Predictive audiences vs. lookalikes: How to unlock omnichannel

Until recently, single-platform lookalike audiences were the quickest and simplest options for marketers to reach likely-buyers. Now, there are off-the-shelf AI solutions designed to help brands expand beyond those siloed ad platforms.

Predictive audiences vs. lookalikes: How to unlock omnichannel

Lookalike audiences have long been a favorite of marketers, and for good reason. Nearly every ad platform gives advertisers the ability to reach these AI-generated, “likely-to-buy” audiences with a few clicks. It’s a powerful capability that’s included in the cost of advertising on those platforms.

However, single-platform lookalike audiences limit brands’ abilities to create truly unified omnichannel experiences for modern consumers. Furthermore, the deprecation of third-party cookies is limiting the effectiveness of the ad platforms’ native AI-driven audience targeting features.

Brands are now looking beyond single-platform lookalike audiences and adopting alternative forms of predictive audiences to future-proof their targeting capabilities across the dozens of channels and devices that consumers interact with on a daily basis. Read on to explore how the two approaches stack up.

Lookalike audiences for single-platform campaigns

The main benefit of single-platform lookalike audiences is that they’re very easy to use and costs are usually baked into the price of advertising on each platform. Anyone advertising on Facebook can create a lookalike audience with just a few clicks: upload or define your seed audience, select your target audience size, and you’re good to go. Facebook’s algorithms will take it from there to build a lookalike audience of people who share similarities with your seed audience.

However, when it comes to the nuts and bolts of building a lookalike audience, platforms like Facebook face some substantial limitations. Their machine learning algorithms process an enormous trove of data to surface new audiences that share the most traits with yours, but much of that data is collected using third-party cookies, which are on track to be fully phased out in the next couple of years. Already, performance on Facebook is plummeting in the wake of iOS 14 updates that limit cross-channel tracking.

Additionally, ad platforms take most of the control over where marketers can reach the audiences they create on each platform, making it nearly impossible for brands to reach the same audiences and deliver consistent experiences across multiple platforms. Lookalike audiences created on Facebook will differ from those created on Snapchat, Pinterest, Tik Tok, etc.

Predictive audiences for omnichannel strategies

There’s no denying that when it comes to most marketers’ channel mix, more really is better. Marketers leveraging three or more channels see an eye-popping 494% higher order rate than those using a single channel for their campaigns. And predictive targeting is crucial to maintaining effective performance while scaling your spend.

Rather than wasting ad spend on bad fits, or trying to reach disparate groups spread out across online and offline sources, marketers can use predictive audiences to hone in on the people that are already most likely to become their next best customer—and then deliver harmonious messaging to them on every channel.

Until recently, single-platform lookalike audiences were the quickest and simplest options for marketers to reach likely-buyers. Now, companies like Faraday are providing relatively off-the-shelf AI solutions designed to help brands expand their predictive marketing capabilities beyond those siloed ad platforms—predictive audiences being one of the most widely used applications of these emerging AI platforms.

Fundamentally, predictive audiences are built using the same approach as single-platform lookalike audiences; machine learning algorithms crunch the numbers on a broad population—say, every adult in the US—to predict the likelihood that each individual will exhibit a specific behavior. Isolate those who are most likely to do so and you’ve got your predictive audience.

Here are a few examples of what that might look like in practice:

  • A bank might predict who’s likely to apply for a mortgage
  • A wellness brand might predict who’s likely to purchase their best-selling moisturizer
  • A subscription meal prep service might predict who’s likely to subscribe to their premium package
  • A solar installer might predict who’s likely to book a free consultation

You can also build predictive audiences that target actions even earlier in the buyer’s journey, like clicking on an ad or adding an item to the cart. Likewise, there's the option to predict lower-funnel actions including repeat purchases, upgrades, or unsubscribes.

Unlike with single-platform lookalike audiences, third-party AI platforms like Faraday allow users to create predictive audiences with cookieless data sources, and generally support cross-platform activation. This enables brands to deliver more consistent experiences across multiple channels, devices, and customer journey stages.

With predictive audiences, an audience that’s interacting with your carefully curated Facebook campaign will also see the same campaign on Snapchat, Pinterest, and Tik Tok for a totally seamless experience. If they’ve subscribed to your emails or joined your membership program, you can also extend your touchpoints to email marketing, SMS alerts, app notifications, and more.

For marketing teams looking to take an even more nuanced approach, predictive audiences unlock all sorts of other doors. You might want to launch a campaign for a particular segment of a predictive audience on one channel, having crafted a totally personalized experience for that group. Then, you can expose that audience to another variation of that messaging on a second channel, and experiment with a third version on yet another one. Incrementally moving your audience down the funnel like this gives you even more control over the buyer’s journey.

Get started with predictive audiences today

You know you can log on to Facebook to build lookalikes anytime—so where do you get predictive audiences?

There are plenty of software solutions out there that allow you to create predictive audiences in just a couple clicks, making it just as easy as Facebook does. Here at Faraday, we’ve built predictive audiences for hundreds of brands, encompassing countless practical use cases and the kind of tangible results that make this strategy a no-brainer.

See how Burrow used predictive audiences