Prediction Market Trading Fees & Revenue Models for Operators

A prediction market charges a fee of a few percent or less on traded volume, the way an exchange does, not a house margin the way a bookmaker does. The revenue base is a maker-taker fee, a spread, or a settlement fee applied as the market resolves, and it is genuinely a net-revenue base rather than a house edge baked into every line. A fee of a few percent or less on traded volume is the norm rather than a double-digit margin, fee design directly governs liquidity, and affiliate revshare is paid out of net fee revenue rather than out of GGR or NGR. This guide is about operator revenue and how the partner economics sit on top of it; it is not an affiliate-commission-structure article.

The revenue base is trading fees on volume, not a house edge, and affiliate revshare is computed against that net fee revenue.

The fee models at a glance

A prediction-market venue generates revenue through three recurring mechanisms, and most platforms blend them. Each charges against volume or settlement, not against a built-in margin, which is the property that distinguishes an exchange revenue model from a sportsbook one.

Public interest in this is high enough that "kalshi vs polymarket fees" is a notable commercial-intent query, which tells you operators and traders alike care about the precise number. We will not invent precise rates here - they change and differ by venue - but the structural point holds: these are fees on activity, and the documentation for venues like Kalshi and Polymarket is the authoritative source for current rates. On a CFTC-regulated venue these fees apply to an event contract whose price reads as an implied probability, and the fee schedule itself sits inside the venue's compliance and disclosure obligations.

Designing a fee model for a new venue follows five ordered steps:

1. Set the maker-taker schedule: a low or rebated maker fee to reward resting orders and a higher taker fee on liquidity-removing trades.
2. Decide whether to charge a settlement fee on the resolving side, which defers cost to market resolution.
3. Choose how much of the cost sits in the spread versus an explicit posted fee, since a tighter book lowers the implicit charge.
4. Tier the fees by volume so high-frequency participants who deepen the order book pay less at the margin.
5. Define the net fee revenue base that affiliate revshare and CPA terms will be computed against, net of any agreed deductions.

Maker-taker fees: the workhorse model

Maker-taker is a fee model that charges a lower fee, or pays a rebate, to the maker who posts a resting order and adds liquidity, and a higher fee to the taker who removes liquidity by hitting an existing order. The structure is well established across financial exchanges, and prediction markets borrow it directly because it solves their central problem of thin order books.

The maker-taker structure predates prediction markets by decades in traditional derivatives venues, and regulators have addressed it directly in public statements on exchange fee models.

The design intent is to subsidize the behaviour the venue needs most. By making it cheap to post liquidity into the central limit order book, the operator deepens the book, tightens spreads, and improves execution for everyone. The fee tier a trader pays can also scale with volume, rewarding the high-frequency participants who keep markets liquid.

Spread and settlement fees

Two revenue mechanics operate beyond explicit maker-taker fees: the spread and the settlement fee. The spread, the gap between the best bid and best ask, is an implicit cost the trader pays and a function of liquidity rather than a posted fee, so tighter books mean smaller spreads and a better deal for users. A settlement fee is charged when a market resolves, often on the winning side, and is the clearest analogue to a betting-exchange commission on net winnings.

A settlement fee defers the charge to market resolution rather than to each trade.

The mix an operator chooses shapes behaviour. Settlement-weighted fees defer the cost to resolution and can feel cheaper to active traders; execution-weighted maker-taker fees price every trade. Most venues blend the two, and the blend is itself a product decision that affects both volume and the net-revenue base the affiliate program will earn against.

Why this is a net-revenue base, not a house edge

A prediction-market fee is a charge on trading activity, not a house edge. A sportsbook builds margin into its odds, the vigorish, and profits from the difference between true probability and offered odds, taking outcome risk in the process. A prediction-market exchange does not take the other side; it matches traders and charges a fee on their activity, the way a stock exchange does.

The vigorish has no equivalent at a fee-based exchange, where revenue comes from activity rather than outcome risk.

That distinction drives the whole revenue conversation. There is no gross win to skim; there is net fee revenue earned on volume. It is closer to a brokerage or exchange P&L than to GGR or NGR. For the side-by-side with the betting-exchange commission model, see our prediction markets vs betting exchanges comparison.

How fee design drives liquidity

Fee design determines the operator's primary liquidity lever, and getting it wrong is self-defeating. Set fees too high and you suppress the very volume the fee is charged on, thin the order book, and widen spreads until traders leave. Set maker incentives well and you deepen liquidity, which improves execution, which attracts more volume, the flywheel an exchange lives on.

Deeper liquidity tightens spreads and pulls in the volume the fee is charged on.

Because the fee is the revenue base, there is a direct tension between per-trade take and total volume. The operators who get this right treat fee tiers as a growth instrument - lower or rebate maker fees to seed open interest, and let volume rather than rate carry the revenue. That same logic should inform how generous the affiliate program can be, because partner payouts come out of the same net fee pool.

How affiliate revshare maps onto net fee revenue

Affiliate revshare relies on net trading-fee revenue as its base for a prediction market, which is the structural difference from a sportsbook program that pays on NGR. A revenue share on the fees generated by a partner's referred traders maps cleanly onto the exchange P&L: the partner earns a slice of the fees their users pay, net of any agreed deductions.

A prediction-market commission therefore tracks the fee pool a partner's traders generate, not a gross-win figure.

The operational requirement is a commission engine that can take net fee revenue per partner as its input, apply tiered revshare or CPA terms, and reconcile that against actual settlement. That is exactly what Track360's commission management and finance and payouts layers do - they let an operator define the net-fee denominator once and run revshare, CPA or hybrid terms against it, with deep-linked tracking feeding attribution. For how the on-chain version of this is tracked, see our crypto prediction markets guide.

The takeaway for operators

Prediction-market monetization functions as exchange monetization: maker-taker fees, spread and settlement fees on volume, charged as a net-revenue base rather than a house edge. Fee design is inseparable from liquidity, so price young books to seed depth rather than to maximize per-trade take, and remember that the affiliate program is paid out of the same net fee pool. Build the commission engine to compute revshare on net fee revenue, not on GGR or NGR, and the partner economics will line up with how the venue actually earns.