You spent budget on Meta Ads, but clicks aren't converting into sales? The problem isn't the ad — it's who sees it. A perfect copy and award-winning creative are useless if shown to people who will never buy from you. We, at Meteora Web, see this daily: clients who come with campaigns already poorly spent, hundreds of euros wasted on audiences that are too broad or misaligned. Targeting makes the difference. In this guide we dive into the three main tools: custom audiences, lookalike audiences, and interest targeting. Why they work, how to create them, and where most people get it wrong.
Custom Audiences — Your Database is Gold
A custom audience is a set of users you have already contacted outside Meta: customers who purchased, website visitors, people who used your app, or those who interacted with your content (videos, lead forms, Page). It's the most precise targeting because it works on your own data. If you run a physical store in Sciacca and upload a list of loyal customers, you can show them exclusive offers on Meta. Nothing beats that.
Why It Works
Because you re-engage users who already know you. Acquisition cost drops, conversion rate rises. We have an e-commerce clothing client: using a custom audience of customers from the last 90 days reduced CPA by 40% compared to cold campaigns. The reason is simple: trust is already built.
How to Create Them – Step by Step
You can create a custom audience from several sources. Here are the most commonly used:
- Customer list: upload a CSV file with emails, phone numbers, or user IDs. Meta matches them to users on the platform. Ensure data is clean and up-to-date. No duplicates or non-consensual emails.
- Website visitors: you need the Meta Pixel (or Conversions API) installed. Then in Audience Manager create an audience based on pages visited, time on site, specific events (e.g., “Added to cart but not purchased”).
- App activity: track in-app events via the Meta SDK.
- Engagement: people who watched your video (e.g., 50%), interacted with your Page, or opened a lead form.
We recommend starting with the “All website visitors” audience for general retargeting, then segmenting by behavior (e.g., “Abandoned cart”).
Common Mistakes
- List too small: under 1,000 users Meta struggles to find matches. Aim for at least 5,000 for lookalikes.
- Outdated data: a customer list from 2 years ago is almost useless. Refresh every 30–60 days.
- Ignoring privacy: if you collect data without consent, you violate GDPR. Use only clean lead magnets and opt-ins. We come from accounting and data management: compliance is a hidden cost if you mess it up.
Lookalike Audiences — Expand Without Losing Quality
A lookalike is an audience created by Meta starting from a “seed”. The system analyzes the characteristics of people in the seed and finds similar users across the platform. It's the best tool for targeted prospecting instead of spray-and-pray.
The Secret is the Seed
The quality of a lookalike depends entirely on the quality of the seed. A list of all customers is weak. A list of customers who spent over €500 and purchased 3 times is strong. A custom audience of website visitors who completed a conversion also works. The more specific and large the seed, the better Meta learns.
Sizing: 1% vs 10%
Meta lets you choose the percentage of the population to cover. 1% is the most similar to the seed, but reaches fewer people (about 800,000 in the US for the whole country). 10% is much broader, but similarity dilutes. Rule of thumb: for conversion campaigns, use 1% or 2%. For awareness or lead generation, you can go up to 5%. Beyond that, you risk drifting too far.
How to Create in Ads Manager
- Go to Audience Manager → Create Audience → Lookalike.
- Choose the seed: a custom audience (e.g., “High-value customers” or “Purchasers last 90 days”).
- Select the target country (if you sell only in Italy, choose Italy).
- Set the percentage (start with 1%).
- Give it a clear name, e.g., “Lookalike 1% – High-value customers – Italy”.
A tip: never use a lookalike generated from a seed of only visitors without conversions. Visitors can be random. Instead, use a conversion event like “Purchase” or “Lead”.
Interest Targeting — Understand Behaviors
Interests on Meta are not just “hobbies”. They are groupings based on user profile, pages followed, interactions, and off-platform purchases (via partners). There are declared interests (e.g., “Fitness enthusiast”) and behavioral ones (e.g., “Buyer of high-end sportswear”).
How to Find the Right Interests
Don't rely on intuition. Use Meta Audience Insights (free tool) to analyze a reference audience. For example, enter a competitor's Page or a broad interest like “Coffee shop”, and see other associated interests. Alternatively, start from your ideal customer: if you sell professional coffee machines for offices, look for interests like “Coffee machine”, “Business equipment”, “Modern office”. Combine with behavioral data like “Procurement manager at SMB” (if available).
Combining Interests and First-Party Data
The real leap is to use interests as an additional filter on lookalikes. For example: create a lookalike from your customers and then overlay an interest like “Coffee equipment”. This narrows down to an audience similar to your customers but with a specific interest. We used this technique for a client in the food & wine sector: lookalike 1% + interest “Wine tourism” halved CPA compared to pure lookalike.
Combined Strategy — A Three-Level Funnel
There's no single “best” targeting. There's the right targeting for each funnel stage.
- Retargeting stage: custom audiences (abandoned cart, recent visitors, past customers).
- Warm prospecting stage: lookalikes from high-value customers. Expand gradually (1% → 3% → 5%) if ROAS holds.
- Awareness or scaling stage: interests + wider lookalikes, or pure interests if you don't have first-party data yet.
The key is measurement. We crunch numbers: CPA, ROAS, average order value. If a 1% lookalike gives you CPA €20 but a 5% lookalike gives you CPA €30, you're using data poorly. Go back, refine the seed.
In Summary – What to Do Now
- Clean your customer database: export a list with emails, first and last name. Upload it to Audience Manager to create a basic custom audience.
- Install Meta Pixel and Conversions API (if you haven't already). Create an “All website visitors” audience and an “Abandoned cart” one.
- Identify a high-LTV customer segment (e.g., those who bought more than 3 times in the last year). Create a 1% lookalike from that seed.
- Use Audience Insights to find 3–5 relevant behavioral interests and test them in a separate campaign.
- Combine lookalike + interest for hybrid targeting. Monitor CPA for at least 2,000 impressions before judging.
If you're also working on organic growth, check our Instagram for Business guide to build a solid social base that can later feed your lookalikes.
For official resources, see Meta's Custom Audiences documentation and Lookalike Audiences guide.
Sponsored Protocol
