Lookalike from purchase data vs email list

Quick Answer

A lookalike from purchase data almost always outperforms a lookalike from a raw email list because the seed reflects people who actually paid you. An email list lookalike is broader — it includes leads, freebie subscribers, and unconverted trialists. Use purchase data when you want efficiency and a tight CPA. Use email list lookalikes when your purchase volume is too low to build a healthy seed (under 500 conversions), or when you want top-of-funnel reach.

Why purchase lookalikes win on signal quality

Meta builds lookalikes by analysing patterns in your seed audience — demographics, interests, behaviours, device, time-of-day activity, page likes, even purchase history elsewhere on the platform. The richer the signal, the better the model.

A purchase event tells Meta: this person spent money on your category. That is the strongest possible signal. An email subscriber may have downloaded a free PDF three years ago and never opened anything since. Garbage in, garbage out.

Step-by-step setup

1. Audit both seed sources first. In Ads Manager, build a custom audience from your purchase event (last 180 days) and another from your email list. Note the size of each.
2. Clean the email list before uploading. Remove unsubscribes, bounces, and anyone inactive for 12+ months. Aim for at least 1,000 emails post-clean.
3. Hash and upload via Customer File. Audiences > Create > Customer List. Include first name, last name, email, phone, and country for the best match rate.
4. Build separate 1% lookalikes from each. Name them clearly, e.g., LAL_1%_PurchasersL180D and LAL_1%_EmailListClean.
5. Run them in separate ad sets with the same creative, budget, and placement. This isolates the audience variable.
6. Compare CPA after 7 days and 50+ conversions per ad set.

Real examples

DTC coffee brand, £24 AOV. Purchase lookalike CPA: £19. Email list lookalike CPA: £33. Email list contained too many one-time competition entrants. Switched all prospecting budget to purchase lookalike and added the email list as a retargeting layer.

Online course, $497. Purchase volume was only 220 in 90 days — too small to build a stable lookalike. Email list lookalike (12,000 people) actually performed better because the seed was big enough for Meta to find patterns. Once purchase data crossed 500, the purchase lookalike took over.

Pet supplies brand, £55 AOV. Purchase lookalike CPA: £14. Stacked email list lookalike (2% to 5%) for top-of-funnel awareness CPA: £22. Combined approach drove 38% more revenue than purchase lookalike alone.

When to use each

• Purchase data lookalike: You have 500+ purchases in the last 180 days. You want efficiency and a tight CPA. You sell a focused product with a clear ICP.
• Email list lookalike: Purchase volume is below 500. You are running awareness or lead-gen campaigns. You need broader reach. Your email list is well-segmented and recently cleaned.
• Both: Run them in parallel with separate ad sets. Use the email list lookalike as a top-of-funnel layer and the purchase lookalike for warm-down.

Match rate matters

Your email list custom audience is only as useful as the percentage that Meta can match to real accounts. A 60% to 70% match rate is normal. Below 40%, your seed will be too small to build a meaningful lookalike. To improve match rate, include phone numbers, full names, country, and date of birth where you can.

FAQs

Can I combine purchase events and email lists into one custom audience?
Yes. Build a custom audience that includes anyone who purchased OR subscribed. The lookalike will be broader but potentially noisier. Test it as a third option.

Should I use lifetime purchasers or just the last 90 days?
Last 90 to 180 days. Older purchasers may not match your current product or brand direction.

What if my email list is huge but conversions are low?
That is a sign your list quality is poor. Segment to engaged subscribers only (opened in last 60 days) before building the lookalike.

How often should I rebuild these lookalikes?
Meta refreshes lookalikes automatically every 3 to 7 days. You only need to rebuild if you change the seed source or want to test a new time window.

Does GDPR affect what I can upload?
If you are in the EU/UK, you need lawful basis to share email data with Meta. Most brands rely on legitimate interest or consent at signup. Document your basis.

Test creative against the right audience

A lookalike audience is only half the equation — the other half is whether your ad creative resonates with that audience. Pix-Vu lets you test ad creative, hooks, and landing page variants against your target audience before you spend a penny on Meta. Find the winners faster at https://pix-vu.com.