Supercharging User Acquisition with Fusion™ AI

Since Facebook Ads first introduced lookalike modeling in 2013, the landscape of ad targeting has undergone a significant transformation. This innovative feature allowed marketers to reach new potential customers by identifying those who resembled their best existing customers, revolutionizing digital advertising. Lookalike modeling quickly became a cornerstone of targeted advertising strategies, enabling brands to expand their reach and increase conversion rates by focusing on audiences most likely to engage with their offerings.

However, as the digital ecosystem has evolved, so too have the challenges of accurately identifying and targeting the right audiences. Traditional lookalike models, while effective, often rely on basic, predefined metrics that capture only surface-level similarities. As a result, they can miss the deeper, more complex patterns that truly define customer behavior. This is where Fusion steps in, offering a new, data-driven approach to lookalike modeling that goes beyond conventional methods, unlocking richer insights and delivering more precise audience targeting.

Fusion is an AI model designed to revolutionize the way businesses approach lookalike modeling. Unlike traditional methods that rely on predefined metrics and basic data attributes (such as demographic information, past purchase behavior and other surface-level features), Fusion employs a fully data-driven, unsupervised learning approach, offering a level of depth and precision that conventional systems simply cannot match. By leveraging deep learning techniques, our model is capable of autonomously discovering complex, high-dimensional patterns within customer data, allowing it to uncover the latent factors that truly define customer behavior. 

For example, Fusion doesn’t just look at what customers have purchased in the past; it delves into the intricate web of correlations and relationships within the data. It might uncover that a particular group of customers, who at first glance appear dissimilar, actually share a subtle, yet significant, preference that could make them ideal candidates for a new product line. This deeper understanding enables businesses to create more accurate and nuanced representations of their customer base, leading to more effective lookalike modeling and, ultimately, better marketing outcomes.

Another key advantage of Fusion is its scalability and adaptability. As the market evolves and new data becomes available, our model is designed to be applied seamlessly to new data sets, and then to continuously learn and improve. This dynamic capability is particularly valuable in today’s data environment, where privacy regulations often limit the type of data that can be used in audience modeling. Unlike traditional systems, which often require manual updates and tweaks to remain effective, Fusion can automatically process any type of consumer data and adapt to new information, ensuring that advertising efforts are always aligned with the latest insights.

In addition to its ability to uncover hidden patterns, Fusion also excels at integrating diverse types of data. Today’s customers interact with brands across multiple channels and generate a wide array of data – from transactional records and social media activity to text and image data. Traditional methods often struggle to combine these disparate data sources in a meaningful way, leading to fragmented insights and missed opportunities.

Fusion, on the other hand, is designed to process and integrate multimodal data, providing a more holistic view of each customer. This capability is deployed across Fantix’s data federation, a network of sources that make their data viewable by Fantix’s federated machine learning tools to train AI models. The federation includes credit card and open banking transactions with product and SKU-level granularity, fintech app data, sports intelligence, demographic information and many other data sets that Fusion can analyze and draw targeting audiences from. 

The coupling of Fusion’s AI-driven modeling with the depth and width of data in Fantix’s federation produces great results for advertisers across a wide spectrum of ad solutions: from CTV to programmatic and from Meta Ads to Snapchat and X, targeting audiences generated with Fusion have yielded CAC reductions as high as 45%.

Fintech and Insurtech players have been the earliest adopters of Fusion’s technology, and will benefit from it particularly in light of Meta’s upcoming expansion of special ad categories to include financial products and services. As we head into Q4, other leaders are beginning to experiment with Fusion and are poised to bring the benefits of federated audience building to the e-commerce and sports gaming industries, as well.