For the first decade of legal US sports betting, the CRM problem was simple: acquire bettors, retain them, reactivate them when they lapse. The customer base was homogeneous enough that a single acquisition funnel, a single retention model, and a single engagement vocabulary could serve the whole database. That era ended in December 2025.
When DraftKings and FanDuel both launched standalone prediction markets apps within weeks of each other, they created something the US market had never seen: hybrid operators. One company. Two fundamentally distinct product lines. Two different regulatory regimes. Two different user demographics responding to entirely different value propositions. And, critically, one CRM team expected to manage all of it.
This article examines why the sportsbook CRM playbook structurally fails prediction markets users, what the economic incentives for cross-product migration actually look like, and what a hybrid CRM architecture must do differently to capture the margin opportunity without destroying the demographic advantage prediction markets have built.
The Structural ShiftDecember 2025: When Sportsbooks Became Hybrid Operators
The transition happened faster than most operators anticipated. DraftKings launched its Predictions product in 38 states in December 2025 — offering sports and finance markets in states where its sportsbook doesn’t operate, and finance markets only in states where it does. The geographic logic reflects the fundamental regulatory bifurcation: DraftKings Predictions is a CFTC-regulated product, and its availability is not constrained by state gaming licenses. It can go places the sportsbook cannot.
FanDuel took a different structural path. Its Predicts product launched in 5 states — Alabama, Alaska, South Carolina, North Dakota, and South Dakota — through a partnership with CME Group. Rather than building or acquiring exchange infrastructure, FanDuel relies on CME for the regulated exchange layer while retaining customer acquisition and user experience. This partnership architecture creates CRM integration dependencies that vertically integrated operators do not face: data-sharing constraints between FanDuel and CME Group may limit how deeply prediction markets CRM can be personalized.
The most aggressive structural bet came from an operator most people weren’t watching. Underdog Sports exited traditional sports betting entirely in late 2025 and acquired the Aristotle CFTC-registered exchange (DCM + DCO) on March 9, 2026 — becoming the only operator to run both native exchange infrastructure and a customer-facing prediction markets product without relying on CME Group. Underdog has made a calculated bet that being a pure-play prediction markets operator is more defensible than trying to straddle both worlds.
Then, on March 2, 2026, DraftKings launched what it calls a unified super app — merging sportsbook, casino, lottery, and predictions under a single wallet. DraftKings’ 2026 revenue guidance stands at $6.5–$6.9 billion, with Adjusted EBITDA of $700–$900 million. Predictions is projected to contribute hundreds of millions annually in standalone revenue within a few years. The financial stakes of getting the CRM architecture right are substantial.
Why the Sportsbook CRM Playbook Fails Prediction Markets Users
DraftKings’ stated strategy is to apply its sportsbook CRM and acquisition playbook directly to prediction markets. On the surface, this seems reasonable — the company already has the customer acquisition infrastructure, the brand equity, and the CRM tooling. Why build a separate system?
The answer lies in the demographic data. Prediction markets are not attracting the same users as sportsbooks. They are attracting users that sportsbooks structurally could never acquire.
The Kalshi demographic shift is not a rounding error. It represents a fundamentally different user psychology. Prediction markets attract knowledge-economy participants: users who are drawn to analytical formats, who frame risk as a function of information quality rather than entertainment, and who respond to messaging about skill, accuracy, and financial returns. The vocabulary of sportsbook CRM — odds, handicaps, parlays, “the action” — lands differently with this audience, or doesn’t land at all.
CEO Jason Robins confirmed the demographic divergence in practical terms: prediction markets are currently capturing “low-margin customers” from the sportsbook database. The users migrating across product lines are not high-value sportsbook bettors; they are users at the bottom of the sportsbook value pyramid. This has a direct and underappreciated implication: the current sportsbook-to-predictions migration is happening at exactly the wrong place in the customer value stack. A CRM that treats cross-product migration as a single undifferentiated goal will accelerate the wrong migrations and miss the economically important ones entirely.
The right segmentation framework distinguishes between four user types that require completely different CRM treatments:
| User Segment | Profile | CRM Priority |
|---|---|---|
| Sportsbook-only | High-margin, odds-literate, bonus-responsive | Retain; surface PM only for major sports events |
| Predictions-only | Analytically motivated, new demographic | Retain on PM terms; do not push sportsbook framing |
| High-value crossover candidates | Engaged sportsbook users with finance or analytics interest signals | Highest CRM priority — migrate upmarket |
| Low-margin crossover migrants | Low-frequency sportsbook users already drifting toward PM | Retain on PM; accept the migration; optimize for PM LTV |
The Economic Case for Cross-Product Migration CRM
The margin differential between prediction markets and sportsbook is not marginal. Prediction markets carry 10–30% higher adjusted gross margins than sportsbook operations, driven entirely by the absence of state gaming taxes. Because DraftKings Predictions is CFTC-regulated, it does not fall under state gambling tax regimes. Every dollar of gross revenue from a prediction markets user is structurally more profitable than the equivalent sportsbook dollar.
CEO Jason Robins has put the full industry opportunity at $10 billion in annual gross revenue across the prediction markets category. DraftKings’ internal target for its own Predictions product is hundreds of millions annually — a standalone business line, not a rounding error on the sportsbook P&L. Industry analysis estimates 15–25% cross-product conversion rates for engaged users in a unified super app architecture, which defines the ROI boundary for hybrid CRM investment: if you can convert even 15% of your engaged sportsbook base to also trade predictions, the margin uplift on those users is substantial.
But the economic incentive for migration creates a CRM design trap. Traditional sportsbook CRM funnels are optimized for bonus responsiveness, odds engagement, and parlay conversion — none of which translate to a CFTC-regulated financial product. The compliance language is different. The offer mechanics are different. The mental model the user needs to adopt is different. Applying a sportsbook bonus funnel to a prediction markets acquisition attempt doesn’t just underperform; it actively damages the prediction markets brand by framing a knowledge-based product in gambling-adjacent language that alienates the analytically oriented demographic it’s trying to attract.
The Super App ParadoxOne Wallet, Two Regulatory Regimes — The CRM Cannot Be Unified
DraftKings’ one-wallet super app architecture, launched March 2, 2026, is the most ambitious product bet in US sports betting history. Merging sportsbook, casino, lottery, and predictions under a single wallet directly echoes iGaming super app designs that have dominated European and Asian markets for years. In those markets, the model works because a single regulatory regime governs all products. In the US, DraftKings is attempting something those operators never had to solve: a super app that spans two fundamentally different regulatory worlds simultaneously.
The CFTC and state gaming commissions operate under entirely different legal frameworks. Bonus mechanics permissible under one regime may be prohibited under the other. responsible gambling rules, self-exclusion requirements, and marketing restrictions differ materially between CFTC-regulated prediction markets and state-licensed sportsbooks. A CRM trigger that fires a bonus offer to a user engaging with sportsbook content cannot be routed to prediction markets without legal review of whether that offer structure is even compliant under CFTC rules.
The practical implication: the “one-wallet” experience is a UX decision. The CRM engine underneath must be two separate systems. They can share the user identity layer and the event-trigger layer, but every offer, every communication, and every compliance touchpoint must be routed through the correct regulatory logic before it reaches the user.
FanDuel’s CME partnership model adds a third dimension to this problem. The data-sharing boundary between FanDuel and CME Group — a futures exchange with its own regulatory obligations and data governance requirements — means FanDuel’s CRM may not have the same depth of user behavioral data for Predicts that it has for sportsbook. Personalization depth on the prediction markets side may be structurally limited by what data CME Group can share back across the partnership boundary. Vertically integrated operators like DraftKings avoid this constraint; FanDuel must design around it.
Event-Driven TriggersLive Sports: The One CRM Hook That Works Across Both Products
Amid all this structural complexity, there is one powerful point of CRM convergence: live sports events are the primary acquisition and engagement trigger for both product types, and they behave identically across the divide.
This convergence is not a coincidence. Sports dominate prediction market trading even on platforms with no sportsbook heritage: more than 80% of Kalshi’s trading volume is sports-related, and US-based Polymarket activity is effectively 100% sports. The prediction markets category has not displaced sports betting psychology — it has attracted users who want to apply a different framework (skill, knowledge, analysis) to the same subject matter (sports outcomes).
For hybrid operators, this is the single most actionable CRM insight: the event-driven trigger infrastructure you already built for sportsbook — real-time alerts keyed to game start, halftime, injury reports, weather changes, and final scores — is directly reusable for prediction markets. The difference is not the trigger. The difference is what the trigger sends.
A sportsbook CRM sends: “Chiefs vs. Eagles kicks off in 2 hours — bet the spread.”
A prediction markets CRM sends: “Chiefs vs. Eagles kicks off in 2 hours — predict the winner and earn on your analysis.”
Same event. Same trigger. Different vocabulary, different framing, different regulatory wrapper, different product surface. This is the cross-product CRM opportunity in its simplest form: a shared event bus that routes the same sports moment to both product surfaces simultaneously, with product-appropriate content generated for each. DraftKings recorded $54 billion in sportsbook handle in 2025, an 11% year-over-year increase despite only half the US population having legal access to sports betting. That event engagement infrastructure is the foundation on which hybrid CRM can be built — not replaced.
The New PlaybookWhat a Hybrid CRM Architecture Actually Requires
Hybrid operators that try to run a single undifferentiated CRM across sportsbook and prediction markets will achieve one of two failure modes: they will apply sportsbook language to prediction markets users and fail to retain the analytically motivated demographic that makes the product economically interesting, or they will apply prediction markets framing to sportsbook users and dilute the engagement mechanics that drive high-margin sportsbook revenue. Neither outcome justifies the investment in the category.
The architecture that avoids both failure modes has four distinct components:
1. Two Acquisition Funnels
Sportsbook onboarding must emphasize odds literacy, the excitement of live action, and the thrill of having skin in the game. Prediction markets onboarding must emphasize knowledge as a competitive advantage, skill-based returns, and the analytical format. These are not the same user motivation, and they cannot be served by the same onboarding sequence. Operators that share a single acquisition funnel across both products will see higher drop-off rates on whichever product the funnel is not designed for.
2. Two Retention Models
Sportsbook retention leans on bonuses, parlay mechanics, live-bet nudges, and the cadence of the sports calendar. Prediction markets retention leans on market depth, payout transparency, and the sense that the platform rewards analytical rigor. bonus abuse and bonus-driven churn are endemic in sportsbook CRM; the prediction markets user profile is less bonus-sensitive and more sensitive to market quality and resolution integrity. A retention model built around deposit bonuses and free bet offers will not retain the knowledge-economy user that prediction markets are built to attract.
3. Four-Segment User Classification
The segmentation framework outlined earlier — sportsbook-only, predictions-only, high-value crossover candidates, and low-margin crossover migrants — is the operational foundation for hybrid CRM. Each segment requires a different communication strategy, a different offer structure, and a different success metric. Treating all four as a single addressable market produces averaged-down results for all of them. The high-value crossover candidate — an engaged sportsbook user with analytical interests — is the most economically important segment and the least likely to be identified by a CRM system not designed to look for them.
4. A Shared Event-Trigger Layer
The one point of legitimate unification is the real-time event bus. Live sports events, injury reports, weather delays, and final scores are relevant to both product types and should be routed to both surfaces simultaneously. The event-trigger layer is where the efficiency gains of a unified super app are actually realizable. Everything downstream of the trigger — the content, the offer, the compliance wrapper, the product surface — must be handled by product-specific systems. But the trigger itself is shared infrastructure, and it is where hybrid CRM delivers its highest leverage.
What This Means for CRM Platforms and Content Generation
The hybrid CRM challenge is not primarily a platform problem — it is a content and segmentation problem. Most enterprise CRM platforms (Braze, Optimove, Salesforce Marketing Cloud) are technically capable of running bifurcated communication tracks with separate compliance wrappers. The bottleneck is not the sending infrastructure. The bottleneck is generating the right content, at the right time, with the right vocabulary, for the right segment, at a scale that makes individual personalization economically viable.
A sportsbook CRM team managing 5 million users across two product types cannot manually produce 50 distinct content variants per user segment per sports event. The math doesn’t work without AI content generation at the personalization layer. The event-trigger infrastructure fires in real time; the content it delivers must be generated in real time, calibrated to whether the recipient is a sportsbook user being shown prediction markets for the first time, a high-value crossover candidate being nudged toward a specific prediction market, or a predictions-native user being retained through market depth and analytical framing.
This is precisely where the hybrid CRM opportunity becomes a content infrastructure problem. The four-segment framework requires four distinct content vocabularies. The two-product architecture requires two distinct compliance wrappers. The event-trigger layer requires real-time content generation that can respond to a breaking injury report or a live game moment with product-appropriate messaging for each segment simultaneously. Operators that build this capability will measure it in cross-product LTV and in the 15–25% cross-product conversion rates that industry analysis suggests are achievable in a well-architected super app.
SourcesData Sources & Attribution
- Next Event Horizon: DraftKings Q4 2025 Analysis — DraftKings $54B handle, Predictions state availability, Super Bowl trading volume, revenue guidance
- DeFi Rate: DraftKings Super App Analysis — 10–30% margin advantage, 15–25% cross-product conversion estimate
- Fast Company: Kalshi & Polymarket Feature — Kalshi female user share (13% to 26%), sports as 80%+ of Kalshi trading volume
- CME Group Press Release, December 2025 — FanDuel Predicts launch in 5 states via CME partnership
- DraftKings Q4 2025 earnings materials — CEO Jason Robins on “low-margin customers”, $10B industry opportunity, hundreds of millions standalone Predictions revenue target