Meal delivery is a convenience-driven vertical where customers expect variety and personalization. Ordering patterns are habitual (weekly favorites), goal-oriented (healthy eating or a specific diet), or occasion-based (busy weekdays or family dinners).
This article shows how to use Smart Recommender in Insider One to inspire meal discovery, simplify choices, and grow order value. It includes an A/B test walkthrough for meal add-ons and strategies for each page type.
Before you start: Make sure your product catalog includes the attributes this article filters on, such as Category, Cuisine Type, Season, and Bundle ID. Create missing custom attributes on the Product Attributes page and mark them as filterable.
What you can achieve with Smart Recommender
Meal discovery on the homepage
Highlight trending dishes, seasonal picks, or new arrivals such as "Protein-Packed Meals."
Easy reorders
Show regulars the meals they ordered before so they can reorder quickly.
Sides, drinks, and desserts
Recommend add-ons on product or cart pages to raise average order value (AOV).
Re-engagement
Show inactive customers past favorites alongside new options.
Checkout extras
Suggest sauces, protein boosts, or desserts before checkout.
Taste-based personalization
Use Attribute Affinity with cuisine type, diet, or preparation time.
A/B testing
Test strategies, widget placement, and visual design.
Walkthrough: Boost AOV with meal add-ons
Many customers order with a clear goal, such as a quick lunch. A well-timed side, drink, or dessert raises order value and satisfaction.
This walkthrough tests two cross-sell strategies on meal product detail pages for customers with a low AOV:
Variant A: Recommend beverages.
Variant B: Recommend desserts.
Step 1: Create the recommendation strategies
Navigate to Recommendation Strategies and click Create. Configure the first strategy as follows:
Select Product Detail Page as the page type.
Select Purchased Together as the algorithm.
Enter the number of products to recommend.
Exclude products purchased in the last four weeks.
Add a filter for the category attribute: Category + is + Beverages.
Create the second strategy with the same settings and change only the category filter:
Category + is + Desserts.
Note: Keep every other setting identical between the two strategies. If more than one setting changes, you cannot tell which change caused the difference in results.
Step 2: Launch the campaign
Navigate to Web Smart Recommender and click Create.
Select the integration method for the widget.
On the Segments step, select customers with an AOV below your threshold (for example, 20 USD) under Purchasing Behavior.
On the Rules step, use Page Rules to target meal product detail pages only.
On the Design step, assign one strategy to each variant and set the traffic allocation for each variant.
Click Edit Design to open the Advanced Product Card Designer and customize your product cards.
Select the locales and stores where the campaign appears.
Review the campaign details, confirm that every setting matches your goal, and click Launch.
Note: Every attribute you show on a product card must exist in your product catalog. To show more information, create custom attributes on the Product Attributes page.
Step 3: Track campaign metrics
Navigate to Smart Recommender Analytics.
Find your campaign in the Campaign and Variant Metrics table. When the test period ends, click the campaign name.
Compare Direct Revenue, Average Order Value (AOV), and Conversion Rate for the two variants. This shows whether beverages or desserts raise AOV more for low-spending customers.
If one variant clearly wins, set its traffic allocation to 100% so every visitor sees the better strategy.
After you find a winner, test other add-ons, such as soups, sauces, or protein boosts.
Step 4: Optimize your campaigns
After the campaign has run long enough to collect data, use Smart Recommender Analytics to improve it:
Engagement funnel: Review product impressions, click-through rate, add-to-cart rate, and conversion rate to see where shoppers drop off.
Campaign and variant performance: Compare AOV, Conversion Rate, and Direct Revenue in the Campaign and Variant Metrics table, then refine or scale the winning variant.
Product-level impact: Open Top 100 Product Analytics to see which recommended products perform best. Give extra visibility to strategic products that perform below expectations.
Category trends: Use Category Analytics to find the categories that drive the most conversions, and decide whether to prioritize them or promote weaker ones.
Use cases by page type
Home page
Seasonal menus
Show the most viewed seasonal dishes, such as salads and smoothies in summer.
Create a strategy with the Most Popular algorithm.
Turn on the Enhance recommendations based on Attribute Affinity toggle. This personalizes by diet and cuisine.
Add the filter Season + is + Summer.
If you do not have a Season attribute, create it on the Product Attributes page.
Tip: A/B test seasonal imagery or the widget position.
Meal sets for first-time customers
Help new customers start with sets such as Intro to Keto or Family Dinner Essentials.
Create a strategy with the Top Sellers algorithm.
Add the filter Category + is + Meal Sets.
Target new users on the Segments step.
Note: To recommend the rest of a bundle, create a custom attribute called Bundle ID on the Product Attributes page and give every item in a bundle the same value (for example, Keto Bundle 02).
Tip: A/B test headlines such as "Top Selling Menus" and "Your First Order, Sorted."
Recently viewed meals
Remind returning customers of dishes they explored before.
Create a strategy with the Recently Viewed algorithm.
Add the filter Category + is not one of + Beverages, Sauces, Desserts to keep add-ons off the homepage.
Target returning users on the Segments step.
Note: Recently Viewed and User Based rely on each user's past behavior. Target returning users so these widgets always have data to show.
Tip: A/B test Recently Viewed against User Based.
Product detail page
Picks based on recent browsing
Show each customer meals that match their most recent interactions.
Create a strategy with the User Engagement algorithm.
Add the filter Category + matches the item they're currently viewing.
Tip: Try headlines such as "Not This One? Try These."
More from this cuisine
Recommend more dishes from the same cuisine, such as Korean or Mediterranean.
Create a strategy with the Trending Products algorithm.
Add two filters: Cuisine Type + matches the item they're currently viewing, and Category + matches the item they're currently viewing. Connect them with an OR connector.
If you do not have a Cuisine Type attribute, create it on the Product Attributes page.
Tip: Use a dynamic title such as "More Korean Dishes."
Category page
Trending in this diet type
Show dishes gaining popularity this week in the diet category the customer is browsing.
Create a strategy with the Trending Products algorithm.
Add the filter Category + contains + the category they're currently viewing.
If your categories are hierarchical, select the Expand the category filter to include recommendations from the next category if there aren't enough products to display checkbox, so the widget still fills when the current category has few products.
Turn on the Enhance recommendations based on Attribute Affinity toggle.
Discounted meals in this category
Show the best deals in categories such as Keto or High-Protein.
Create a strategy with the Highest Discounted algorithm.
Add the filter Category + matches + the category they're currently viewing.
Leave Attribute Affinity off.
Tip: Add labels such as "20% Off" and A/B test the widget at the top versus the bottom of the page.
Cart page
Complete your meal
Suggest sides, soups, snacks, or drinks to round out the order.
Create a strategy with the Most Popular algorithm.
Add the filter Category + is one of + Side Dishes, Drinks, Snacks, Desserts.
Leave Attribute Affinity off.
Tip: Try headlines such as "Round Out Your Meal."
Reach free delivery
Show items that help customers reach your free delivery threshold.
Create a strategy with the Checkout Recommendation algorithm and set the spend threshold (for example, 40 USD).
No filters are required. Attribute Affinity is optional.
Tip: Use copy such as "Add 5 USD more for free delivery." If you do not offer free delivery, use Purchased Together instead.