Flavors of Lookalike Audiences: From Pixel-Based to Predictive Value-Based Seeds

As consumer brands strive to make their marketing campaigns more effective, lookalike audiences have become a crucial tool for reaching new customers who share similar traits and interests to their existing ones. However, the way in which these audiences are created has undergone significant changes in recent years. (Read our earlier article here on lookalike audiences and signal loss.)

In this post, we'll explore the evolution of lookalike audiences from pixel-based to predictive value-based seed, and why this shift is important for marketers at consumer brands.

Pixel-Based Lookalike Audiences

Pixel-based lookalike audiences are created using data collected by pixels on your website. Pixels collect data on users who visited specific pages or completed certain actions, such as adding an item to their cart or making a purchase. This data is then used to create a lookalike audience of users who share similar traits and interests to the original group.

However, over time, signal loss has made it increasingly difficult to rely solely on pixel data. Changes to browser tracking, consumer opt-outs, and other factors have reduced pixels’ signal volume (pixels don’t fire on all pages or with all actions) which means that ad platforms have an incomplete picture of users' behaviors.

Customer Data as Seed

Following signal loss, a better seed for lookalike audiences is your first- and zero-party data. This data includes information on your existing customers, and by including CRM data, also those who engage with your brand online and offline. By sending this data to ad platforms from your CRM or e-commerce platforms, you can create a lookalike audience based on your most up-to-date customer data.

Value-Based Lookalike Audiences

Value-based lookalike audiences assign values to your customer seed list, providing direction to the ad platforms on which seeds equate to more value for your business. The ad platform then finds more people that fit that enhanced customer list and thus, builds more powerful lookalike audiences. Values can be the average order value of the customer, or even predictive LTV which is a modeled, future view of your LTV value. Value-based lookalike audiences can also be useful to suppress likely low lifetime value customers from your broad campaigns.

Predictive Value-Based Seed

But what if you could take it a step further and create a lookalike audience based on more than just available data, but also with additional filters and the predicted value of those customers? This is where predictive value-based seed comes in.

With predictive value-based seed you can layer on predictions based on who your users are and what they are doing outside your brand, implemented by segmenting your data based on filters. You can filter on customer interests and behavior, such as focusing on holiday gifters or discount purchasers. You can even build lookalike audiences with customers likely to repurchase in the next 90 days to improve long-term ROAS.

You can also refine your seed by exploring where your customers spend money on brands outside of your own. For example, let's say you run a luxury home decor brand and you notice that many of your customers also make purchases from Williams Sonoma, a high-end kitchenware retailer. By analyzing this pattern, you can create a predictive value-based seed that identifies customers who have previously made purchases from both your brand and Williams Sonoma. You can then create a lookalike audience that targets individuals who share similar characteristics to these high-value customers and align your creative based on the prospect’s interests and affinities, expanding your reach to find new customers who are more likely to make a purchase from your brand.

Advanced Matching

By using a predictive value-based seed, businesses can leverage the advanced matching capabilities offered by Prospect Data Platforms, resulting in yet another advantage. Ad platforms like Meta maintain a "graph" of all its users, including all the information and insights they have on each person. However, with consumers having multiple touch points beyond ad platforms (devices, accounts, etc.), it can be challenging to match all your seed customers with the platforms’ users. Prospect data platforms like Angler AI can help by acting like an on-boarder, filling in missing information or attributes to your customer seed list. When you share that on-boarded information, the probability of matching the seed customer with the user on the ad platform increases, in turn helping the ad platform better advertise your product to its user. This can help improve the accuracy of your lookalike audiences and make your marketing campaigns more effective.

Angler AI's Prospect Data Platform

The evolution of lookalike audiences from pixel-based to predictive value-based seed has been driven by the need to find new and more effective ways to reach potential customers. To make the process of crafting predictive value-based seeds more accessible to marketing teams, Prospect Data Platforms like Angler AI can help. With Angler, marketers can easily create predictive value-based seeds and seamlessly push them to ad platforms like Meta for campaign optimization, while reducing the toil for marketing teams. Request a demo now to learn more about how Angler can help you improve your marketing efficiency and scale your business in a cookieless world.