Kount vs Sift vs Forter: $160K Fraud-Tool Gap [2026]

A Steam refund request, a disputed PlayStation Store charge, a Fortnite V-Bucks purchase a parent “didn’t recognize” on the credit card statement: every one of those ends up on a fraud analyst’s desk, and every one of them costs a publisher real money whether it was theft or not. Digital goods and subscription chargeback rates climbed 59% in a single year, from 0.34% in 2023 to 0.54% in 2024, according to aggregated industry data compiled by Chargeflow. Game publishers sit squarely in that exposed category: no physical delivery, no signature, no proof a package ever left a warehouse. Just a login, a purchase, and a bank that has to decide whether to believe the merchant or the cardholder.

That’s the gap four companies have built businesses around: Kount, Sift, Forter, and Riskified. Each sells a system that scores a transaction in real time and tells a game storefront whether to approve it, hold it for review, or decline it outright. None of them were built exclusively for gaming, but all four count digital-goods and marketplace merchants among their core customer base, and gaming’s mix of account takeover, friendly fraud, and global card-not-present sales makes it one of the harder verticals to get right. This comparison breaks down what each platform actually does, what it costs, what the benchmark data says, and which one fits a given publisher’s risk profile in October 2026.

The stakes for getting this wrong have gone up in 2026, not down. Card networks tightened their monitoring thresholds this year, chargeback volume is still climbing globally, and the share of disputes tied to friendly fraud rather than outright theft keeps growing, which means a publisher’s fraud stack now has to do two jobs at once: catch the bot-driven card-testing and account-takeover attempts before they clear, and assemble the evidence to win disputes over purchases that were actually legitimate. Those are different engineering problems, and the four vendors in this piece solve them with meaningfully different architectures, price tags, and risk-transfer models.

·

Google · Preferred Sources

Don't miss new tech stories on Google

Add Tech Insider once in the Google app and our stories appear in your news suggestions.

Add Now

Why Game Publishers Need Dedicated Fraud Detection Software in 2026

Video game storefronts share a structural weakness with every other digital-goods seller: the thing being sold has no shipping address. A stolen credit card buying a $400 gaming laptop at least has to go somewhere, which gives a fraud team an address to flag. A stolen card buying 10,000 V-Bucks, a season pass, or in-game currency clears instantly and the “goods” are consumed within minutes, long before a bank even notices the card is compromised. That asymmetry is why card-not-present transactions post chargeback rates of 0.6% to 1.0%, roughly double the 0.5% rate for card-present purchases, per Chargebacks911’s 2026 statistics roundup.

Account takeover compounds the problem. Credential-stuffing attacks against gaming accounts don’t stop at login. Once an attacker is inside a Steam, Xbox, PlayStation, or Epic account with a saved card on file, the endgame is almost always a fast burst of microtransactions before the real owner notices. Sift’s own fraud index pegs the broader e-commerce account-takeover attack rate at 2.5% in Q2 2025, with projected ATO losses climbing to $17 billion industry-wide, up from $13 billion in 2023, according to figures cited in a 2026 ecommerce fraud playbook from Digital Applied. That same report found 83% of surveyed organizations had experienced at least one ATO incident. None of those figures are gaming-exclusive, but gaming’s combination of saved payment methods, resellable virtual goods, and young, often first-time cardholders makes it an attractive target inside that broader number.

Then there’s friendly fraud, the slipperier problem: a legitimate purchase that the cardholder later disputes anyway. A teenager buys a battle pass on a parent’s saved card, the parent doesn’t recognize the line item on the statement, and files a dispute instead of asking their kid. Visa itself defines friendly fraud as exactly this: a dispute over a transaction the cardholder or someone in their household actually made. Merchants in Chargebacks911’s 2026 Chargeback Field Report estimate friendly fraud accounts for 43.8% of their chargeback losses on average, while Visa’s own figures put the industry-wide share at 75% and Chargebacks911’s internal data at 86%. Juniper Research forecasts friendly fraud will represent 22% of global chargebacks in 2026, climbing to 28% by 2031. None of that is a problem antivirus software or a firewall solves. It’s a pattern-recognition and evidence-assembly problem, and that’s exactly what Kount, Sift, Forter, and Riskified are built to handle.

Friendly Fraud vs. Theft: Two Problems, One Fraud Stack

It’s worth separating the two failure modes a game publisher’s fraud stack has to catch, because the fixes are different. True fraud is a stolen card, a compromised account, or a bot farm running card-testing scripts against a checkout endpoint to find which stolen numbers still work. Friendly fraud is a legitimate purchase disputed after the fact, whether out of genuine confusion, buyer’s remorse, or a deliberate attempt to get a free game. Industry breakdowns from Chargeflow group chargebacks into three buckets: true fraud, friendly fraud, and merchant error, and digital-goods sellers get hit disproportionately by the second category because there’s no delivery record to settle the argument.

Card networks have started building tools specifically for that evidence gap. Visa’s Compelling Evidence 3.0 framework, live for Visa-branded cards, lets a merchant defeat a reason-code 10.4 card-not-present fraud dispute by showing at least two prior undisputed transactions on the same card, made 120 to 365 days before the disputed purchase, that share at least two matching identifiers such as IP address, device ID, billing details, or account history. For a game publisher, that means transaction history, login device fingerprints, and account activity logs become the “proof of delivery” a physical retailer gets for free from a shipping carrier. This is precisely the data layer Kount, Sift, Forter, and Riskified are built to collect, structure, and hand back to merchants during a dispute. None of the four vendors publish gaming-specific chargeback statistics, and this piece won’t invent any, but the mechanics of CE 3.0 line up almost exactly with what a game account’s purchase history already contains.

Kount: Equifax’s Payment-Screening Engine

Kount built its name as an independent fraud-prevention vendor before Equifax acquired it, reportedly for approximately $640 million in February 2021. It is now marketed in some current material as Kount 360 by Equifax, and a recent software guide lists Kount Command for fraud decisioning, Kount Control for account protection and Kount Central for chargeback management.

Kount’s pricing is custom-quote only, with no published rate card. The company says cost depends on a merchant’s transaction volume and specific risk goals, according to a 2026 vendor comparison from Chargeback.io. That opacity is typical across this category, but it means a smaller or mid-market mobile game studio has to go through a sales cycle just to get a ballpark figure, whereas Sift and Forter at least have third-party estimated ranges circulating in the market (more on those below). For a publisher already running on Equifax identity-verification rails elsewhere in the business, Kount’s main pitch is consolidation: one vendor, one data relationship, one combined view of payment and identity risk.

Sift: Behavioral Signals Across the Full User Journey

Sift positions itself less as a single chargeback-prevention tool and more as a fraud layer that spans the entire account lifecycle. As the company puts it on its own platform page, Sift is built to “detect and prevent fraud across key moments in the user journey, from account creation and login to payment and post-transaction review, while reducing friction for trusted users,” a framing that matters for gaming specifically because account creation and login fraud (fake accounts, credential stuffing, bot sign-ups for promo abuse) is often a bigger operational headache for a free-to-play title than payment fraud alone. That lifecycle framing now extends to AI tooling: Sift’s platform includes AI tools for fraud detection and decisioning, the Fraud Attack Defense Suite, whose first component, Attack Detection Agent, is aimed at detecting coordinated fraud attacks earlier, and the Sift MCP Server.

On pricing, Sift runs a subscription-plus-event-volume model rather than Forter’s or Riskified’s percentage-of-GMV structure, and it doesn’t bundle a chargeback guarantee the way its two competitors do. Third-party deal-data estimates compiled by AI Toolbox’s 2026 comparison put typical mid-market annual Sift contracts in the $40,000 to $100,000 range, making it the cheapest of the three quote-based platforms at the entry tier, though without the built-in financial backstop a guarantee provides. Sift frames the tradeoff directly on its platform page: “Effective payment fraud prevention helps lower chargebacks, reduce manual review, and improve approval rates for legitimate transactions”, a lower cost of ownership in exchange for the publisher carrying more of the chargeback risk itself.

Forter: Decision-as-a-Service With a Bundled Guarantee

Forter’s pitch is the inverse of Sift’s: pay more, but transfer more of the financial risk onto the vendor. Forter’s own pricing page claims the company is “the only company that guarantees both chargeback and approval rates,” positioning its guarantee not just as insurance against fraud losses but as a commitment that legitimate customers won’t get wrongly declined either, which matters to a publisher trying to avoid losing real paying players to an overly conservative fraud filter.

That guarantee comes bundled into the price rather than sold as an add-on, and Forter charges as a percentage of gross merchandise volume on a quote-only basis. The same AI Toolbox dataset referenced above estimates mid-market Forter contracts in the $60,000 to $180,000 annual range, a meaningfully higher floor than Sift, reflecting the cost of that bundled guarantee. For a storefront selling digital currency or DLC at high volume with thin per-transaction margins, that percentage-of-GMV structure means the fee scales directly with revenue, which is either a feature or a drawback depending on how a studio’s finance team wants to budget for it.

Riskified: The Public Company Alternative

Riskified is the only one of the four that trades on a public exchange, listed as NYSE: RSKD, which gives outside observers a transparency the other three vendors don’t offer. As of October 7–8, 2026, Riskified shares hit a 52-week high near $8.17–$8.22, putting its market capitalization in the $1.06 billion to $1.19 billion range depending on the exact share count used, according to MarketBeat’s October 7 report. The same week, Riskified raised its full-year 2026 revenue forecast to $400–410 million and lifted its adjusted EBITDA guidance to $33–39 million.

Riskified’s guarantee works similarly to Forter’s but with an important carve-out: it covers unauthorized-card and stolen-payment-information chargebacks, but explicitly excludes disputes like “item not received” or “item not as described,” categories that matter less for a pure digital download but more for marketplaces selling physical collector’s editions or merchandise bundles alongside digital codes. Riskified charges a percentage of approved GMV plus an annual minimum floor, and the AI Toolbox estimate puts mid-market contracts at $70,000 to $200,000 a year, the highest of the three quote-based vendors in that dataset, with pricing typically running 15–30% above Sift because the guarantee is bundled rather than optional. As Riskified states on its own fraud prevention page, “Our performance guarantee, including full chargeback protection, allows you to sell to anyone, anywhere,” and elsewhere the company frames detection accuracy as a direct lever on customer experience on its chargeback guarantee page: “Accurate fraud detection reduces false positives and allows orders to ship sooner.”

Full Specs Comparison: Kount vs. Sift vs. Forter vs. Riskified

The table below lines up what’s publicly confirmable about each vendor as of October 2026. Employee and revenue figures for Kount, Sift, and Forter come from third-party aggregator estimates rather than official company disclosures, since none of the three private companies publish headcount or revenue, so those rows are marked accordingly. Riskified’s figures are drawn from its public market data since it’s the only exchange-listed vendor of the four.

Attribute	Kount	Sift	Forter	Riskified
Ownership	Equifax subsidiary (since Feb. 2021)	Private, venture-backed	Private, venture-backed	Public, NYSE: RSKD
Core data approach	Device ID + identity/credit network (via Equifax)	Behavioral signals across full user journey	Decision-as-a-service with bundled guarantee	Approve/decline decisioning + GMV guarantee
Chargeback guarantee	Not standard, case-by-case	Not bundled by default	Yes, bundled into price	Yes, bundled (excludes “not received”/”not as described”)
Pricing model	Custom quote, volume-based	Subscription + event volume	Quote-only, % of GMV	% of approved GMV + annual minimum
Est. mid-market annual cost*	Not publicly estimated	$40,000–$100,000	$60,000–$180,000	$70,000–$200,000
Est. employee count*	~292	~408	~670	~801
Public financials	Rolled into Equifax’s reporting	Private	Private	Public 10-Q/earnings guidance
Account-lifecycle coverage	Payment-focused	Signup, login, payment, post-transaction	Primarily payment/checkout decisioning	Primarily payment/checkout decisioning
Best fit by company size	Enterprise already on Equifax identity rails	Mid-market to enterprise wanting lifecycle coverage	Enterprise wanting a bundled financial backstop	Enterprise wanting public-company transparency
Headquarters market	U.S. (Boise, reported)	U.S.	U.S./Israel dual presence	U.S./Israel dual presence
Known 2025–2026 development	Pricing/ops folded under Equifax	Launched Attack Detection Agent, the first component of its Fraud Attack Defense Suite (October 2026)	Guarantee-based pricing page updated 2026	Raised FY2026 revenue guidance Oct. 8, 2026

*Employee counts and Sift/Forter/Riskified mid-market pricing estimates are drawn from third-party aggregator data (ReviewBolt, Vendr-sourced estimates via AI Toolbox), not official vendor disclosures, and should be treated as directional rather than exact.

Pricing Breakdown: What a Game Publisher Actually Pays

None of the four vendors in this comparison publish a self-serve rate card, which is standard for enterprise fraud-prevention software but frustrating for a smaller studio trying to budget before a sales call. The clearest pattern in the available data is that pricing scales with how much financial risk the vendor is willing to absorb. Sift, which doesn’t bundle a standard chargeback guarantee, comes in cheapest. Forter and Riskified, both of which bundle guarantees into the price, run meaningfully higher, and Riskified tops the range in the available estimates, consistent with its guarantee-plus-minimum-floor pricing structure.

Vendor	Pricing structure	Est. mid-market annual range	Guarantee included	Best for
Kount	Custom quote (volume-based)	Not publicly benchmarked	Case-by-case	Enterprises already using Equifax identity services
Sift	Subscription + event volume	$40,000–$100,000/yr	No (optional add-on varies)	Cost-conscious mid-market, full lifecycle coverage
Forter	Quote-only, % of GMV	$60,000–$180,000/yr	Yes, bundled	High-volume storefronts wanting approval-rate guarantees
Riskified	% of approved GMV + annual minimum	$70,000–$200,000/yr	Yes, bundled (with exclusions)	Enterprises wanting the most financial risk transferred

That $160,000 spread between Sift’s entry-level estimate and Riskified’s top-end mid-market estimate is the real decision point for most studios: it’s effectively the price of a bundled chargeback guarantee plus a higher baseline of white-glove account management. A publisher processing low millions in annual digital-goods revenue, with a trust and safety team that can handle manual review escalations, can reasonably run Sift leaner. A publisher processing hundreds of millions through microtransactions, where even a small percentage of unresolved fraud losses is a material line item, is the more natural Forter or Riskified customer. One industry fraud-stack breakdown circulating in 2026, from AI-Pedias’ fintech fraud detection roundup, estimates a combined gaming/iGaming fraud stack pairing Sift with identity and device-trust tools like Sumsub and iovation at roughly $300,000 a year for multi-account and bonus-abuse detection, while a marketplace-style stack combining Sift, Forter, and Persona for trust-and-safety coverage runs closer to $500,000 a year. Those figures come from one third-party analysis rather than vendor disclosures, but they illustrate how fast costs stack once a publisher layers multiple point solutions instead of consolidating on one platform.

Benchmarks: What the Chargeback Data Actually Shows

None of the four vendors publish independently audited, head-to-head accuracy benchmarks against each other, and any article claiming an exact “Forter catches X% more fraud than Sift” figure is almost certainly fabricating a number no public source supports. What does exist, across multiple independent industry reports, is a consistent macro picture of where chargebacks are trending, and it explains why demand for all four platforms has grown.

Mastercard and Datos Insights’ 2025 chargeback research projects global dispute volume will hit 261 million in 2025 and climb to 324 million by 2028, with roughly 74% of disputes escalating to a full chargeback rather than getting resolved earlier in the process, according to figures compiled by Chargeback.io’s 2026 statistics roundup. LexisNexis Risk Solutions data cited in the same roundup puts the all-in cost of fraud at $4.61 for every $1 actually lost to merchants once investigation labor, lost goods, and bank fees are factored in. That multiplier is the real business case for any of these four platforms: even an imperfect automated filter that catches a meaningful share of fraudulent transactions before they ship pays for itself by avoiding that $4.61 overhead on each one.

On the regulatory side, Visa tightened its Acquirer Monitoring Program in 2026: starting April 1, 2026, the threshold for what Visa classifies as an “excessive” fraud-and-dispute ratio dropped from 2.2% to 1.5% of settled transactions across the U.S., Canada, and the EU, based on figures reported by Shopify’s 2026 chargeback management guide. That same reporting notes Visa’s monitoring calculation counts both chargebacks and fraud reports together, and a single transaction can be counted twice: once as a fraud alert and once as a dispute. A lower threshold means publishers have less room for error before landing in a monitoring program that brings higher processing fees and closer scrutiny from acquiring banks, which raises the stakes on getting fraud detection right in the first place rather than cleaning up after the fact.

Account Takeover and Card Testing: The Gaming-Specific Attack Pattern

Gaming accounts are an unusually attractive target for credential-stuffing bots for a reason that has nothing to do with the games themselves: saved payment methods plus resellable digital inventory equals liquid value an attacker can cash out fast. A compromised account with a saved card attached doesn’t need to be drained through a bank transfer. It can be used to buy premium currency, skins, or season passes that get resold on secondary markets within hours, often before the account owner even notices the login alert.

Card testing follows the same mechanical pattern seen across retail and travel: bots run small, automated transactions against a checkout endpoint using lists of stolen card numbers, filtering for which numbers still clear before using the “good” ones for a larger charge elsewhere. Game storefronts with low-dollar digital goods (a $0.99 in-app purchase, a cheap cosmetic item) make convenient, low-friction testing grounds precisely because the charges are small enough to clear without triggering a card issuer’s fraud alerts. This is the layer where Sift’s full-lifecycle approach, scoring signup and login behavior rather than just the payment step, is structurally different from a pure checkout-decisioning tool: a card-testing bot often reveals itself through account creation velocity and device fingerprint patterns well before it ever reaches a payment field.

Mobile, Console, and PC Storefronts: Where Fraud Risk Concentrates Differently

Fraud exposure isn’t evenly distributed across platforms, and a publisher shipping the same title on mobile, console, and PC is effectively running three different risk profiles under one brand. Mobile storefronts see the highest volume of small-dollar transactions and the heaviest card-testing traffic, because app store checkout flows are frequently optimized for one-tap purchases with saved payment methods already on file, which is convenient for a legitimate buyer and just as convenient for a bot testing stolen card numbers. Console storefronts (PlayStation Store, Xbox, Nintendo eShop) sit in the middle: transaction values run higher on average, since a console purchase is more likely to be a full game or a season pass than a $0.99 consumable, but the closed hardware ecosystem and first-party account systems add friction that pure card-testing bots find harder to automate at scale.

PC storefronts and direct-to-consumer web checkouts carry a different risk entirely: they’re the platforms most exposed to friendly fraud and refund-policy abuse, because a PC buyer purchasing directly from a publisher’s own site, rather than through a platform holder’s app store, isn’t protected by that platform’s built-in dispute-handling process. A refund dispute on a direct PC purchase lands squarely on the publisher’s own payment processor and, by extension, on whichever fraud-detection vendor the publisher has integrated at that checkout. This is part of why none of the four vendors in this comparison position themselves as a single drop-in replacement for a platform holder’s built-in protections. Kount, Sift, Forter, and Riskified are typically deployed at the specific checkout surfaces a publisher controls directly, such as its own web store, launcher, or mobile backend, rather than inside Valve’s, Sony’s, or Microsoft’s own payment infrastructure, which those platform holders secure themselves.

Visa Compelling Evidence 3.0: The Rule That Changes the Math for Digital Goods

CE 3.0 deserves its own section because it’s the single regulatory change in this space most directly relevant to game publishers, and it’s one most studios’ finance teams have likely never heard of. The framework lets a merchant preemptively defeat a Visa reason-code 10.4 (card-not-present fraud) dispute by submitting evidence of at least two prior undisputed transactions on the same card, made between 120 and 365 days before the disputed charge, sharing at least two matching data points such as IP address, device ID, billing address, or account identifiers, according to TaskUs’s 2026 breakdown of the rule.

For a physical retailer, assembling that evidence package means digging through shipping records and signature confirmations. For a game storefront, the equivalent evidence, login history, device fingerprints, prior purchase timestamps, and account tenure, already lives in the platform’s own database, assuming it’s being captured and structured correctly. This is exactly the data layer Kount, Sift, Forter, and Riskified are built to maintain continuously rather than reconstruct after the fact during a dispute. A publisher without one of these systems in place is effectively trying to win a CE 3.0 dispute by manually pulling server logs under a deadline. A publisher with structured transaction and device history already indexed has the evidence package assembled before the dispute even arrives.

5 Real-World Use Cases for Fraud Detection in Gaming

These scenarios reflect the general attack and dispute patterns documented across the fraud-prevention industry’s own public reporting, applied specifically to how a game storefront’s checkout and account systems work.

• Mobile free-to-play currency purchases. A mobile title selling premium currency in $0.99–$99.99 tiers is a classic card-testing target: small transaction sizes clear easily, and a successful test can be immediately followed by a larger purchase on the same validated card before a bank flags it.
• Season pass and battle pass disputes. A parent disputing a $9.99–$19.99 seasonal pass their child purchased is the textbook friendly-fraud case Visa’s own definition describes: a transaction the household actually authorized, disputed after the fact rather than refunded through the publisher’s own support channel.
• Account takeover via credential reuse. A player who reuses a password breached elsewhere becomes a target for automated login attempts. Once inside, an attacker with a saved payment method can burn through in-game currency purchases in minutes, a pattern Sift’s lifecycle-wide monitoring is specifically designed to catch at the login stage rather than the payment stage.
• Digital marketplace resale fraud. Platforms that allow trading or reselling in-game items (skins, cosmetics, virtual currency) face a secondary fraud surface where stolen accounts are used purely to launder value through item trades rather than direct purchases, a pattern closer to marketplace fraud than simple payment fraud.
• Chargeback disputes on bundled physical/digital releases. A collector’s edition that ships a physical box alongside a digital code creates exactly the “item not received” dispute category that Riskified’s guarantee explicitly excludes, which is why publishers running hybrid physical/digital SKUs need to read guarantee exclusions carefully rather than assume blanket coverage.

Pros and Cons: Kount vs. Sift vs. Forter vs. Riskified

No single platform wins on every axis, and the right choice depends heavily on how much financial risk a publisher wants to keep in-house versus pay a vendor to absorb.

Kount

• Pros: Backed by Equifax’s identity and credit data network, useful for publishers already consolidating vendor relationships around Equifax services.
• Cons: No public pricing benchmarks at all, even third-party estimates, and weaker public documentation of gaming-specific use cases than its three competitors.

Sift

• Pros: Full account-lifecycle coverage from signup through post-transaction review, the lowest estimated entry cost among the quote-based vendors, and published fraud-index research adds some transparency into its data.
• Cons: No bundled chargeback guarantee by default, meaning the publisher retains more direct financial exposure to fraud losses that slip through.

Forter

• Pros: Bundled guarantee covers both chargebacks and approval rates, directly protecting revenue from both fraud losses and false declines.
• Cons: Meaningfully higher estimated cost than Sift, and percentage-of-GMV pricing scales with revenue growth, which can get expensive fast for a breakout hit title.

Riskified

• Pros: Only publicly traded vendor of the four, giving outside visibility into financial health via quarterly earnings, and raised its own 2026 revenue guidance, signaling business momentum.
• Cons: Highest estimated mid-market pricing range of the four, and a guarantee that explicitly excludes “item not received” and “item not as described” disputes, which matters for hybrid physical/digital releases.

Migration Guide: Switching Fraud-Detection Vendors Without Breaking Checkout

Moving from one fraud-prevention platform to another, or adopting one for the first time, is a payments-infrastructure change, not a simple plugin swap, and it touches live checkout traffic. The sequence below reflects how enterprise merchants typically stage this kind of migration to avoid a gap in fraud coverage.

1. Export at least 12 months of historical transaction, chargeback, and dispute-outcome data from the current system or payment processor, since every vendor’s onboarding model-tuning process depends on historical labeled data.
2. Map every checkout touchpoint (web storefront, mobile app, console store API, third-party key resellers) that currently calls the existing fraud API, since missing an integration point creates an unprotected gap.
3. Run the new vendor in shadow mode, scoring live transactions in parallel without actually blocking or approving based on its output, for a minimum of 30–60 days to validate its decisions against known outcomes.
4. Compare shadow-mode false-positive and false-negative rates against the incumbent system’s historical performance before cutting over any real traffic.
5. Negotiate the chargeback-guarantee terms (if applicable) explicitly around the publisher’s specific SKU mix, flagging any hybrid physical/digital bundles that might fall into excluded dispute categories.
6. Migrate a single, lower-risk SKU or region first (e.g., one DLC pack or one secondary market) rather than cutting over the entire storefront at once.
7. Monitor the Visa Acquirer Monitoring Program ratio and Mastercard dispute-ratio program status weekly during the transition, since a spike in unresolved disputes during cutover can trigger monitoring-program flags that outlast the migration itself.
8. Fully decommission the old vendor’s API calls only after at least one full billing and dispute cycle (typically 60–90 days) has cleared on the new system with acceptable performance.

A sample webhook payload structure for a fraud-scoring decision, representative of the shape these APIs typically return (not an exact vendor spec), looks like this during the shadow-mode validation phase:

{
  "transaction_id": "txn_8f2a1c",
  "decision": "review",
  "risk_score": 72,
  "signals": {
    "device_id_known": false,
    "account_age_days": 3,
    "ip_country_mismatch": true,
    "velocity_flag": "high"
  },
  "recommended_action": "manual_review",
  "guarantee_eligible": false
}

How to Choose: Matching the Vendor to Your Storefront

A console-exclusive AAA publisher selling full-price titles with occasional DLC has a very different fraud profile than a mobile free-to-play studio selling $0.99 currency packs at massive volume, and the right vendor choice follows that difference. High-volume, low-dollar-value microtransaction storefronts benefit most from a system tuned for speed and lifecycle-wide signal detection, since the economics don’t support heavy manual review on each transaction, which favors Sift’s model. Enterprise publishers running large seasonal content drops with significant single-transaction value, where a wrongly declined legitimate purchase during a launch window is itself a costly problem, benefit more from Forter’s or Riskified’s bundled guarantees, since the cost of the guarantee is easier to justify against the revenue at stake. Publishers already deep in the Equifax ecosystem for other identity or credit services have a case for Kount purely on vendor consolidation grounds, independent of any head-to-head performance claim.

The Verdict: Which Fraud Platform Fits a Game Publisher in 2026

There’s no universal winner here, and any comparison claiming otherwise is oversimplifying a decision that depends on transaction volume, SKU mix, and how much chargeback risk a studio wants to carry itself. What the data does support: Sift is the most defensible default for a mid-market publisher that wants full-lifecycle fraud coverage (signup through post-transaction) without paying for a bundled guarantee it may not need, at an estimated $40,000–$100,000 a year, the lowest entry point among the three vendors with public pricing estimates. Forter is the stronger pick for a publisher that wants both chargeback and approval-rate guarantees baked into the contract and can absorb the higher estimated $60,000–$180,000 range. Riskified, the only publicly traded option at a roughly $1.1 billion market cap as of early October 2026, suits an enterprise that values financial transparency and is willing to pay the highest estimated range, $70,000–$200,000, for a bundled guarantee, as long as its SKU mix doesn’t lean heavily on the “item not received” disputes the guarantee excludes. Kount remains the hardest to benchmark against the other three given the complete absence of public pricing data, making it a reasonable evaluation candidate mainly for publishers already committed to the Equifax ecosystem.

The bigger structural point sits above any single vendor choice: with Visa’s Acquirer Monitoring Program threshold now at 1.5% instead of 2.2%, and friendly fraud on track to represent 22% of global chargebacks in 2026 per Juniper Research, a publisher without one of these systems in place is operating with less margin for error than it had even a year ago. Picking the “best” vendor matters less than picking one and integrating it properly before the next monitoring-program cycle starts counting.

Frequently Asked Questions

What’s the difference between fraud detection software and a chargeback guarantee?

Fraud detection software scores transactions and recommends approve, decline, or review decisions. A chargeback guarantee is a separate financial commitment, usually bundled into the price for platforms like Forter and Riskified, where the vendor reimburses the merchant for losses on transactions it approved that later turned out to be fraudulent. Not every fraud-detection platform includes one by default. Sift, for example, doesn’t bundle a standard guarantee.

Which is cheaper: Kount, Sift, Forter, or Riskified?

Based on third-party mid-market estimates, Sift is typically the least expensive at $40,000–$100,000 per year since it doesn’t bundle a standard chargeback guarantee. Forter runs $60,000–$180,000 and Riskified runs $70,000–$200,000, both reflecting the cost of a bundled guarantee. Kount doesn’t have public pricing estimates available, so it can’t be reliably ranked on cost against the other three.

Why do game publishers specifically need fraud detection software?

Digital goods have no physical delivery record, which removes a common evidence source retailers use to dispute fraudulent chargebacks. Combined with account takeover risk from saved payment methods and friendly fraud from shared family cards, game storefronts face a higher card-not-present chargeback rate (0.6%–1.0%, per Chargebacks911) than card-present retail.

What is Visa Compelling Evidence 3.0 and why does it matter for gaming?

CE 3.0 lets a merchant defeat a card-not-present fraud dispute by showing two prior undisputed transactions on the same card, made 120–365 days earlier, sharing matching identifiers like IP address or device ID. Game platforms already capture this kind of account and device history, making CE 3.0 evidence easier to assemble than for merchants with no digital account trail.

Is friendly fraud really that common for game purchases?

Industry-wide (not gaming-specific) data suggests it’s significant: Chargebacks911’s 2026 field report found merchants estimate friendly fraud accounts for 43.8% of their chargeback losses on average, and Juniper Research projects friendly fraud will make up 22% of global chargebacks in 2026. No vendor in this comparison has published a gaming-exclusive friendly-fraud percentage, so treat industry-wide figures as directional context rather than a gaming-specific statistic.

Can a small indie studio afford any of these platforms?

All four vendors target mid-market to enterprise merchants, and none publish self-serve, low-volume pricing. A small indie studio selling primarily through an established storefront (Steam, a console store, or a mobile app store) is typically already covered by that storefront’s own built-in fraud tooling and wouldn’t need a separate contract with Kount, Sift, Forter, or Riskified until it reaches meaningful direct-sales volume of its own.

Does switching fraud vendors risk blocking legitimate customers during the transition?

It can, which is why enterprise migrations run the new vendor in shadow mode (scoring transactions without acting on them) for 30–60 days before cutover, and migrate one lower-risk SKU or region first rather than switching an entire storefront at once.

Do these vendors guarantee protection against every type of chargeback?

No. Riskified’s guarantee, for example, explicitly excludes “item not received” and “item not as described” disputes, covering mainly unauthorized-card and stolen-payment-information cases. Publishers selling hybrid physical/digital bundles should read guarantee exclusions carefully rather than assume blanket coverage.