A value-based lookalike is a Meta lookalike audience built from a customer list that includes a value for each person, such as how much they have spent with you. Meta then looks for new people who resemble your higher-value customers more than your lower-value ones.
How a value-based lookalike works
An ordinary lookalike treats every person in the source equally. A customer who bought once for £15 counts the same as one who has spent £2,000 over three years. A value-based lookalike adds a column of customer value to the source, so Meta weights its search towards the patterns shared by people at the top of your list.
Building one takes three steps:
- Export customers from your CRM or shop platform with identifiers (email, phone, name, postcode) and a value, usually customer lifetime value or total revenue.
- Upload the list in Ads Manager as a customer list custom audience, marking the value column when asked. Meta hashes the identifiers before matching.
- Create a lookalike from that source, choosing the United Kingdom and an audience size, from the closest 1% of the population up to a broader 10%.
Meta needs a minimum number of matched people from one country in the source, and more matched people with a real spread of values gives it a better picture. A list where everyone has the same value adds nothing over a normal lookalike.
Why it matters
Not all customers are worth the same. A subscription box, a B2B supplier or a clinic offering treatment plans may find that a small share of customers brings most of the revenue. Finding more of those people is worth far more than finding more one-off buyers, and a value-based lookalike is one of the few prospecting tools that aims directly at that.
Its role has changed, though. With Advantage+ audience, Meta increasingly treats lookalikes as suggestions rather than hard limits, and broad targeting with good conversion data often matches them. A value-based lookalike is still useful as a signal and as a test, but it is not a guaranteed upgrade.
There is a UK GDPR side. Uploading customer data to Meta is processing personal data for advertising. You need a lawful basis, your privacy notice should say that you share customer details with advertising platforms for matching, and you should respect anyone who has objected to direct marketing. The ICO’s guidance on direct marketing and data sharing is the place to start.
Common mistakes
- Using order value instead of customer value Which rewards one large purchase over loyal repeat buyers.
- Including refunds and staff orders that distort the values.
- A stale list that is never refreshed.
- No comparison. Launching the lookalike without testing it against broad targeting and a normal lookalike.
- Expecting it to fix weak tracking. If purchases are not reported reliably, the campaign cannot learn which of the new people are worth reaching, however good the source list.
How to act on it
Pull a clean customer export, decide on one honest measure of value, and remove refunds, tests and duplicates. Confirm your privacy notice and lawful basis cover the upload. Build a 1% UK value-based lookalike, then run it against your current prospecting set-up with equal budgets, judging on cost per purchase and return on ad spend over several weeks. Refresh the list every few months. I set up and test audiences like this in my Facebook ads management service.