1% vs 2% vs 5% lookalikes — when to use each

Quick Answer

Use a 1% lookalike when your seed list is small, well-defined, and high quality (think paying customers or top 10% LTV). Use a 2% to 3% lookalike when you have already saturated your 1% and need to scale without losing too much signal. Push to 5% (or higher) only once your CPA on the smaller segments has crept above your target and you need fresh inventory. The right size is the one that keeps your CPA inside acceptable range while delivering enough volume to spend your daily budget.

What the percentage actually means

Meta builds lookalikes by ranking every adult in your chosen country by similarity to your seed audience. A 1% lookalike is the top 1% most similar people. A 5% lookalike includes everyone in the top 5%. The 5% audience contains the 1%, the 2%, the 3%, and the 4% — so if you run them together you will overlap heavily.

Think of it as concentric rings. The smallest ring is the most similar to your customers, and accuracy drops as the ring widens.

Step-by-step setup

1. Open Audiences in Meta Ads Manager. Click Create Audience, then Lookalike Audience.
2. Pick your source. Use a custom audience built from your highest-quality data — paying customers, repeat buyers, or value-based events.
3. Choose the location. Lookalikes are country-specific. If you sell in the UK, US, and Australia you will need three separate audiences.
4. Set the size. You can pick a single percentage (e.g., 1%) or a range (e.g., 1% to 5% as one combined audience). For testing, build them as separate slices: 0% to 1%, 1% to 2%, 2% to 5%.
5. Name them clearly. Use a format like LAL_1%_UK_PurchasersL90D_v1 so you can track which is performing.
6. Wait for population. Audiences take 6 to 24 hours to populate. Do not launch ad sets until they show ready.

Real examples

Ecom skincare brand, £40 AOV. Started on a 1% lookalike of last 90 days purchasers. CPA hit £18 with 800 daily spend. Tried scaling to £1,500/day and CPA jumped to £31 because of audience saturation. Switched to a stacked 2% to 3% lookalike layered with the 1% as exclusion — CPA settled at £21 and they could spend £2,200/day cleanly.

B2B SaaS, $89/month. 1% lookalike of paid signups was tiny (the seed was only 1,400 customers in the US). Audience size came back at 2.1M. CPA was great but daily spend ceiling was £180. Moved to a 3% lookalike for prospecting and used the 1% as a high-intent retargeting layer.

Local fitness studio, £60 trial. 1% lookalike of trial-to-member converters was too narrow geographically. The studio used a 5% lookalike of all leads inside a 10-mile radius and let Meta narrow with location targeting. CPL dropped from £14 to £8.

When each size wins

• 1% wins when your seed is rich (high-LTV customers, value-based events, repeat purchasers) and you need precision over volume. Best for niche products and high-consideration purchases.
• 2% to 3% wins when you are spending more than your 1% can absorb without rising costs, or when your seed audience is small (under 1,000 people).
• 5% wins when you are running large budgets, have already extracted value from smaller percentages, or you sell a broad consumer product where precision matters less.

FAQs

Does the 5% lookalike contain the 1%?
Yes. Lookalikes are cumulative. To avoid overlap, exclude the smaller percentages when running larger ones in the same campaign.

How many people are in a 1% lookalike?
It depends on the country. In the US a 1% lookalike is roughly 2.1 million people. In the UK it is about 530,000. In smaller markets like Ireland or New Zealand it can be under 50,000.

Should I test all three percentages at the same time?
Only if your daily budget is high enough to give each ad set a fair chance (around 50 conversions per week). Otherwise, start with 1% and scale outward when CPA hits your ceiling.

What if my 1% lookalike has fewer than 100,000 people?
That is too small for reliable delivery. Either widen to 2% or 3%, or use a broader seed audience.

Does Advantage+ make these percentages obsolete?
Not completely. Advantage+ Audience uses lookalikes as a signal rather than a hard constraint. You still build them the same way, but Meta chooses how strictly to honour them.

Stop guessing what your customers actually look at

Lookalike audiences are only as good as the visuals you show them. If your creative does not match the moment, even a perfect 1% lookalike will burn budget. Pix-Vu lets you test ad creatives, landing pages, and product images against real audience reactions so you can ship the variants Meta will reward. Start free at https://pix-vu.com.