Scaling with lookalikes: 1% → 10% progression

Quick Answer

Scaling lookalikes from 1% to 10% used to be the standard play. In 2026, with broad targeting now matching or beating high-percentage LALs in most accounts, the question is whether to bother with the progression at all. The honest answer: 1-3% LALs still work for narrower verticals; above 5% you may as well be running broad. Most accounts at scale should run 1% LAL alongside broad and skip the middle tiers entirely.

The Framework

1. Build LALs from your highest-value seed

Don't build LALs from 'all purchases' — build them from your top 25% of customers by AOV or LTV. Smaller seed, but higher quality, and the resulting LAL will perform better.

2. Start with 1% in your strongest geo

1% LAL of top customers in your home market is the best LAL audience you can build. Use it as your first test. If 1% can't beat your other audiences, larger percentages won't either.

3. Test 3% as the next step, not 5%

3% gives you roughly 3x the audience size and is still concentrated enough to be meaningfully 'similar' to your seed. 5% starts to dilute. 10% is essentially a broad audience with extra steps.

4. Run LAL and broad in parallel and let Meta arbitrate

Don't treat LAL and broad as alternatives. Run 1% LAL in one ad set and broad in another. Meta will allocate spend based on performance. Often broad wins on volume and 1% wins on ROAS.

5. Watch for LAL-broad overlap

1% LAL audiences usually have 60-80% overlap with broad targeting in modern Meta. If broad is converting well, LAL is mostly hitting the same people. Use the auction overlap report to verify.

6. Refresh seed audiences quarterly

If you're using LAL, refresh the seed audience every 3 months. Stale seeds produce stale LALs. New buyers should keep the model current.

Real Numbers from the Field

An ecommerce home goods brand had been running a 5-tier LAL stack (1%, 2%, 3%, 5%, 10%) plus broad for 18 months. We tested collapsing to just 1% LAL + broad. Total spend held at $9k/day, but blended ROAS improved 12% because the middle-tier LALs had been cannibalising broad without adding incremental conversions. The simpler structure was easier to manage and performed better.

Frequently Asked Questions

Are lookalikes still useful in 2026?

Yes, in narrower scenarios than before. 1% LALs of high-value customers still consistently beat broad in some verticals. But the era of 'stack 5 LAL tiers and scale' is over.

Can I use LAL with Advantage+ Shopping?

ASC handles audience selection itself, but you can include LAL audiences as 'audience suggestions' which ASC will consider. Performance gain is marginal.

Should I build LALs from different events (purchase vs add-to-cart)?

Purchase-based LALs almost always outperform earlier-funnel LALs. Stick with purchase or top customers — don't waste time on LALs of pageviewers.

How big should my seed audience be for a good LAL?

Minimum 1,000, ideally 5,000+. Below 1,000, the LAL becomes statistically noisy and underperforms broad.

What about lookalikes from email lists?

Customer list LALs (especially top customers by LTV) are often the strongest LALs you can build. Upload lists quarterly and refresh.

Stop Guessing What Your Ads Look Like at Scale

When you're spending £5k+ a day, you can't manually screenshot every ad on every device, in every placement, in every locale. Things break silently. Creators ship the wrong aspect ratio. Localised text overflows. Safe zones get clipped by Reels UI.

Pix-Vu lets you preview every ad creative across Facebook, Instagram and Reels placements before launch — desktop, mobile, all aspect ratios, all crop variants. Spot the broken creative before you spend £20k discovering it the hard way.

Try it free at https://pix-vu.com — built for media buyers who can't afford a single wasted impression.