Custom Audiences

When a company runs advertising, it has two fundamental options. It can target as broad a mass as possible and hope that some portion of the viewers is interested. Or it can use the data it already has about its own customers and address those people directly. This is exactly what lies behind the term Custom Audiences: the advertiser uploads a list — for example, the email addresses of existing customers — to an advertising platform. The platform matches this list against its own user accounts. The result is an audience made up of real, already known people, not statistical assumptions about an average person.

Why Custom Audiences make ad budgets more efficient

Broadly scattered advertising is expensive because a large part of the budget reaches people who will never buy. Custom Audiences solve this problem. Anyone running an ad for a product can show it specifically to people who have already viewed the product but haven’t bought it yet. This noticeably lowers the cost per purchase.

The principle becomes especially powerful in combination with so-called Lookalike Audiences — audiences that a platform assembles automatically because their members are statistically similar to one’s own customers. The Custom Audience serves as the template here. This allows a company’s own customer base to be used as a compass to find new prospects.

Data matching in the background

Technically, the process runs through a comparison known as matching. The company exports, for example, a list of email addresses from its own system. Before the list is uploaded, software converts the addresses into so-called hash values: short strings of characters that act like a fingerprint for each address, but from which the original address can no longer be recovered. The platform compares these hash values with the stored values of its own user accounts. Anyone who appears in both lists ends up in the Custom Audience.

Besides email addresses, phone numbers, user IDs, and website visitors also work as source data. Website visitors are captured via a small piece of code embedded on the company’s site — on Facebook this piece of code is called the Meta Pixel, while on Google it’s called a tag. Anyone who visits the site leaves a trace this way, which is later used to build the audience.

A common misconception is that the company hands over real addresses to Facebook or Google in the process. That’s not quite true. The matching happens exclusively via the hash values, and the platform never learns which email address belongs to which account. However, these protective mechanisms are not a free pass: in Europe, the General Data Protection Regulation, or GDPR, regulates what data companies are even allowed to collect and share in the first place.

Custom Audiences in products and headlines

The best-known implementation is the Custom Audiences feature from Meta, the company behind Facebook and Instagram. It was introduced in 2012 and is considered one of the main reasons Facebook became attractive as an advertising platform for small and medium-sized businesses. Similar tools are offered by Google Ads, TikTok, and Pinterest under various names.

Custom Audiences tend to make the news mainly when privacy advocates question the practice. In 2018, it emerged that some companies had used phone numbers that users had only provided for account security on Facebook for targeted advertising. Meta subsequently changed the rules for permissible data sources. Since then, the topic has been a recurring fixture in the discussion about where useful targeting ends and unwanted surveillance begins.

For consumers, this means: anyone wondering why, shortly after visiting an online shop, ads for that exact shop appear on Instagram is, in all likelihood, dealing with Custom Audiences. The experience often feels like coincidence, but it is actually the result of a precisely planned matching process.