How Klaviyo Product Recommendations Work

You've seen them in emails — the "You might also like" section or "Recommended for you" product grid. When they work, they drive real revenue. When they don't, they show a customer who just bought dog food a recommendation for cat litter.

Klaviyo's product recommendation engine is one of its most powerful features, but most stores either use the default settings (which are mediocre) or don't understand what the engine is actually doing behind the scenes.

Let me pull back the curtain on how Klaviyo's recommendations actually work, and how to configure them so they make you money instead of making you look clueless.

The Recommendation Types

Klaviyo offers several recommendation algorithms. Each one uses different data to decide what products to show.

Best Sellers

This is the simplest. It shows your top-selling products based on order volume over a time period. Everyone sees the same products.

When to use it: in emails to new subscribers who you don't have purchase data for yet. For a brand awareness campaign where you want to showcase your hits. On generic promotional emails.

When NOT to use it: in post-purchase flows (they might have already bought these), in win-back campaigns for repeat customers (they've seen your best sellers — show them something new).

Trending Products

Similar to best sellers but weighted toward recent performance. Products that are selling more than usual right now. This captures seasonal trends and viral moments.

When to use it: newsletters, new arrival campaigns, and seasonal pushes. It's particularly good for fashion and lifestyle brands where trends shift frequently.

Personalized Recommendations

This is where Klaviyo's engine gets interesting. It looks at the individual recipient's browsing history, purchase history, and engagement data, then recommends products that similar customers have purchased.

The algorithm works on collaborative filtering — the same logic Netflix uses for movie recommendations. "People who bought X also bought Y." If a customer bought a moisturizer and a cleanser, and other customers who bought those same products also bought a serum, the algorithm recommends the serum.

When to use it: post-purchase flows, win-back campaigns, browse abandonment, and any flow where you have behavioral data on the recipient.

Limitation: it requires data. For new subscribers with no purchase or browsing history, personalized recommendations fall back to best sellers. The more data Klaviyo has on a customer, the better the recommendations get.

Recently Viewed

Shows products the recipient recently viewed on your website. This requires the Klaviyo tracking pixel on your site (which should already be installed if you're using Klaviyo with Shopify).

When to use it: browse abandonment emails ("Still thinking about these?"), and as a secondary recommendation block in campaigns.

This is one of the highest-converting recommendation types because the customer already expressed interest in these specific products.

Cross-Sell Recommendations

These recommend products that complement the recipient's recent purchase. If they bought a camera, show camera accessories. If they bought a dress, show shoes and bags.

Klaviyo can generate these automatically based on what other customers who bought the same product also purchased. You can also manually configure cross-sell relationships in your product catalog (or through your Shopify data).

When to use it: post-purchase flows (email 2 or 3, after the order has been delivered), and upsell campaigns.

Configuring Recommendations in Email Templates

In Klaviyo's email editor, you add recommendations by inserting a "Product Block" and setting it to "Dynamic" rather than static.

Here's what you can configure:

Number of products. How many products to show. For most emails, 3-4 products is optimal. More than 6 creates decision fatigue and makes the email too long.

Recommendation type. Select from the algorithms above (best sellers, personalized, recently viewed, etc.).

Catalog filter. Restrict recommendations to a specific collection or exclude certain products. For example, you might want to only recommend products from the same category as the customer's last purchase. Or exclude products that are out of stock. Or exclude the product they just bought.

Fallback behavior. What happens when Klaviyo can't generate personalized recommendations (e.g., new subscriber with no data)? Set a fallback to best sellers or a curated collection.

Sort order. By relevance (default), price (low to high or high to low), or newest.

The Product Feed Connection

Recommendations are only as good as your product data. Klaviyo pulls product information from your Shopify catalog (or whatever platform you use) through a catalog feed.

Make sure your catalog feed includes:

• Product images (high quality — these show in emails)
• Accurate pricing (including sale prices)
• Inventory status (to avoid recommending out-of-stock products)
• Product categories/collections (for filtering)
• Product descriptions (used by the algorithm for content-based recommendations)

In Klaviyo, go to Content, then Products to verify your catalog is syncing correctly. If products are missing or have outdated prices, fix the sync before relying on recommendations.

Common issues:

• Draft products showing in recommendations (set your feed to only include published/active products)
• Products with no images appearing as blank cards
• Out-of-stock products being recommended (filter these out in your recommendation block settings)

Advanced Configuration

Exclude Recently Purchased Products

By default, Klaviyo might recommend a product the customer just bought. This is especially awkward in post-purchase flows. "You just bought this moisturizer. Want to buy it again?" Three days after purchase? No.

In the recommendation block, add a filter: "Exclude products the recipient has purchased in the last X days." For non-consumable products, set this to 90-180 days. For consumable products, set it based on the typical repurchase cycle (30-45 days for monthly supplements, for example).

Price Range Filtering

If a customer typically buys products in the $30-50 range, showing them a $200 product in recommendations is unlikely to convert. You can filter recommendations by price range relative to the customer's average order value.

This isn't a built-in filter in Klaviyo's recommendation block, but you can achieve it by creating collections based on price tiers and filtering recommendations to the appropriate collection.

Category Affinity

Some customers only buy from one category. A skincare-only customer probably doesn't want to see supplement recommendations. Use Klaviyo's conditional content to show category-specific recommendation blocks:

• If customer has purchased from skincare collection, show skincare recommendations
• If customer has purchased from supplements collection, show supplement recommendations
• Default: show best sellers across all categories

New Products Priority

If you launch new products frequently, you want recommendations to include new arrivals. Add a secondary recommendation block below personalized recommendations that shows "New This Week" or "Just Launched" products. This ensures new products get visibility even when the algorithm hasn't built enough purchase data to recommend them yet.

Flow-Specific Recommendation Strategy

Welcome Flow

Email 1: No recommendations (focus on the welcome message and incentive). Email 2: Best sellers (they don't have purchase data yet, so personalized won't work). Email 3: Trending products or curated "staff picks."

Post-Purchase Flow

Email 1 (order confirmation): No recommendations — keep the focus on the order. Email 2 (3-5 days after delivery): Cross-sell recommendations related to their purchase. Email 3 (14 days after delivery): Personalized recommendations based on purchase + browsing history.

Browse Abandonment Flow

Email 1 (1-2 hours after browsing): Recently viewed products. Email 2 (24 hours later): Personalized recommendations that include similar products to what they viewed.

Win-Back Flow

Email 1: Best sellers or trending (remind them why your brand is worth coming back to). Email 2: Personalized recommendations based on their historical purchases (with recently purchased items excluded).

Measuring Recommendation Performance

Klaviyo shows revenue attribution for recommendation blocks. Track:

• Click rate on recommendation blocks. How often do people click the recommended products? Benchmark: 2-5%.
• Revenue per recommendation click. When someone clicks a recommendation, what's the average order value?
• Overall flow/campaign revenue. Compare revenue from emails with recommendations vs. without. The lift should be measurable.
• Recommendation accuracy. Qualitatively review the recommendations for a sample of profiles. Do they make sense? Is the algorithm recommending relevant products?

If your recommendation blocks have a click rate below 1%, either the products aren't relevant or the placement in the email isn't prominent enough.

The Bottom Line

Product recommendations in Klaviyo are not a "set it and forget it" feature. The default settings work, but they don't optimize. Configure your catalog feed properly, choose the right algorithm for each flow and campaign, filter out irrelevant products, and measure performance.

Done right, recommendation blocks add 10-25% incremental revenue to your email program. Done wrong, they make your brand look like it doesn't know its own customers.

If you want help configuring Klaviyo's recommendation engine to actually move the needle, book a call with us. We'll audit your current setup and show you what's leaving money on the table.