Facebook Lookalike Audiences Australia: How to Find More of Your Best Customers
Facebook Lookalike Audiences are one of the most misunderstood and under-utilised tools in paid advertising.
Here’s the concept: You have 500 customers. You’re happy with them—they buy, they stay loyal, they refer others. You upload them to Facebook. Facebook finds 10,000 people similar to them. You show ads to those 10,000. A significant percentage becomes customers.
This is lookalike audiences. And when it works, it’s one of the cheapest ways to scale customer acquisition.
But it only works if:
- You start with the right “source audience.”
- You understand what size lookalike to create (1% vs. 2–10%).
- You know how to layer them with other targeting.
This guide covers how lookalike audiences work, how Meta builds them, what source audiences actually convert, how Australian businesses are using them, and what’s changed since iOS 14.
What Is a Lookalike Audience and How Does It Work?
You give Facebook a “source audience” (your best customers, people who bought from you, or people who watched your videos). Facebook then finds people in its database who share similar characteristics:
- Demographic: Age, gender, location, education, job.
- Behavioural: Interests, online habits, purchase history, engagement patterns.
- Psychographic: Values, personality traits, lifestyle choices.
Facebook doesn’t reveal exactly how it matches (it’s proprietary), but the general logic:
“These 500 people who bought from you tend to be women 25–40, interested in fitness and wellness, live in Sydney, and engage with health content. Find 10,000 people who match this profile.”
Meta’s lookalike algorithm then finds those people and ranks them by similarity.
Why Lookalike Audiences Work So Well
Cost per acquisition is lower than cold audiences because:
- The audience already has traits of your best customers.
- Higher conversion rate (4–8% vs. 1–2% for cold audiences).
- Lower cost per result (20–40% cheaper than broad interest targeting).
Real example: A Melbourne beauty brand found that cold Instagram audiences cost $30 per lead. Their lookalike audience (sourced from purchasers) cost $12 per lead. Same ad, different audience.
The difference: intent and affinity. Cold audiences have no reason to care. Lookalike audiences have characteristics of people who already do care.
Source Audiences: What Actually Works
Not all source audiences are equal. Some create high-converting lookalikes. Others don’t.
Source Audience Ranking (Best to Worst)
Tier 1 (Best):
- Purchasers (last 90 days). People who bought from you. These are your most valuable customers. Create a lookalike from them first.
- High-value customers (if you can segment). Customers with >$1,000 lifetime value.
- Engaged video viewers (watched 50%+ of a video). High intent and attention span.
Tier 2 (Good):
- Website visitors (last 30 days). People who visited but haven’t purchased. Lower quality than purchasers but still warm.
- Email list (engaged subscribers). People who opened 3+ emails. High engagement, lower quality than purchasers.
- Engagement audience (liked, commented, shared posts). People who engaged with your content.
Tier 3 (Okay):
- Page Followers / Instagram Followers. Broad audience, mixed intent.
Tier 4 (Weak):
- Website visitors (last 365 days). Too broad; includes people who visited once and forgot about you.
- People who clicked a link. Very low intent.
Rule: Use your BEST customer segment as your source audience, not your broadest. A lookalike of 100 purchasers will outperform a lookalike of 10,000 random website visitors.
Creating a Lookalike Audience in Ads Manager
Step-by-step:
- Go to Ads Manager > Assets > Audiences (left sidebar).
- Click Create Audience > Lookalike Audience.
- Choose your source audience:
- If sourcing from customers: Go to Custom Audience > Customer File and upload your customer list (CSV of emails, phone numbers, or IDs). Wait for the audience to build (usually 24–48 hours). Then create lookalike from it.
- If sourcing from website visitors: Go to Custom Audience > Website Traffic and select visitors from last 30 days. Create lookalike from it.
- If sourcing from engagement: Go to Custom Audience > Engagement and select people who watched your video. Create lookalike from it.
- Choose your country (Australia).
- Choose your audience size:
- 1% Lookalike: The closest match to your source audience. Smallest audience. Highest quality. Best for scaling profitable campaigns.
- 2% Lookalike: Slightly broader. More people, slightly lower quality.
- 5% Lookalike: Broader. More volume, lower conversion rate.
- 10% Lookalike: Broadest lookalike. Most volume, lowest quality.
- Name it (e.g., “1% Lookalike – Purchasers Q1 2026”) and save.
- Wait 24 hours for Meta to build the audience. Size will be shown (e.g., “1.2 million people”).
Done. Now you can target this audience in campaigns.
1% vs. 2–10%: Which Size to Use
This is where most Australian advertisers mess up. They jump straight to 10% lookalikes.
1% Lookalike
Audience size: Roughly 1% of your country’s population. For Australia, that’s ~250,000 people.
Quality: Highest. These are people most similar to your source audience.
Cost per result: Lowest (because quality is highest).
Conversion rate: 4–8%.
Use when: You’ve found a profitable campaign and want to scale. You know the unit economics work.
Risk: Might run out of audience (smaller pool). You could saturate 1% lookalikes and see rising costs.
2–5% Lookalikes
Audience size: Slightly broader. 2–5% of Australia. 500k–1.25M people.
Quality: Good. Still similar to your source audience, but with some variation.
Cost per result: Slightly higher than 1%, but still much lower than cold audiences.
Conversion rate: 3–5%.
Use when: You’re testing lookalikes for the first time. You want a balance of size and quality.
Advantage: Bigger audience than 1%, so you’re less likely to saturate it. Good for medium-spend campaigns.
10% Lookalike
Audience size: Broadest. 10% of Australia. 2.5M+ people.
Quality: Lowest. Less similar to source audience, but still warmer than cold interest targeting.
Cost per result: Higher, but still typically lower than cold audiences.
Conversion rate: 2–3%.
Use when: You have a very profitable campaign and need massive scale. You’ve tried 1% and 5% and maxed them out.
Risk: Cost per result is higher. If your margins are tight, 10% might not be profitable.
Strategy: Ladder Your Lookalikes
Start narrow, expand as you scale:
Month 1: Test 1% Lookalike at $20–$30/day for 7 days. If cost per result is acceptable, move to Month 2.
Month 2: Keep 1% running, add 2–5% Lookalike at $20–$30/day. See which performs better.
Month 3: Scale winner. Pause underperformer. Continue until costs rise above your target.
Month 4+: Add 10% Lookalike only if 1–5% are saturated (rising costs).
This way, you always prioritise quality over volume.
Creating Multiple Source Audiences (Advanced)
If you have different customer segments, create different lookalikes.
Example: A Brisbane fitness brand with three customer types:
- High-ticket customers (paid $2,000+ for coaching program). 50 people. Create 1% lookalike.
- Mid-ticket customers (paid $500 for 3-month program). 200 people. Create 1% lookalike.
- Low-ticket customers (bought $49 digital guide). 2,000 people. Create 1% lookalike.
Now you have three lookalike audiences, each pulling from a different customer segment. Test which converts best.
Most likely: High-ticket lookalike converts better (though audience is smaller). Low-ticket lookalike has more volume but lower conversion.
Allocate budget accordingly: 50% high-ticket, 30% mid-ticket, 20% low-ticket lookalike.
Layering Lookalike Audiences with Other Targeting
Lookalikes work even better when combined with other targeting.
Lookalike + Age Targeting
Example: 1% Lookalike from purchasers, but only women 25–40.
This narrows the audience but increases quality. Useful if your source audience skews a specific age/gender.
Lookalike + Location
Example: 1% Lookalike, but only people in Sydney metro.
Useful for local businesses or if your source audience is geo-concentrated.
Lookalike + Interest Stacking
Example: 1% Lookalike + interested in fitness + interested in wellness.
This is powerful. You’re saying: “Find people like my best customers AND who are interested in fitness.”
Much smaller audience, much higher intent.
Lookalike + Engagement
Example: 1% Lookalike + people who watched 50%+ of a video.
You’re targeting people like your best customers, who have also shown video engagement. Very high quality.
Best practice: Layer lookalikes with 1–2 additional targeting parameters. More than that and audience becomes too small.
iOS 14 Impact on Lookalike Audiences
Apple’s iOS 14 privacy changes affected lookalike audiences, but less severely than they affected pixel tracking.
What changed:
- Meta can no longer see precise demographic data from iOS users.
- Lookalike audiences are slightly less accurate (maybe 10–20% lower conversion rate than pre-2021).
- Smaller source audiences (under 100 people) are harder to find matches for.
What didn’t change:
- Lookalikes still work.
- Cost per result is still lower than cold audiences.
- They’re still one of your best scaling tools.
What you should do:
- Use larger source audiences (500+ people, not 50).
- Combine lookalikes with other targeting to layer in intent.
- Test more frequently (iOS tracking is less precise).
Real Australian Examples
Example 1: Melbourne E-Commerce (Fashion)
Business: Online women’s fashion retailer.
Source audience: 200 customers who purchased in last 90 days.
Lookalike: 1% Lookalike Australia.
Audience size: 250k people.
Test: $30/day for 7 days.
Results:
- CPM: $6
- Clicks: 420
- CPC: $0.71
- Purchase conversions: 8
- Cost per purchase: $26.25
- Average order value: $95
- ROAS: 3.6:1
Verdict: Profitable. Scale to $50/day. Test 2–5% lookalike simultaneously.
Example 2: Sydney B2B Consulting (Leads)
Business: Management consulting firm.
Source audience: 150 clients (high-ticket, >$50k projects).
Lookalike: 1% Lookalike Australia, layered with job titles: “Director, VP, C-Suite.”
Audience size: 80k people (small because job title filter).
Test: $30/day for 7 days.
Results:
- Impressions: 2,100
- Clicks: 140
- CPC: $1.50
- Lead form submissions: 12
- Cost per lead: $175
- Close rate: 25% (3 deals)
- Deal size: $100k average
- Revenue: $300k
- CAC: $17.50k across all channels (consulting, not just ads)
Verdict: Profitable at enterprise deal sizes. Scale to $100/day. Lookalike + job title targeting is highly effective for B2B.
Example 3: Brisbane Personal Training (Lead Gen)
Business: Personal training studio.
Source audience: 80 past and current clients.
Lookalike: 1% Lookalike, layered with interests: fitness, wellness, health, personal training.
Audience size: 120k people.
Test: $20/day for 7 days.
Results:
- CPM: $8
- Clicks: 280
- Cost per click: $0.50
- Form submissions: 25
- Cost per lead: $28
- Trial conversion rate: 40% (10 new clients)
- Average lifetime value: $2,500
- Profit: $25,000 – $1,400 (ad spend) = $23,600
Verdict: Highly profitable. Scale to $50/day. Plan to test 2–5% lookalikes next week to find even more customers.
Common Lookalike Mistakes
- Using a weak source audience. You source from “all website visitors.” Result: low conversion lookalike. Always use your best customers.
- Jumping to 10% lookalikes. You test 10% first. Cost per result is too high. Fail. Don’t abandon lookalikes; test 1% first.
- Not giving lookalikes enough budget. You allocate $5/day. Not enough data by day 7. Always minimum $20–$30/day for 7 days.
- Over-layering targeting. You layer lookalike + 5 interests + age + gender + job title. Audience becomes too small. Stick to 1–2 additional parameters.
- Not refreshing source audience. You created a lookalike from Q1 customers. Now it’s Q4. Create a new lookalike from Q3–Q4 customers. Your business has evolved.
- Confusing lookalike with interest targeting. Lookalikes are fundamentally different from interest-based audiences. Don’t mix them up.
Your Lookalike Testing Checklist
- [ ] Identified your best customer segment (not all customers).
- [ ] Created a customer list (emails, phone numbers, or IDs).
- [ ] Uploaded to Facebook as a custom audience.
- [ ] Waited 24–48 hours for audience to populate.
- [ ] Created 1% lookalike audience.
- [ ] Named it clearly (“1% Lookalike – Purchasers Q2 2026”).
- [ ] Created a campaign targeting this lookalike.
- [ ] Set budget: minimum $20–$30/day.
- [ ] Ran for minimum 7 days.
- [ ] Checked cost per result. Is it below your target?
- [ ] If yes: Scale or test 2–5% lookalike. If no: Test different source audience.
When Lookalikes Don’t Work
Lookalikes can fail if:
- Your source audience is too small (under 100 people). Meta can’t find reliable matches.
- Your source audience is too broad (all website visitors). Lookalike is diluted. Use only your best segment.
- Your product has no clear customer profile (niche with scattered demographics). Lookalike won’t find reliable matches.
- Your conversion rate on-site is already low (under 1%). Lookalike won’t fix this. Fix your product first.
- Your margins are razor-thin (under 2:1 ROAS required). Even cheaper traffic won’t help.
In these cases, focus on retargeting and Google Ads (higher intent). Lookalikes work best for clear customer archetypes with reasonable margins.
Ready to scale with Lookalike Audiences? Anitech builds and tests lookalike audiences for Australian businesses. We identify your best customer segments and create lookalikes that scale profitably.
Let’s find more of your best customers and build a lookalike strategy that works.
Internal links: