How to Set Up Value-Based Lookalike Audiences on Meta

Value-based Lookalike Audiences on Meta use the lifetime value (LTV) of your existing customers as a signal to help Meta find new users who resemble your highest-spending buyers rather than just any buyer, typically delivering higher average order values and better LTV on acquired customers. Last updated: February 2026

Table of Contents

• What Are Value-Based Lookalikes?
• Standard Lookalike vs Value-Based Lookalike
• Requirements for Value-Based Lookalikes
• How to Create a Value-Based Lookalike
• Preparing Your Customer Value Data
• Testing and Scaling Value-Based Lookalikes
• Common Mistakes and How to Avoid Them
• FAQ

What Are Value-Based Lookalikes?

A standard Lookalike Audience treats all customers equally. Whether someone spent $30 or $3,000 with you, they count the same in the source audience. The algorithm finds people who resemble all of your buyers without any weighting for how good those buyers are.

A value-based Lookalike, also called a "value-optimized Lookalike," tells Meta which customers in your source audience are most valuable by including a monetary value alongside each record. Meta then builds the lookalike by weighting the model toward users who resemble your highest-value customers.

The result: Meta finds new users who are more likely to become high-LTV customers, not just any customer. For DTC brands focused on long-term profitability rather than just acquisition volume, this is a meaningful distinction.

According to internal MHI Media data, brands using value-based Lookalikes see 20-30% higher average order values on first purchase compared to standard Lookalike campaigns, though the audience size may be slightly smaller at equivalent similarity percentages.

Standard Lookalike vs Value-Based Lookalike

Standard Lookalike:

• Source: Customer list or purchase event (all customers weighted equally)
• Meta's model: Find people who look like any of your buyers
• Best for: Maximum reach, early-stage brands, lower-priced products

Value-Based Lookalike:

• Source: Customer list with LTV column
• Meta's model: Find people who look like your high-value buyers
• Best for: Brands with clear LTV variation, subscription or repeat purchase models, higher-priced products

The right choice depends on your business model. If 80% of your LTV comes from 20% of your customers, value-based Lookalikes help Meta find more customers like that 20%. If your customers are relatively homogeneous in value (most buy once at similar prices), the benefit is smaller.

Requirements for Value-Based Lookalikes

To create a value-based Lookalike, you need:

1. Customer list with value data: A CSV file with at minimum: email address and a customer_value column. Meta uses the value column to weight the source audience. 2. Minimum 100 matched customers: Meta requires at least 100 people in the matched portion of your list (after hashing and matching to Meta accounts). For meaningful performance, aim for 1,000+ matched records. 3. Business Manager with a connected ad account: Value-based Lookalikes are created in Audiences within Business Manager. Ensure your Business Manager is properly configured. 4. Value data that reflects actual customer worth: Do not use purchase count or first order value if LTV is more meaningful for your business. The value you pass should reflect what each customer is actually worth to your business over time.

How to Create a Value-Based Lookalike

Step 1: Prepare Your Customer File

Create a CSV with these columns:

• `email` (required for matching)
• `customer_value` (the LTV or value metric you want Meta to use)
• Optional additional columns: `phone`, `fn` (first name), `ln` (last name) for improved match rates

Example file structure:

email,customer_value,fn,ln
john@example.com,450.00,John,Smith
jane@example.com,125.00,Jane,Doe

Values should be positive numbers in your local currency. Meta normalizes the values internally; you do not need to standardize them to a specific scale.

Step 2: Create a Customer List Custom Audience

• In Meta Ads Manager, go to Audiences
• Click "Create Audience" then "Custom Audience"
• Select "Customer list"
• Choose "Yes, include customer value in my list"
• Upload your prepared CSV
• Map the columns: identify which column is email, which is customer value
• Accept Meta's data terms
• Name the audience clearly: "Customer List - LTV Weighted - [Month/Year]"
• Click "Create"

Allow 24-48 hours for the audience to populate and match.

Step 3: Create the Value-Based Lookalike

• In Audiences, find your newly created customer list audience
• Click the three-dot menu and select "Create Lookalike"
• Select your target location (country)
• Select audience size (1% recommended for initial testing)
• You will see the option to create a "Value-based lookalike" if your source audience has value data
• Select this option
• Click "Create Audience"

If you do not see the value-based option, your source audience may not have sufficient matched users or the value column may not have been properly mapped during upload.

Preparing Your Customer Value Data

The quality of your value data determines the quality of your value-based Lookalike. Here are the main approaches:

Total lifetime revenue: Sum of all orders per customer. Straightforward and reflects actual business value. Best for most DTC brands. `customer_value = sum of all historical orders for this email` Predicted LTV: If you have LTV modeling (some Shopify apps provide this), use predicted 12-month or lifetime value instead of historical revenue. This is more forward-looking but requires modeling infrastructure. AOV-weighted approach: If you do not have long purchase histories, use average order value. Not as powerful as true LTV but better than treating all customers equally. Segmented value approach: Instead of a continuous value field, use tier scores. High-LTV customers get a value of 100, mid-LTV get 50, low-LTV get 10. This coarser approach still provides directional weighting.

Export customer value data from Shopify via the customer export feature, or from your CRM/email platform. Append LTV data from your analytics system or order history.

Testing and Scaling Value-Based Lookalikes

Initial Test Setup

Run your value-based Lookalike against your standard purchase Lookalike in a controlled comparison:

Ad Set A: 1% Lookalike from standard purchase Custom Audience (180 days) Ad Set B: 1% Value-Based Lookalike from LTV-weighted customer list

Same creative, same budget, same campaign, same time period. Compare over 14-21 days.

Key metrics to compare:

• CPA: Which acquires customers cheaper?
• Average Order Value (AOV) of acquired customers: Does value-based find higher-AOV buyers?
• 30-day repurchase rate (if trackable): Do value-based acquired customers return faster?

Scaling Winners

If value-based Lookalike wins on CPA or delivers meaningfully higher AOV:

• Expand to 2% and 3% sizes
• Refresh the source list monthly (re-upload current customer LTV data)
• Consider adding to your primary prospecting rotation

Common Mistakes and How to Avoid Them

Uploading outdated customer data: LTV changes as customers make more purchases. Re-upload your customer list monthly to keep the Lookalike model current. Using only new customer data: Include all-time customers for maximum list size and LTV variance. New customers have less LTV data. Confusing order count with value: A customer who made 5 purchases of $20 each ($100 total) is different from one who made 2 purchases of $200 ($400 total). Ensure your value column reflects total revenue, not order count. Not testing value-based against standard: Assume nothing. Run the test. In some categories, standard Lookalikes perform equally well. In others, value-based delivers significant improvement. Data beats assumptions. Setting customer_value to 0 for any customers: Zero or negative values confuse the algorithm. Exclude customers with zero LTV from the upload or assign a minimum value of 1.

---

FAQ

Can I use pixel purchase events for value-based Lookalikes? Not directly. Value-based Lookalikes require a customer list with value data uploaded as a Custom Audience. Pixel-based purchase events track occurrence but do not pass individual customer LTV to Meta in a way that supports value weighting. What is the minimum number of customers needed? Meta requires at least 100 matched customers, but meaningful performance requires 1,000+ matched users. Below 1,000, the model has limited data to work with. How does Meta use the value data? Meta normalizes the values and uses them to weight the model toward characteristics shared by higher-value customers. The algorithm does not see or store the raw values; it uses them only to calibrate the lookalike model. Can I create value-based Lookalikes for different countries? Yes. When creating the Lookalike, select your target country. If you want value-based Lookalikes in multiple countries, create separate audiences selecting each country. The model is informed by your customer list regardless of the target country. How often should I update my customer list? Monthly is ideal. LTV data changes as customers make repeat purchases. Stale lists mean your Lookalike model is optimizing for an outdated picture of who your best customers are. Does Meta share what characteristics it finds in my high-value customers? No. Meta does not disclose the specific signals it uses to build Looakalike models. Audience insights tools can reveal some demographic and interest patterns in audiences, but Meta does not break down the Lookalike modeling logic. Is value-based Lookalike available for all advertisers? Value-based Lookalikes are available to accounts that meet Meta's minimum requirements (sufficient matched users with valid value data). They are not available in all regions or for all ad account configurations. If you do not see the option during audience creation, ensure your customer list file has a properly mapped value column.