ShopOps weekly

You can now A/B test your entire theme - and a store went from 12 seconds to under 1

June 12, 2026

Hello there,

Google recently launched something called Preferred Sources - you choose which sites you want surfaced in your AI search results.

If you regularly search for Shopify-related topics, it's worth setting up. You get more results from sources you already trust, and less noise from sites you've never heard of.
One of those sources could be us.

Here's how to add shopexperts.com in 10 seconds:

  1. Go to google.com/preferences/source

  2. Search for shopexperts.com

  3. Check the box ✔

That's it. Next time you Google something Shopify-related, we're more likely to show up - labelled as a source you chose.

Genuinely useful for you. Also, obviously, good for us. Felt worth being upfront about both!

A few things this week that change how you can operate and build.

  • Shopify Rollouts - schedule, gradually roll out, and A/B test your entire theme or checkout. Native, no extra tools

  • Vibe coding meets Shopify - v0, Replit, and ManusAI now connect directly to Shopify stores

  • Wednesday's outage - if your store was down for a couple of hours, here's what Shopify said

  • Quick win - a store went from a 12-second load time to under 1 second. Here's exactly what was fixed

If this was forwarded to you, subscribe here 🔒

Shopify Updates You Can't Miss

1. You can now A/B test, schedule, and gradually roll out your entire theme or checkout

Shopify Rollouts now supports scheduling, gradual publishing, and A/B testing - not just for individual elements, but for your entire theme, checkout configuration, and customer account setup.

In practice, this means you can:

  • Schedule a theme switch to go live at a specific date and time

  • Activate a BFCM theme for one week and automatically revert when it ends

  • Roll out a new checkout or branding update to 10% of visitors before going all-in

  • Run a proper A/B test between two completely different themes or checkout setups

  • Test different CTAs or layouts per market, without touching your live store

Shopify also automatically creates a copy of your published theme when an experiment starts, so you can keep editing your live theme independently while the test runs. Multiple experiments can run simultaneously.

Why it matters: Most stores change things and hope. This removes "hope" from the equation. BFCM theme prep alone is worth the setup - but the real value is making checkout changes testable before they go live to everyone.

What to do: Go to Markets → Rollouts in your Shopify admin. Start with something low-stakes - a CTA text change on your checkout, or a layout variation on your customer account page. Get familiar with how experiments are structured before BFCM prep season starts.

2. Describe your store to an AI and have it built – v0, Replit, and ManusAI now integrate with Shopify

Three of the most popular vibe coding tools now connect directly to Shopify. Describe your business in plain language, they generate the store, and it connects to Shopify. From there, you're selling.

This builds on the same Universal Commerce Protocol infrastructure that opened to all developers a few weeks ago. Each new tool that connects is another way someone can build a Shopify storefront without touching Liquid or writing a line of code.

Why it matters: For merchants, this mostly matters indirectly right now — it lowers the floor for anyone building a new Shopify store, which accelerates what's possible on the platform. For founders considering building custom storefronts or tools, the path just got significantly shorter.

What to do: If you're planning a new market, sub-brand, or landing experience and want to prototype fast, these tools are worth exploring. Shopify.com/build-with-ai has the overview.

3. Shopify had a two-hour outage on Wednesday

Shopify experienced an infrastructure issue on Wednesday morning that affected admin access, checkouts, storefronts, POS, and support for approximately two hours. The issue was resolved the same morning.

If your store was offline or orders weren't processing during that window, that's why. No further explanation was provided publicly beyond confirmation that it was an infrastructure issue.

What to do: Review Wednesday's orders and sessions in Analytics to see if there's a gap. If you run subscription billing or scheduled automations, check that any jobs which should have fired during that window completed correctly afterward.

A Danish Cheese Shop Loaded in 12 Seconds. Then This Happened.

A Shopify store running a custom Booster theme. Content-heavy homepage. E grade on GTmetrix. 12.3-second load time. 337 requests. 6MB page size.

No visual changes were made. Everything stayed exactly as it looked.

Here's what was actually causing it:

  1. The hero image - the most important element on the page - was set to lazy load, meaning it was the last thing to arrive

  2. Multiple render-blocking scripts in the <head> were holding up everything else

  3. One app script was executing twice on every single page load without anyone knowing

  4. Trustpilot was spawning 10+ child requests the moment the page opened

After fixing the priority loading, deferring non-critical scripts, removing the duplicate, and loading Trustpilot only on user interaction:

Before → After

  • GTmetrix grade: E → B ✅

  • Load time: 12.3s → 975ms (↓92%)

  • Page size: 6.07MB → 678KB (↓89%)

  • Requests: 337 → 37 (↓89%)

  • LCP: 2.0s → 632ms (↓68%)

Same store. Same design. Just the things that were silently breaking it, fixed.

Every Week, New AI Tools Are Being Built Into Shopify. Here's How to Shape What They Become.

v0 builds stores. Voice AI handles support calls. AI agents manage customer queries. AI visibility tools track whether your products appear in ChatGPT results.

The founders building these tools are actively looking for merchants to talk to before they launch. App Store Research is where those conversations happen - short paid calls, 30 to 60 minutes, and you get early access and direct input into what gets built.

Over $1 million in incentives paid out to merchants on the platform. New projects go live daily.

A few open right now:

  • Voice AI for customer support - validate whether voice agents can handle support calls for your store ($150)

  • AI margin recovery in logistics - an AI tool that finds overcharges hiding in your shipping invoices ($250)

  • AI shopping assistant for enterprise - feedback on an AI assistant built natively for Shopify ($100)

  • Influencer marketing tools - what teams running creator programs are still missing ($100)

  • Apparel & fashion bundles - smart bundle widget built for fashion, activewear, and swimwear brands ($120)

  • Shopify POS staff goals - validate a goal-setting app built for POS retail teams ($150)

Takes under a minute to sign up. You pick the projects. You pick the time.

New Expert Listings

❋ Awais Javed - Front-End Developer · ⭐ 5.0 (18 reviews)

→ Custom theme development, Liquid programming, and performance optimization
→ GA4 + GTM setup, Klaviyo integrations, and conversion rate improvements
→ Speed and SEO overhauls - including Core Web Vitals and render-blocking fixes
→ AI-assisted workflow for faster turnaround and fewer revisions

A strong fit if your store has performance issues you suspect but haven't diagnosed - or if you need clean, precise development work delivered on time. (See the Quick Win above.)

❋ Has Merit - Shopify Agency · ⭐ 5.0 (2 reviews)

→ Conversion-driven storefront design and custom development
→ Brand identity and visual systems built for recognition and growth
→ CRO audits and data-grounded UI/UX improvements
→ Strategy, design, and development all in-house - no handoffs between teams

A strong fit if you're planning a redesign or A/B testing your storefront layout and want a team that thinks about brand and conversion as one problem, not two separate briefs.

Know someone who belongs in the directory? Send them our way.

❤️ Quick Pulse Check

This week's question: How do you typically roll out changes to your storefront?

A) Publish and see what happens - we don't A/B test
B) We test occasionally, usually for big changes
C) Testing is a regular part of how we make decisions

Reply with A, B, or C. (We read every response)

Last week we asked: "Have you ever run ads on Reddit?"

Nearly half have never tried it but are curious - which makes sense given last week's news. The 19% who tried it and didn't repeat is probably the most honest data point in the set. Attribution is genuinely hard on Reddit right now (a reader replied last week noting UTM stripping and mixed-model attribution as the real pain), so "didn't see results worth repeating" might partly mean "couldn't measure results accurately." Worth factoring in if you're evaluating it.

See you next Friday,

– Jon & the shopexperts team

Let us know how we did today (one click 👇)

Login or Subscribe to participate in polls.