US & Global
Running B2B ad campaigns on Meta often gets dismissed because native job title and industry targeting options can feel broad compared to LinkedIn's expensive ecosystem. However, Meta's algorithm excels when fed precise data. To solve the targeting gap without paying exorbitant cost-per-click rates, we built an acquisition loop that extracts targeted B2B decision-makers from Apollo.io and feeds them into Meta as a seed list to generate a Lookalike Audience. Here is a look at the process behind setting up this segment.
The performance of any Lookalike audience relies entirely on the purity of the underlying seed data. If you feed the algorithm irrelevant contacts, it simply optimizes toward similar irrelevant profiles. A single bad batch pollutes the entire matching engine, which is why we don't just dump broad keyword searches into a CSV. Instead, we layer filters intentionally.
For this campaign supporting Vertical atQuo, we focused on finding decision-makers across Latin America who need modular marketing execution without long-term agency retainers.
.gif)
When building a search in Apollo, you can take two paths: setting up filters manually or using Apollo's native AI search prompt.
If you choose the manual route, you select filters step-by-step from the sidebar. We scoped job titles to leaders holding direct budget authority, such as CMOs, VPs of Marketing, Heads of Marketing, Marketing Directors, Founders, and CEOs. We set company size specifically to 11 to 200 employees, which hits the sweet spot of businesses with active execution needs that prefer agile, fixed-scope marketing modules over adding internal headcount. Geographically, we targeted primary Latin America across sectors like Software, IT Services, Finance, and Healthcare.
Alternatively, you can let Apollo handle the setup by entering a natural language prompt directly into the search bar. Writing a clear prompt like "Find CMOs, Marketing Directors, and Founders at Software, Finance, and Healthcare companies with 11 to 200 employees in Latin America" automatically populates the corresponding filters across your workspace.
Next comes filtering, where strict data quality takes priority. We toggled the option to require an existing company domain, stripping out inactive projects or dead websites. We also explicitly filtered for verified emails only. This guarantees that every identifier handed over to Meta belongs to an active inbox.
Equally important was what we left out. Using explicit industry exclusions, we omitted marketing, advertising, PR, and design agencies. Reaching decision-makers at competing agencies wastes ad spend on people who will never buy our service modules.
When exporting from Apollo, understanding how export credits work helps optimize budget. Unlocking standard corporate email records consumes 1 export credit per contact. If your credit budget allows, unlocking phone numbers using phone credits (8 credits per phone number) offers an even stronger signal. Decision-makers rarely register social accounts with work emails, but mobile numbers link directly to personal WhatsApp, Instagram, and Facebook profiles, driving significantly higher match rates in Meta.
Whether exporting phone numbers or standard emails, Meta requires a minimum of 100 matched profiles in a single country to build a Lookalike Audience. Because industry benchmarks for B2B lists typically sit between 25% and 40% on Meta1, starting with a seed list of a few hundred verified contacts ensures you comfortably clear Meta's threshold.
To maximize the match rate beyond primary emails, we exported a complete matrix of identifier columns like: First Name, Last Name, Primary Email, Company Name, City, State, and Country. Providing these location and naming parameters gives Meta's matching engine multiple data points to connect a B2B work identity to an active personal profile.
.png)
Once the CSV export is ready, the next phase takes place in Meta Ads Manager under Audiences. We navigate to Create Audience > Custom Audience > Customer List. On the setup screen, we select No for customer value, accept Meta's Custom Audience terms, and upload our file.
Meta's matching engine automatically scans the CSV headers and pairs them with native identifiers. In our setup, five core fields map directly: First Name, Last Name, Email, City, and Country. Reviewing these fields ensures no unmapped columns slip through. Clicking Import & Create completes the upload, generating our primary seed audience.
With the seed audience established, we move directly to creating the Lookalike Audience. Selecting our newly uploaded customer list as the source, we set the audience size slider to 1%. This tells Meta's algorithm to analyze the core behavioral and demographic traits of our seed contacts and find the top 1% of similar users across the population.
Once submitted, Meta marks both the Custom Audience and the Lookalike Audience in the dashboard as Populating initially before transitioning to Ready shortly after. This allows you to attach them directly to your campaign ad sets.
With our 1% Lookalike audience generated, the pipeline setup is complete and ready to power our upcoming Meta ad experiments.
1 Benly.ai. Meta Ads Custom Audiences Guide: Match Rates and Best Practices. Available at: https://benly.ai/learn/meta-ads/custom-audiences-guide
Strategy, design, content and growth.
