Azzam Bilal Chamdy
By Ronan Shields | June 19, 2026
In the high-stakes world of sports betting, where customer acquisition costs (CAC) can skyrocket and the lifetime value (LTV) of a user is the ultimate metric of success, traditional performance marketing has reached a point of diminishing returns. For years, the industry playbook was static: define a demographic, identify the channels they frequent, and bid aggressively for their attention.
However, a radical shift is underway. Fanatics Sportsbook, a titan in the digital gaming landscape, has recently abandoned the "audience-first" dogma in favor of a "system-led" approach. By partnering with ad-tech firm Cognitiv, Fanatics has demonstrated that machine learning (ML) is far more capable of identifying high-value customers than even the most sophisticated human marketing team. The result? A staggering 19% increase in projected customer lifetime value.
The Traditional Playbook: Why Demographic Targeting is Failing
For decades, the standard operating procedure for digital marketers has been built on the "persona" model. Whether it was the "25-to-34-year-old sports enthusiast" or the "weekend parlay bettor," marketers spent countless hours and millions of dollars constructing these profiles. The hypothesis was simple: if you find the right person, they will become a loyal customer.
But this model suffers from a fundamental flaw: it relies on human assumption. Marketers often conflate correlation with causation, assuming that because a user fits a certain demographic profile, they are predisposed to purchase. In reality, human behavior—particularly in the volatile world of sports betting—is rarely that linear.
"For years, performance marketers have built campaigns around the premise of defining the ideal customer, identifying where that audience spends time, and then placing media buys accordingly," notes industry analysts. But as Fanatics Betting & Gaming has discovered, relying on these manual constructs is no longer sufficient in an era where data-driven machine learning can map behavioral patterns that are invisible to the naked eye.
Chronology of a Paradigm Shift
The transition for Fanatics did not happen overnight. It was the result of a deliberate, iterative testing process that challenged the status quo of their marketing stack.
- Early 2025: The Hypothesis Phase. Fanatics began experimenting with Cognitiv’s ad-tech solutions, initially focusing on standard performance metrics like Cost Per Acquisition (CPA). They attempted to model audiences based on specific betting behaviors, trying to "nudge" the machine toward the right demographics.
- Late 2025: The Recognition of Limits. Despite achieving better results than traditional display ads, the internal team at Fanatics realized they were still hitting a ceiling. They were "getting ahead of themselves on targeting," as Andy Magnes, director of paid marketing at Fanatics Betting & Gaming, later put it. The bottleneck wasn’t the media buy; it was the strategy of pre-defining who those buyers were.
- Early 2026: The Outcome-First Pivot. Fanatics moved to a "full-funnel CTV solution" provided by Cognitiv. Instead of feeding the model a demographic profile, they fed it a business objective: find users who demonstrate high long-term value.
- Q2 2026: Validation. The shift proved immediate and profound. By removing the human-imposed constraints on targeting, the algorithms identified prospective bettors that the internal team had never considered, resulting in a 19% lift in LTV.
Supporting Data: The Power of Predictive ML
The success of this strategy rests on the technical architecture of Cognitiv’s platforms, specifically the integration of "AudienceGPT."
Traditional programmatic advertising functions on a rigid feedback loop. An advertiser sets a parameter (e.g., "Males, 21-45, interested in football"), and the system optimizes the spend to hit that specific cohort. In contrast, Fanatics’ new approach uses machine learning to train custom models against "advertiser-defined outcomes."
The data backing this transition is compelling. By optimizing for LTV rather than CPA, Fanatics shifted the focus of their CTV (Connected TV) spend from "volume of signups" to "quality of play." The 19% increase in LTV indicates that the machine-learning model was not just finding more customers, but finding better customers—users whose betting patterns, engagement frequency, and deposit behaviors correlated with long-term profitability.
Official Perspectives: The "Value-First" Philosophy
The shift in strategy has garnered significant attention from marketing leadership. Andy Magnes, a key architect of the campaign, shared his insights with Digiday regarding the dangers of over-targeting.
"We just started modeling on value," Magnes explained. "You can certainly get ahead of yourself on targeting." According to Magnes, the previous obsession with defining the "ideal" persona was essentially a way of guessing. By allowing the machine to optimize directly against value signals, Fanatics effectively removed the "guesswork" layer from the media buying process.
Jeremy Fain, CEO of Cognitiv, provided the technological context for this success. "Fanatics tells us who the high-value people are, and we go out and find more of those people," Fain said. He emphasized that the role of the marketer is shifting from "audience architect" to "objective setter."
However, Fain is quick to acknowledge that the human element hasn’t been completely erased. "Although at the top [of the marketing funnel], you do need a hypothesis… these prompts end up being that hypothesis." The difference is that in this new model, the hypothesis is the starting point for the AI, not the boundary line for the entire campaign.
Implications for the Broader Advertising Landscape
The implications of Fanatics’ success extend far beyond the sports betting sector. This shift marks a broader evolution in digital advertising—a move toward what industry experts are calling "outcome-based programmatic."
1. The Death of Demographic Segmentation
If a machine can find a high-value customer regardless of age, location, or declared interests, the entire premise of "audience segments" begins to crumble. We are moving toward a future where media is bought against intent and behavior rather than static labels.
2. The Rise of "Natural Language" Marketing
With tools like AudienceGPT, the barrier to entry for complex machine learning is lowering. Marketers can now use natural language to describe their business goals—"find me people who will stay active for six months"—and the system translates that intent into a technical execution. This allows for a more fluid, responsive relationship between creative strategy and technical implementation.
3. The Re-evaluation of the "Full Funnel"
Many marketers treat the funnel as a series of disconnected steps: awareness, consideration, and conversion. Fanatics’ success suggests that when you optimize for a bottom-line outcome (LTV), the funnel becomes a single, continuous loop. By feeding the top-of-funnel CTV efforts with bottom-of-funnel outcome data, the brand ensures that its largest media spend is always pointing toward the most profitable destination.
4. Competitive Advantage in a Saturated Market
In industries like sports betting, where the cost of entry is high and the competition for user attention is fierce, the company that can squeeze the most value out of every marketing dollar wins. By leveraging machine learning to identify high-value users, Fanatics has effectively created a "smarter" marketing machine than their competitors, who may still be relying on outdated demographic targeting.
Conclusion: The Machine as a Strategic Partner
The case of Fanatics Sportsbook and Cognitiv serves as a cautionary tale for those clinging to traditional marketing tactics. The message is clear: the era of the human-defined persona is waning. In its place, we are seeing the rise of the autonomous, outcome-oriented model.
For CMOs and performance marketers, the challenge ahead is to relinquish control of the "who" and focus intently on the "what." By defining clear, measurable business outcomes and allowing the algorithms to map the paths to those results, brands can move beyond the limits of human intuition.
As we look toward the remainder of 2026 and beyond, one thing is certain: those who learn to speak the language of the machine—and define their success by the value of the customer rather than the cost of the acquisition—will be the ones leading the pack. The future of advertising isn’t about knowing who your customer is; it’s about letting the data tell you who they want to be.
Iffa Jayyana
Rifan Muazin
