it is not a cheap route to a headless Shopify storefront.

Who it suits: a WooCommerce, Magento or custom-site brand that wants AI-channel distribution now and a platform decision later. Who it doesn't: anyone already on a paid Shopify plan, because agentic storefronts are already part of what you pay for. The wider picture of who ships what across AI channels is in the agentic commerce guide for Shopify.

Should you use a Shopify theme or go headless?

Use a theme. For the large majority of stores the theme question is the real one, and the answer turns on three behaviours rather than looks: does the theme server-render price, stock and specs as HTML, does it read metafields natively, and does it leave headroom on Core Web Vitals once your apps are installed. Lean Online Store 2.0 themes in the Dawn family pass all three out of the box, which is why they are a sound default. A premium theme earns its fee when it adds merchandising you need, such as configurators or complex variant logic, without breaking those three.

Headless means replacing the Liquid theme with your own frontend while Shopify keeps the admin, catalog, payments and checkout. Hydrogen projects are React Router apps preconfigured for Shopify, and Oxygen, Shopify's edge host, is included at no extra charge on paid plans from Starter to Plus. The licence is free; the engineering team is not.

The theme criteria are worked through in Shopify theme choice for the agentic era; the headless test, and the Oxygen limits to budget for, are in headless Shopify: who actually needs Hydrogen.

How many apps should a Shopify store run?

Fewer than you have. Each app that injects storefront code adds weight to every page, and app-rendered widgets that appear only after JavaScript runs are content that crawlers and shopping agents reading the raw page may never see. App bloat is a speed problem and a legibility problem at once.

Audit by job, not by app name. Write down the jobs the store needs done (reviews, subscriptions, search, email capture, upsell, analytics, shipping rules), map each installed app to the jobs it does, and look for jobs done twice. Then cut in three passes: apps nobody has opened in months, overlap merged into one app per job, and anything a theme setting, a metafield or a native Shopify feature can do for free. Flow, Search & Discovery and Shopify's own customer reports replace more paid apps than most stores realise.

One trap: uninstalling an app doesn't always remove what it added. Old theme edits, snippets and script tags can outlive the app by years, so check the theme code after every removal.

The full method, with job categories and the usual overlaps, is in the 6–8 app rule: a Shopify app audit method.

How much should you customize Shopify checkout?

As little as gets the order placed. Customize the edges of checkout (a trust line, shipping-method clarity, a relevant add-on, a delivery-date field you actually use) and leave the core flow and the payment step alone. Every added field, script or app is a new place for an order to fail.

What you can customize depends on the plan. Shopify's help center says standard checkout customizations are available on Basic and higher, with apps limited to the Thank you and Order status pages; apps that customize the information, shipping and payment pages, and the Checkout Branding API, are Plus only. The old way is closing: Shopify made the additional scripts box view-only on August 28, 2025, and August 26, 2026 was the deadline for non-Plus stores to upgrade their Thank you and Order status pages.

AI channels add a second reason for restraint. Where a shopper pays inside Copilot or Google's AI Mode, they use Shopify-powered direct checkout, and heavy web-checkout customization does not travel with them. A checkout that only converts because of its custom blocks will convert worse in conversation.

• Keep: delivery estimates, clear shipping options, one relevant add-on, trust and returns copy.
• Question: extra form fields, multiple upsell apps, anything that loads a third-party script in checkout.
• Leave alone: payment methods order and the payment step itself.

The scoping test and the customizations that consistently earn their place are in checkout extensibility: what to customize, what to leave.

Should you use Shopify Markets or open a separate store?

Start with Markets. Shopify's international tools let you create markets for countries or regions and manage them from a single store, with local currency pricing, translation, local domains and duties handling, and the pricing page lists all of those as included on every plan (local currency prices and local payment methods require Shopify Payments). For the first two or three expansion countries, one admin, one catalog and one inventory pool beat two of everything.

Managed Markets is the middle path when duties and tax registration are the problem. It is powered by Global-e, which becomes the merchant of record for international orders: it registers for and remits tax, accepts local payment methods, and manages duties and import tax. Shopify limits it to merchants based in the continental United States and certain stores in Canada and the United Kingdom. Check the fees in your admin before you assume it pays for itself.

A second store is the honest answer in three cases: a separate legal entity has to sell in that country, the catalog genuinely conflicts (different products, compliance rules or brands), or a different team runs that business end to end. Unfamiliarity with Markets is not on the list.

Decision tables for each path are in Shopify Markets vs a separate store.

What should Shopify Flow automate first?

The repetitive decisions your team makes the same way every time. Shopify Flow is free on Basic, Grow, Advanced and Plus. A workflow is a trigger (an order is created, inventory changes, a product is added), conditions that filter it, and actions that run. The Send HTTP Request action needs Grow or higher; tasks from custom partner apps need Plus; usage limits follow your plan's API limits.

1. 01

   Hold high-risk orders

   When an order is flagged high risk, hold fulfillment, tag it and alert the person who reviews fraud.

2. 02

   Alert on low stock

   When a variant falls below the reorder point, message purchasing with the SKU and current quantity.

3. 03

   Tag customers and orders

   Tag VIPs, wholesale buyers and first orders so segments and reports stay clean without manual work.

4. 04

   Keep new products tidy

   When a product is created, check required metafields and channel publishing before it goes live.

5. 05

   Send a daily out-of-stock digest

   One message with every sold-out product still published, instead of finding them from customer emails.

Flow is glue, not a fix. It won't repair a broken catalog or a bloated app stack, and checkout logic belongs in Shopify Functions. Before you buy an app to move data between two tools, see whether a Flow workflow, or an HTTP request on Grow and up, already does the job.

The five workflows, plan gates and a 14-day rollout are in Shopify Flow automations that actually cut ops hours.

How fast does a Shopify store need to be?

Core Web Vitals

Google's three page-experience metrics. A good score is Largest Contentful Paint within 2.5 seconds, Interaction to Next Paint of 200 milliseconds or less, and Cumulative Layout Shift of 0.1 or less, measured at the 75th percentile of page loads, separately for mobile and desktop.

Fast enough to pass Core Web Vitals on mobile field data, which is where your customers are. Those thresholds are published by Google on web.dev, and a disciplined Shopify theme can meet them. When a store is slow, the cause is almost always what got layered on top: app scripts, oversized images, third-party pixels, page-builder sections. Shopify's hosting is rarely the bottleneck.

Speed now has a second audience. AI assistants and shopping agents fetch pages on tight time budgets, and many read the HTML without running much JavaScript. They don't score your Core Web Vitals, but the same script-heavy causes that fail Core Web Vitals also hide facts from agents. One fix serves both.

Fix in this order, because each step is cheaper than the next and often makes it unnecessary: the app audit, then images (right-sized, modern formats, the hero image not lazy-loaded), then pixels and tags, then the theme itself. Measure on field data from real users, not a single desktop lab run, and re-check after every app install.

The full sequence is in Shopify speed and Core Web Vitals in the agentic era.

How do you migrate to Shopify without losing rankings?

Rankings are lost to broken URLs and missing redirects, not to the platform switch. A migration that keeps every indexed URL resolving to its closest equivalent with a 301 keeps the bulk of its search equity. The work is inventory and mapping, done before anything goes live.

1. 01

   Inventory every URL

   Crawl the old site and pull URLs from analytics, Search Console and backlinks. You can't redirect what you never listed.

2. 02

   Map and import redirects

   Map each old URL to its new Shopify path and import the CSV under Content → Menus → URL redirects.

3. 03

   Keep structured data at parity

   Product, offer, review and organization markup should say at least what the old site said, with the same facts.

4. 04

   Move content and metadata

   Titles, descriptions, alt text, specs and policies move with the products, not after launch.

5. 05

   Monitor daily for two weeks

   404s, redirect chains, Search Console coverage and revenue by landing page, every day while problems are cheap.

Know the platform's limits before you map. Shopify allows up to 100,000 URL redirects on standard plans and 20,000,000 on Plus, redirects only fire for URLs that would otherwise return a 404, and some paths, such as /products and /collections/all, can't be redirected. AI citations are equity too: a migration that drops the specs and policies assistants quoted can lose recommendations even when rankings hold.

The four-part checklist and launch-week plan are in migration to Shopify without losing rankings or data.

Is Shopify's built-in search enough, or do you need AI search?

Shoppers who use site search are your highest-intent visitors, so search that fails costs you your best traffic first. Before you buy anything, check why it fails. Bad results are usually a data problem: missing attributes, inconsistent product types, no synonyms for the words customers actually type. A smarter engine on top of a thin catalog returns smarter-looking wrong answers.

There are three routes. Shopify's built-in storefront search is on by default and AI-powered, and the Search & Discovery app adds custom filters, product boosts and related-product recommendations; together they suit most stores whose catalog is well structured. A third-party AI search app adds semantic matching and merchandising controls when the catalog is large or the vocabulary is messy. A search API engine is for headless builds. Whichever you pick, treat the zero-results log as a weekly job: every query that returns nothing is a shopper telling you what to stock, name or tag.

Search and agents are complements, not rivals. AI search retrieves; an on-site agent such as Microsoft's Brand Agents advises and can guide a shopper to checkout. Both read the same structured catalog, so the attribute work pays twice.

The setup and the weekly discipline are in on-site AI search on Shopify: a setup that converts.

What should Klaviyo get from your store and your agents?

Structured signals, not raw transcripts. Klaviyo already receives orders, products and browsing events from Shopify; the new question is what to do with what shoppers tell an on-site agent. The useful parts are small and structured: the use case they described, a stated preference (size, fit, budget, frequency), and any need the agent couldn't meet. Captured as profile properties or events, those decide which flow a customer enters. Pasted verbatim into a subject line, they feel like surveillance.

Two cautions before you build. First, Microsoft requires Brand Agents merchants to add a privacy disclosure saying conversation content is shared with Microsoft, so update your policy before you route anything onward. Second, Microsoft's documentation describes the agent dashboard and data downloads, not a native Klaviyo connector, so confirm what your setup can actually export before you design flows around it.

The data flow, what to capture and where the creepy line sits are in Klaviyo and Brand Agents: the data flow that works.

How do Shopify products get into ChatGPT, Copilot and AI Mode?

Through Shopify Catalog and agentic storefronts, and for eligible stores it's already switched on. Shopify's Spring '26 Edition, published June 17, 2026, describes Catalog as a structured product dataset AI agents can search, with eligible merchants included by default, and the Universal Commerce Protocol (UCP), co-developed with Google, as the open standard agents use from discovery to checkout. Shopify's help center says agentic storefronts are active by default for eligible stores and are managed under Sales channels → Agentic.

Agentic storefronts

Shopify's feature that lets shoppers discover and buy your products inside AI channels, currently ChatGPT, Google AI Mode and Gemini, Microsoft Copilot, and Meta surfaces. Products reach them through Shopify Catalog.

Brand Agents is the one that confuses people, because it shows up in Shopify stores but isn't Shopify's. Microsoft describes it as a conversational agent on your own site, in beta, for Shopify and WooCommerce. You install it through the Clarity app, and accepting its terms also publishes your catalog to Microsoft Merchant Center for Bing and Copilot. No Shopify plan is required. Our Brand Agents explainer covers it in full.

None of these channels can recommend what they can't read. The work on your side is the same everywhere: specs in metafields rather than prose, facts and JSON-LD rendered on the server, and completed policies. The build is in building a Shopify product page AI agents can read.

How do the decisions fit together?

Read it from the bottom. The plan sets your limits. Catalog data is what every layer above it, including every AI channel, reads. Apps and Flow decide how much code and manual work sits on top of that data. The storefront and checkout are where humans convert, and the channels are where both humans and agents find you. Problems at a higher layer are usually caused lower down: slow pages by apps, weak AI recommendations by thin data, bad search by both.

That's why our order of work runs bottom-up. It's a sequencing heuristic, not a law, and a store with a broken checkout fixes the checkout first.

1. 01

   Week 1: confirm the plan and the channels

   Check your plan against the limits table and open Sales channels → Agentic to see which AI channels are already live.

2. 02

   Weeks 1–2: run the app audit

   Map apps to jobs, cut overlap, and remove leftover code from apps already gone.

3. 03

   Weeks 2–5: fix the catalog facts

   Move specs into metafields, complete policies, and make sure the product page renders facts and JSON-LD on the server.

4. 04

   Weeks 3–6: fix speed on mobile field data

   Images, pixels, then theme. Re-measure after each change.

5. 05

   Weeks 4–8: wire Flow and trim checkout

   Build the five Flow workflows; remove checkout customizations that don't help the order complete.

6. 06

   Weeks 6–12: expand and retain

   Add markets, tune search from the zero-results log, and route structured agent signals into Klaviyo.