Hat Country had a catalog problem that didn't show up in the numbers.
After 15 years in business, the store had grown from a hat shop into a full western lifestyle retailer: 1,000+ products, 50,000+ customers, and brands like Stetson, Resistol, Bullhide, and Ariat across hats, boots, belts, jackets, jeans, jewelry, and accessories for men, women, and children. By product count and category breadth, this was a complete store.
Most customers still bought a hat and left. The product range had expanded; the shopping experience hadn't. That gap was costing Hat Country revenue on every order.
This case study covers what changed when AI product recommendations were built into the store: a 23% AOV lift, 40x ROI, and nearly 1 in 10 sales influenced by recommendations.
About Hat Country
Hat Country is a Shopify Plus retailer with over 15 years in western wear. What started as a hat-focused shop grew into a comprehensive western lifestyle destination, with 1,000+ products and 50,000+ customers spanning hats, boots, jackets, jeans, belts, buckles, jewelry, and accessories across men's, women's, and children's categories. The catalog naturally supports outfit-level purchases, but for most visitors, the store was still a hat shop.
The challenge: a full catalog, invisible to its own customers
Most buyers thought they were in a hat shop
Hat Country's brand identity hadn't caught up with its catalog. Customers arrived looking for cowboy hats; that was the intent, the search query, the reason they clicked through. They found a hat, bought it, and left. Boots, belts, and jackets sat in the catalog unseen.
Navigation helps, but it doesn't close this gap on its own. A customer who came for a specific Stetson isn't going to click through category menus to find the boots that pair with it. That connection has to be made for them, at the moment it's relevant.
Single-item transactions were the norm
The catalog supports outfit-level purchases. A cowboy hat pairs with boots, a belt, a jacket. The products were there to support a $250 order instead of a $75 one, but without cross-selling infrastructure, most transactions stayed at a single item. Customers weren't indifferent to the rest of the catalog; they simply had no way to discover it.
Manual curation couldn't scale
With 1,000+ SKUs across dozens of categories and multiple major brands, manually maintained product pairings weren't sustainable. New products launched without cross-sell connections, seasonal changes broke existing pairings, and even when curation worked, it couldn't adapt to individual customer behavior.
The solution: AI recommendations across every touchpoint
PersonalizerAI built recommendation models trained specifically on Hat Country's catalog structure, order history, and customer behavior patterns, not off-the-shelf templates applied from another store. The implementation covered four touchpoints.
1. Similar Items — product pages
When a customer views a hat and it isn't quite right (wrong brim width, wrong brand, wrong price point), Similar Items surfaces alternatives with comparable style, material, or features. The customer stays on the site instead of bouncing and finds the right product faster. The same logic applies across every category: alternative boot styles, different buckle options, comparable jacket cuts.
2. Complete the Look — product pages
When a customer views a cowboy hat, Complete the Look surfaces the boots, belt, jacket, and bandana they didn't know Hat Country carried. Most customers don't add every suggested item to cart right then. But they come back knowing Hat Country carries more than hats, and that shift in perception is the point. The widget functions as a catalog discovery tool as much as a sales driver. It's specific and visual: this hat, with these boots, this belt. Not a generic "you might also like" row.
3. Recommended For You and Best Sellers — homepage
Returning visitors see recommendations based on their previous browsing and purchase behavior. A customer who bought a hat last time might see boots on their next visit, a natural continuation of building out their western wardrobe.
New visitors see curated best sellers that establish two things quickly: the quality of the catalog and the breadth of what Hat Country carries. Both matter for first-session conversion. Recommendations on the homepage set the expectation that this is a full western lifestyle store before the customer has clicked on a single product.
4. Checkout Recommendations — Shopify Plus
Hat Country's free shipping threshold is $99. A customer with a $75 hat in cart sees a recommended $30 hatband at checkout. Adding it clears $99. From the customer's perspective, they're saving on shipping. From Hat Country's perspective, AOV just went up.
The recommendations are relevant to cart contents, the threshold creates a natural incentive, and the customer makes the add-on decision themselves. It reads more as useful information than an upsell.
Note: Checkout recommendations via Checkout Extensibility are available exclusively on Shopify Plus.
The results
+23% AOV lift. Average order value increased through cross-category discovery and checkout upselling. Customers who had previously bought single hats started adding boots, belts, and accessories, either in the same session after seeing Complete the Look, or on return visits after discovering the full catalog.
40x ROI. Every dollar spent on PersonalizerAI returned forty. The implementation covered its cost within weeks, and the returns compound as the AI refines its understanding of Hat Country's customers over time.
9.5% of total sales influenced. Nearly 1 in 10 sales involved a PersonalizerAI recommendation, making it a core revenue channel across every touchpoint in the customer journey, not occasional incremental lift.
"We’ve had an outstanding experience with Personalizer AI. Shaswat went above and beyond for us, building custom recommendation models tailored to our catalog and customer behavior. That level of support is rare, and it made a real difference.
The results speak for themselves. Our average order value climbed, and we’ve seen roughly a 15% lift in revenue since implementing the tool. It’s been reliable, easy to work with, and surprisingly impactful for how quickly we got everything up and running.
If you’re looking for a recommendation engine backed by someone who actually listens, delivers, and iterates with you—not just an app you install and forget—this one’s worth your time."
— Mike, Owner at Hat Country
What a 23% AOV lift means at different revenue scales
A 23% lift in average order value means additional revenue from every single order, extracted from existing traffic without acquiring new customers. At different store sizes, the math looks like this:
Monthly revenue baseline | Additional monthly revenue |
|---|---|
$50,000 | +$11,500 |
$100,000 | +$23,000 |
$500,000 | +$115,000 |
$1,000,000 | +$230,000 |
Based on Hat Country's verified 23% AOV lift applied to different revenue scales. Actual results vary by catalog, traffic, and vertical.
What other Shopify merchants can take from this
The Hat Country result comes down to a few factors that apply to most multi-category Shopify stores.
The most important one: catalog expansion only works if customers can find what you added. Hat Country had grown into a full western lifestyle store, but most customers came for a hat and left without realizing the rest existed. Navigation menus and category pages don't surface products at the moment of purchase intent. Recommendation widgets on product pages and at checkout do. If customers can't discover what you carry, the result is the same as not carrying it.
Free shipping thresholds are also underused as an AOV tool. When checkout recommendations are calibrated to the threshold, a suggested add-on starts feeling like a way to save money rather than spend it. That framing change matters for conversion, but it only works when the recommendation is genuinely relevant to what's already in the cart — otherwise it reads as an obvious upsell and gets ignored.
And on the AI model question: the specificity matters more than most merchants expect. Hat Country's catalog has internal structure, how hat brims relate to crown shapes, how boot styles pair with specific jacket cuts, how belt widths match buckle sizes. Models trained on that specific catalog and order history reflect those relationships. Generic templates applied from other stores don't. A 1,000+ SKU catalog with frequent new arrivals also can't be maintained manually; the catalog changes faster than any curation process can keep up with.
Is your store leaving similar revenue on the table?
Hat Country's result came from a catalog that supports multi-item purchases, a customer base that needed guidance to discover it, and recommendation models built around that specific store's structure. Many Shopify stores with broad catalogs are in the same position: customers who buy single items, and no infrastructure connecting the rest of the catalog to their intent.
In 20 minutes, the PersonalizerAI team will audit your product discovery, show you projected revenue lift based on your actual store data, and walk you through how recommendations would work across your catalog.
Built for scaling Shopify brands with 100+ SKUs and active ad spend.