A value-based lookalike audience tells Meta to find people who resemble your best customers, not just any customer. You give Meta a source with a value for each person, usually lifetime spend, and Meta weights its search toward people like the high-value end of the list. A standard lookalike audience treats every source person as equal.
How it works
- Build a value-based source. The most common is a customer list uploaded with a value column, such as each customer's total spend over 12 months or customer lifetime value. Meta can also build value-based sources from pixel or app events that carry purchase values, where offered.
- Create a lookalike from that source and choose the country and size, from 1% (closest match) to 10% (broadest).
- Use it as an audience in original targeting, or as a suggestion in Advantage+ audience.
Meta requires a minimum source size for lookalikes, at least 100 people from one country, and recommends much larger sources for better results.
What goes in the value column
| Option | Pros | Cons |
|---|---|---|
| Total revenue per customer | Easy to export | Ignores margin differences |
| 12-month spend | Recent and comparable | Short history for new brands |
| Contribution margin per customer | Closest to real profit | Needs cost data per order |
| Predicted LTV from your tools | Forward-looking | Only as good as the model |
Use one consistent definition and currency. Exclude test orders, staff orders and wholesale buyers, which can distort the weighting.
Worked example
Example: a coffee subscription brand has 4,000 customers.
| Group | Customers | 12-month spend each |
|---|---|---|
| Top 20% | 800 | About ₹12,000 |
| Middle 50% | 2,000 | About ₹3,000 |
| Bottom 30% | 1,200 | About ₹600 |
A standard lookalike gives every customer the same weight, so one-off buyers shape it as much as loyal subscribers. A value-based lookalike leans toward people like the top 800. If those people are hard to find, costs per first order may be higher, but the customers may be worth more. Judge it on LTV:CAC, not only on first-order CPA.
When to use value-based lookalikes
- Customer value varies a lot, for example subscriptions, consumables or a catalogue with very different price points.
- You have enough customers with history for the value column to mean something.
- You optimise for value with a ROAS goal or highest value bidding, so the audience and the bid strategy point the same way.
Limits to keep in mind
- Lookalikes have become less decisive as Meta's automated targeting has improved, and in Advantage+ audience they are only a suggestion.
- A tiny value-based source can produce a lookalike that is no better than a standard one.
- Refresh the source regularly. A list from two years ago describes past customers.
Our lookalike audiences guide covers source building and testing in more depth.
Frequently asked questions
What is a value-based lookalike audience?
It is a Meta lookalike built from a source that includes a value per customer, such as lifetime spend. Meta weights its search toward people similar to your highest-value customers instead of treating all customers equally.
How do I create a value-based lookalike on Facebook?
Upload a customer list with a value column, such as 12-month spend, as a custom audience and map that column to customer value. Then create a lookalike from it, choosing the country and size. Some pixel or app sources with purchase values can also be used.
How many customers do I need for a value-based lookalike?
Meta's minimum for any lookalike source is 100 people from one country, but a source that small rarely helps. Use as many customers with real purchase history as you have, ideally refreshed regularly.
Are value-based lookalikes better than standard lookalikes?
They can be when customer value varies a lot, because Meta looks for people like your best buyers. When most customers spend similar amounts, the difference is small. Judge them on customer value and LTV:CAC, not only on first-order cost.
