Every November, small shops publish a new holiday theme, a fresh homepage or a gift-guide menu, and find out a week later that the add-to-cart button sits below the fold on phones. With Black Friday weeks away, Shopify has a cheaper way to catch the obvious mistakes first: SimGym, an app that sends AI shoppers through your store before real customers see the change.
It is useful. It is also, in Shopify's own words, not a substitute for real buyers. Here is how to use it for a holiday redesign without trusting it more than it deserves.
What SimGym is
SimGym is a Shopify app in "AI Research Preview". It creates AI shoppers with human-like profiles that browse your store, open products and collections, add items to the cart and leave written feedback on the experience. Shopify says it built the shopper personas from patterns across purchases on its platform. It opened to all eligible merchants in March with no waitlist; you install it from the Shopify App Store.
It can do two things:
- Analyze a theme: a summary of the shopping experience, strengths, areas to improve, recommendations ranked by likely impact on conversion, and the AI shoppers' feedback.
- Compare two themes: your published theme against a draft theme. The "winner" is the one with the higher add-to-cart rate among the AI shoppers.
Before you start: will it run on your store?
Shopify's help page lists the requirements:
- A Liquid storefront (the normal Shopify theme setup). Hydrogen and headless stores are not supported.
- Shopify Network Intelligence turned on.
- The store is not in private (password) mode.
- At least one product listed, and for a comparison, at least two themes installed including the published one.
Cost: each month comes with an allowance of simulations that does not roll over. Past it, a single-theme analysis is $0.75 and a two-theme comparison $1.50. For a small shop testing a few holiday variants, that is pocket money compared with a weekend of lost sales.
A holiday-redesign routine
- Build the holiday version as a draft theme. Duplicate your live theme and make the changes there: banner, gift guide, new menu, sale badges. Do not edit the live theme.
- Run "Analyze a theme" on the draft first. Pick a focus area if your change is local: Home page, Products, Collections, Cart or Search. A new mega menu? Focus on Collections. A new product-page layout? Products.
- Read the shopper feedback, not just the score. The written comments are where you find things like "could not find the size guide" or "the discount was not clear until the cart". Treat each as a lead to check yourself on a real phone.
- Fix the obvious problems and then run "Compare two themes": live theme against the fixed draft.
- Read the comparison as directional. Shopify calls these "directional performance indicators". A clear gap in add-to-cart rate is a signal; a small one is noise. If both themes score zero, the help page says the result is inconclusive.
What it cannot tell you
Shopify is unusually honest about this. Its help page says results "might differ from actual buyer behavior" and that simulations work best for stores with existing products and content. Aaron Glazer, Shopify's director of product, told Modern Retail there is "no replacing humans", and that the only way to fully understand the impact of a change is an A/B test.
A few gaps worth keeping in mind:
- AI shoppers do not pay. The comparison is on add-to-cart, not orders. Shipping costs, delivery dates and payment options at checkout, where many real carts die, are not what the winner is picked on.
- They are not your customers. The personas come from Shopify-wide patterns. Your regulars, who know where everything is, and your gift buyers, who arrive from an Instagram ad on a phone, behave differently.
- No feelings about your brand. A theme that converts AI shoppers can still look off to people who love your shop for its look.
Then check with real people
Use SimGym to throw out the bad ideas cheaply, then confirm the survivor with real traffic:
- If you have the traffic: run a proper split test with an A/B testing app for a week or two before the big weekend, measuring orders, not clicks.
- If you do not: publish the new theme a couple of weeks before Black Friday, not the night before. Compare add-to-cart and conversion rates in Shopify Analytics against the previous weeks, and keep the old theme one click away in your theme library.
- Either way: ask three real people (a friend, a regular, someone who has never seen the shop) to buy a specific gift on their phone while you watch. Ten minutes each, and it catches what no simulation will.
SimGym earns its place as the first filter: fast, cheap and good at spotting buried buttons and confusing navigation. The final say still belongs to the people who actually pay.
