Meta Ads offers two core scaling strategies: lookalike audiences (LLA) and broad targeting. Both expand reach beyond your existing customer list, but they work in opposite directions. Choosing the wrong one wastes budget; choosing the right one accelerates growth. This post clarifies when each makes sense and what to measure.
The Core Difference
A lookalike audience mirrors your best customers. Meta finds people who share traits (location, behavior, interests) with users in your seed audience. You start with a known group (customers, high-value visitors, leads) and expand outward.
Broad targeting removes targeting constraints entirely. You set a location and age range, then Meta shows your ad to anyone in that demographic. The algorithm learns who to prioritize based on your conversion event (purchase, signup, add-to-cart). No seed audience required.
The tradeoff: Lookalikes are more precise upfront but limited by seed quality. Broad targeting is less predictable initially but scales further when conversion data accumulates.
When to Use Lookalike Audiences
Lookalikes work best when you have a strong seed audience and want predictable performance early.
You have 500+ high-intent users to seed from
A lookalike needs critical mass to work. Meta recommends 500–1,000 users minimum; below that, the similarity signal weakens. Your seed should be people who actually converted (customers, not just clickers). If you only have 200 customers, a lookalike based on 200 will underperform compared to broad targeting because the pattern is too thin.
You're scaling a proven offer
Lookalikes shine when you've already validated product-market fit in one segment. You've sold to dentists in the US. Now you want to reach dentists in Canada or expand to other healthcare professionals who look like your existing customers. The seed audience proves the offer works; the lookalike extends it to similar people.
Your conversion window is short
Broad targeting needs time to learn. The Meta algorithm requires 50+ conversions per week to optimize effectively. If your sales cycle is 30 days or longer, broad targeting will spend budget inefficiently for the first 2–3 weeks while it's still collecting data. Lookalikes start with intelligence already built in, so they perform faster when you're in a hurry.
Budget is tight and you can't afford learning phase waste
Lookalikes cost more per impression but convert at higher rates early. Broad targeting is cheaper per impression but has higher waste during the learning phase. If you have $2,000/month, a lookalike campaign might spend that $2,000 on 500 conversions at 4% conversion rate. A broad campaign might spend the same $2,000 on 300 conversions at 2% while learning. The math favors lookalikes when budget is limited.
When to Use Broad Targeting
Broad targeting excels when you want to discover new customer profiles and have budget for the learning phase.
Your seed audience is small or unproven
If you have fewer than 500 conversions or your conversion data is noisy (lots of accidental clicks, low-intent traffic), a lookalike will inherit that noise. Broad targeting ignores the seed and learns directly from your conversion event. A new SaaS product with 50 trial signups should use broad targeting, not a lookalike built from 50 signups.
You want to discover new customer segments
Lookalikes replicate what you already know. Broad targeting explores. If your research shows dentists buy your software, a lookalike reaches similar dentists. But broad targeting might find that veterinarians, optometrists, and chiropractors also convert well. You'll never discover those segments with a lookalike.
You have budget for a learning phase (2–4 weeks)
Meta's algorithm needs 50+ conversions per week to stabilize. That means spending enough budget to hit that threshold. A brand new campaign on broad targeting will waste 20–30% of spend in week one while the algorithm is still training. If your monthly budget is $10,000 and you can absorb $2,000–$3,000 in learning waste, broad targeting is viable. If $2,000 is your entire monthly budget, the waste is too high.
Your conversion event is reliable and high-volume
Broad targeting works when the Meta algorithm has clear, consistent signals to learn from. A purchase event is stronger than a "view product" event. A form submission from a qualified lead is stronger than a page view. If your conversion event fires 100+ times per week with minimal noise, broad targeting will learn fast and scale efficiently.
The Audience Maturity Framework
Think of your business as moving through stages. Each stage favors a different strategy.
Stage 1: Pre-product-market fit (0–100 conversions)
Use broad targeting. You don't have enough data to build a reliable lookalike. Spend is low; the learning waste doesn't matter. Your goal is to find any customer, not optimize for a specific type. Lookalikes here are noise.
Stage 2: Validated product (100–500 conversions)
Broad targeting still works, but lookalikes become viable if your conversion event is clean. If 80% of your conversions are real customers (not accidental clicks), build a lookalike from your best 50–100 customers and run it alongside broad. Test both; one will outperform.
Stage 3: Scaling phase (500–2,000 conversions)
Lookalikes now outperform broad targeting for most businesses. You have enough seed data, proven product-market fit, and can afford higher cost-per-result. Broad targeting still works for discovering new segments, but it's secondary. Run a 70/30 split: 70% lookalike, 30% broad for exploration.
Stage 4: Mature business (2,000+ conversions)
Use multiple lookalikes. Segment your customer base by value (high-LTV vs. low-LTV), geography, or product. Create a lookalike for each segment. Broad targeting becomes a small experimental budget (5–10%) for testing new offers or geographies.
Common Mistakes
Building a lookalike from the wrong seed
A lookalike built from all website visitors (not just customers) will find people who look like clickers, not buyers. A lookalike built from trial signups (not paying customers) will find trial-prone users who never convert. Seed from your highest-intent group: paying customers, high-LTV accounts, or qualified leads that closed. If you don't have 500 of those, use broad targeting instead.
Assuming lookalikes are always better
Lookalikes are not a silver bullet. They perform best when your seed is proven and your business is scaling. Early-stage products, new offers, and new geographies need broad targeting. Forcing a lookalike into an unproven situation wastes budget on a narrow, potentially wrong assumption.
Not refreshing your lookalike seed
A lookalike built from customers acquired 6 months ago will decay. Your product changes, your market shifts, your best customers evolve. Refresh your lookalike seed every 30–60 days with your most recent high-value conversions. This keeps the audience aligned with your current business.
Ignoring conversion data quality
Both strategies depend on clean conversion data. If your pixel fires on every page view, every scroll, or every accidental click, neither lookalikes nor broad targeting will work well. Audit your conversion event first. Make sure it fires only for real intent (purchase, qualified form, high-value action). If your data is noisy, no targeting strategy fixes it.
Quick Decision Tree
Do you have 500+ high-quality conversions (customers, not clickers)? Yes: Try lookalikes. No: Use broad targeting.
Is your product proven in one market and you're expanding to a similar one? Yes: Lookalikes. No: Broad targeting.
Can you afford 2–4 weeks of learning-phase waste? Yes: Broad targeting. No: Lookalikes.
Do you want to discover new customer types or validate your current targeting? Discover new types: Broad targeting. Validate current: Lookalikes.
What to Measure
Lookalikes and broad campaigns need different metrics. Track cost-per-result (CPR), conversion rate, and return on ad spend (ROAS) in the first week and first month. Lookalikes should outperform in week one; broad targeting catches up by week 3–4 as it learns. If a broad campaign hasn't improved by week 4, it's not going to; pause it and try a different angle.
Don't judge a lookalike campaign on impression volume. Lookalikes intentionally show to fewer people at higher conversion rates. Broad campaigns show to more people at lower conversion rates. The winner is the one with better ROAS or lower CPR, not the one with more impressions.
Next Steps
Audit your current conversion data. Count how many high-quality conversions you have in the past 30 days. If it's 500+, test a lookalike alongside your broad campaigns. If it's below 500, focus on broad targeting and prioritize improving your conversion event quality. Once you hit 500, revisit this decision and shift to lookalikes. For help structuring your Meta Ads strategy, consider a Meta Ads audit to identify which approach fits your current stage.
FAQs
Can I use both lookalikes and broad targeting at the same time?
Yes. Run them as separate ad sets in the same campaign or different campaigns. Track performance separately so you know which is delivering better results. Most mature businesses use both, with lookalikes as the primary budget and broad as a secondary exploration channel.
How big should my lookalike audience be?
Meta will show you the estimated audience size when you create the lookalike. For most businesses, 500K–2M people is effective. If Meta estimates below 200K, your seed is too small or too niche. If it's above 5M, your seed might be too broad (not specific enough to your best customers).
What's the minimum seed size for a lookalike?
Meta officially requires 100 people, but performance is unreliable below 500. Between 500–1,000 is the sweet spot for most businesses. Above 1,000, you're usually including some lower-intent users, which dilutes the signal.
Should I exclude my lookalike audience from my broad campaigns?
Yes. If you're running both, exclude the lookalike audience from broad campaigns to avoid overlap and wasted spend. Meta allows audience exclusions at the ad set level.
People Also Ask
How often should I refresh my lookalike audience?
Every 30–60 days with your most recent high-value conversions. This keeps the audience aligned with your current product, pricing, and market. An old lookalike seed will become less accurate over time as your business evolves.
Can I create a lookalike from website visitors instead of customers?
You can, but it usually underperforms. A lookalike from all visitors will find people who look like clickers, not buyers. Use lookalikes seeded from your highest-intent group (paying customers, qualified leads, high-value accounts). If you only have visitor data, use broad targeting instead.
What's the difference between a 1% lookalike and a 10% lookalike?
A 1% lookalike finds people most similar to your seed audience (smallest, most precise). A 10% lookalike expands further, including people less similar to your seed (larger, more exploratory). Start with 1% for tight control; expand to 5–10% if you need more volume and conversion rates are still strong.
Why is my broad targeting campaign not improving after 2 weeks?
It may still be learning. Broad targeting needs 50+ conversions per week to optimize effectively. If your conversion volume is low, the algorithm doesn't have enough signal. Increase budget, improve your conversion event quality, or wait another 1–2 weeks. If it still hasn't improved by week 4, the campaign structure (creative, offer, audience location) may be wrong.
Can I use a lookalike audience for a new product launch?
Only if you have a seed audience from a related product or service. If it's a brand new product with no customer history, use broad targeting. Once you have 500+ customers on the new product, switch to lookalikes for scaling.
Is a lookalike audience better for brand awareness or conversions?
Lookalikes are better for conversions. They're built on intent-based signals (actual customer behavior). For brand awareness, broad targeting or interest-based targeting works better because you're optimizing for reach, not conversion rate.
What if my conversion rate drops after switching from lookalikes to broad targeting?
That's normal. Broad targeting finds new customer types who may have different purchase patterns. Track ROAS (return on ad spend), not just conversion rate. A lower conversion rate with a lower cost-per-click can still deliver better ROAS. Give the campaign 3–4 weeks to stabilize before deciding.
Should I use lookalikes for retargeting campaigns?
No. Retargeting uses pixel-based audiences (people who visited your site). Lookalikes are for cold prospecting. Use lookalikes for new customer acquisition; use retargeting for people who already engaged with your brand.
If this post is wrong, outdated, or you would take a different path
I write from work I have done on real sites. Search products change, and a step that was right when I published can go stale. I can also be wrong about the method.
If you disagree with the approach, the facts, or the outcome, I want the detail. Tell me what is off, what you would do instead, and where you saw it. I use that to correct the post so the next reader is not stuck.
This is not a comment thread. Use Contact me so the note is tied to this post and I can reply.
You are sending feedback for
Lookalike Audiences vs. Broad Targeting: When to Use Each in Meta Ads
Meta Ads
https://hammadshk.com/blog/lookalike-audiences-vs-broad-targeting-when-to-use-each-in-meta-ads
