Ecommerce Fraud Detection Software: Cost and Fit Compared

Ecommerce fraud detection software is the layer that decides, in the time it takes a checkout page to load, whether an order is real enough to ship. For a $3M–$30M Shopify Plus operator, the actual decision is rarely which vendor screens fraud better — most of the named vendors converge on similar detection quality — it is which pricing model, revenue tier and switching cost fit a specific order book. Every ranking comparison names roughly the same seven tools — Signifyd, Riskified, Forter, Kount, SEON, Sift and ClearSale — and stops at features. None states what any of them actually charge, which tier of operator each one is built for, or what manual chargeback review costs in labor while a brand is still deciding. This piece fills those three gaps directly, with every figure traced to the vendor’s own page or marked where no vendor publishes one.

What Does Ecommerce Fraud Detection Software Actually Cost?

Two of the seven vendors compared here publish an actual number on their own pricing page; the other five require a sales call before a dollar figure appears anywhere. SEON’s own pricing page lists a Starter plan at $699 a month for 2,500 fraud checks, with a Premium tier above it priced by quote. ClearSale’s own pricing page states its Growth plans start at $250 a month, moving to a Custom Solutions track for merchants processing 50,000 or more orders a month. Signifyd, Riskified, Forter and Kount each confirm, on their own pricing pages, that cost is charged only on approved orders and varies by order volume, vertical or average order value — mechanically similar structures, and none of the four states the actual percentage anywhere a prospect can read without a login.

The pattern across all seven is a guarantee-based or event-based fee tied to activity a merchant doesn’t fully control until after the fact. The bill moves with order volume and approval rate, not with a flat software-licence line a finance team can budget before the year starts.

Which Revenue Tier Should Buy Which Fraud Detection Tool?

A $3M–$30M Shopify Plus operator’s fit depends more on order volume and average order value than on revenue alone, because every guarantee-based vendor here prices against approved-order activity rather than a flat licence. A brand near the $3M end with a high average order value — jewellery, furniture, a premium supplement subscription — generates fewer orders than a $20M apparel brand at the same revenue, so a percentage-of-order-value fee from Signifyd, Riskified, ClearSale or Forter scales down naturally with order count even as the dollar risk per order stays high. The same brand evaluating SEON’s flat Starter tier should count actual monthly checkout attempts against its published 2,500-check cap before assuming the flat rate is cheaper — a low-order-count, high-AOV brand can sit under that cap for years, while a higher-order-count brand nearer $30M in apparel or beauty can clear it inside a single month and land on Premium’s unpublished quote anyway.

The order-volume signal, read against each vendor’s own stated threshold, is a faster first filter than revenue alone:

Before signing anything, check whether Shopify Protect already covers the exposure. Shopify’s own help documentation describes it as a free program that reimburses eligible fraud-reason-code chargebacks, but the coverage is narrow: it applies to Shop Pay orders only, on physical, fulfilled line items shipped within the US. A subscription brand billing recurring orders outside Shop Pay, or a brand selling digital goods, gets no coverage from it and is evaluating the seven tools compared here from zero, not from a free baseline. A brand that hasn’t yet turned on native rule-based screening at all should start with the Stripe Radar and Shopify Flow settings covered in our ecommerce fraud prevention guide before paying for any of the seven — several of the failure modes a paid tool is sold to fix are free settings changes first.

What Does Manual Chargeback Review Actually Cost in Labor?

No vendor publishes what manual fraud review costs a merchant in staff time, because removing that cost is what their own software is sold to do. The worked table that follows is illustrative arithmetic, not a measured industry number, built on the U.S. Bureau of Labor Statistics’ median hourly wage for claims adjusters, examiners and investigators, $37.51 in May 2025, as the closest published proxy for the wage of a person qualified to read a flagged order and decide whether to ship it. An in-house reviewer doing this work at a smaller ecommerce operation may earn nearer a generalist ops wage than a dedicated claims-examiner wage, so treat this as an upper-bound method rather than a fixed cost.

At an illustrative 4,000 orders a month with a 3% manual-review queue, thirty hours of review time a month comes to roughly $1,125 — about $13,500 a year, before counting the sales lost to slower shipping on every order sitting in that queue, or the customer-service time spent on the false declines a manual reviewer, working faster than the process really allows, waves through or blocks incorrectly. A brand’s real flag rate and review time are its own to measure — pull the count of orders manually held for review last month and multiply by however long that actually takes, rather than trusting either a vendor’s implied time savings or this table’s illustrative inputs.

Signifyd: A Chargeback Guarantee Priced on Approved Orders

Signifyd’s own pricing page states it charges a percentage of order total only on orders it approves, with no charge on an order it declines, and the exact percentage moving with vertical, order volume and average ticket price rather than sitting on a published rate card. The pitch is a 100% chargeback guarantee on anything it approves, which shifts the dispute cost off the merchant for that order entirely.

Not for: a brand that mostly wants better fraud-scoring signal to run its own decisions through, rather than handing the ship-or-not call and the chargeback liability to a third party. Signifyd’s guarantee model pays off financially once approved-order volume is high enough that the guarantee fee reliably costs less than the fraud losses it replaces — a brand with an already-low chargeback rate is paying for insurance against a risk it may not carry much of.

Riskified: Pay-Per-Approved-Order, Built for Enterprise Retail Volume

Riskified’s model, like Signifyd’s, charges a fee on every order it approves and guarantees, scaled to that order’s value rather than a flat per-transaction rate. Riskified’s own pricing and chargeback-guarantee pages require a live quote and do not publish the percentage; the same approved-order-fee structure is corroborated by independent payments-industry coverage, including Cardfellow’s write-up on how the model works, without a public number there either.

Not for: a brand still under a few thousand orders a month expecting the same commercial attention a nine-figure retailer gets. Riskified is an enterprise-oriented fraud platform, publicly known for serving high-volume retail brands, and a fee model scaled to approved-order value does not by itself guarantee a leaner $3M–$10M order book the same pricing leverage or support tier a much larger account negotiates.

Forter: Custom-Quoted Decisioning With an Optional Uncovered Tier

Forter’s own pricing page confirms cost is set by a custom quote rather than a published rate, built from monthly order volume, risk profile, which modules a brand licenses, and contract length. The detail worth knowing before a sales call: Forter offers an “uncovered” contract option that drops the chargeback guarantee and its premium, for a brand that wants Forter’s decisioning without paying for the insurance layer on top of it.

Not for: a brand that wants one number to compare against Signifyd or Riskified before ever speaking to sales. Because Forter’s quote depends on module selection as well as volume, two brands with identical revenue can receive genuinely different quotes, which makes it the hardest of the seven to price-shop without already being partway through its sales process.

Kount: Identity-First Prevention, Positioned Against the Guarantee Model

Kount, acquired by Equifax in February 2021 per Equifax’s own investor relations announcement, prices by custom quote — its own pricing page states cost depends on “business goals and transaction volume, among other factors,” with no figure published anywhere on the page. Kount’s own marketing draws an explicit line against the guarantee vendors above it on this list, stating plainly that it is “a better way to stop chargebacks” than a reimburse-after-the-fact guarantee, and positioning its identity and device-signal decisioning as prevention rather than insurance.

Not for: a brand that specifically wants someone else to absorb chargeback liability rather than just reduce it. Kount’s own positioning is built around stopping a chargeback before it happens, not reimbursing one that already did — a merchant buying Kount still owns the dispute-filing work on whatever fraud gets through, unlike Signifyd, Riskified, Forter or ClearSale, each of which takes on the chargeback itself once it approves an order.

SEON: The One Vendor Here That Publishes a Starting Price

SEON’s own pricing page is the only one among these seven that states an actual monthly figure without a sales call: a Starter plan at $699 a month for 2,500 fraud checks, ten users and fifty custom rules, with a Premium tier above it moving to unlimited checks and custom pricing. That transparency is real and worth crediting. It is also easy to outgrow — a brand running more than roughly 2,500 checks a month is priced out of the published number and into the same quote process as everyone else on this list.

Not for: a $20M–$30M brand with high order volume and a correspondingly high check count. The published Starter price stops being the relevant number well before that revenue level, and the actual Premium quote is unpublished like the rest.

Sift: Event-Based Coverage Across Signups, Logins and Orders, Not Just Checkout

Sift publishes no pricing page at all — there is no public rate, tier or starting figure anywhere on its site. Its own product page for payment fraud describes a broader scope than a checkout-only tool: it screens fake account creation and account takeover attempts alongside transaction fraud, drawing on a Global Data Network that Sift’s own site states processes more than a trillion events a year across the customers it serves collectively (vendor-reported).

Not for: a brand narrowly trying to solve chargebacks on Shopify checkout and nothing else. Sift’s pitch is coverage across the whole customer journey, priced accordingly through a sales process rather than a checkout-sized number, which is more tool — and likely more cost — than a brand with a single, checkout-only fraud problem needs to buy.

ClearSale: Per-Approved-Order Pricing With a Published Volume Threshold

ClearSale charges only for orders approved after review, per its own pricing page — a declined order costs nothing, the same approved-orders-only structure as Signifyd, Riskified and Forter. What ClearSale states that the other three don’t is a description of what sits behind the automated score at its top service tier: a five-layer review process — AI models, fraud rules, a global fraud database, external sources and a secondary human review — rather than a single model making every call alone.

Not for: a brand comfortable relying entirely on automated decisioning with no human layer in it, at least not without checking the tier first. ClearSale’s own description ties that secondary human review specifically to its top “Complete Decision” service level — worth confirming a Growth-plan quote carries the same analyst layer before assuming every plan does.

What Does It Actually Cost to Switch Fraud Detection Vendors?

Switching effort here is rarely the integration — every vendor on this list ships a Shopify app or a documented API, and the technical connection typically drops in inside a day or two. The real switching cost is the guarantee gap the technical swap creates. Visa and Mastercard both allow a cardholder up to 120 days to file a dispute, counted from the transaction or delivery date depending on the reason code, so a chargeback on an order the old vendor approved can still arrive months after that vendor stopped screening new orders. A brand that cancels the old contract the day the new tool goes live is often left holding chargebacks on orders nobody guaranteed by the time the dispute lands, because the old vendor’s guarantee applied only while the contract was active, and the new vendor’s guarantee covers only orders it approved itself. What an ecommerce chargeback actually is, and how the evidence deadline works is worth reading before that gap catches anyone by surprise.

The safer sequence is to run the new tool in shadow — scoring live orders without acting on them — against the incumbent for a full billing cycle, confirm its decisions against orders that have already shipped and settled, and only cancel the old contract once the last batch of orders it approved has cleared its own dispute window. That overlap period means paying two vendors at once for a stretch measured in months, not days, which is the real number missing from every switching comparison, and the one worth budgeting before a contract gets cancelled early to save a month of double billing.

Picking one of these seven tools solves the buying decision. It does not touch the other half of the problem: whatever tool ends up running, are its rules actually catching the fraud a specific order book produces, or only the fraud a generic ruleset assumes every merchant has? Stripe’s own data — vendor-reported — puts 25% of lapsed subscriptions down to payment failure rather than a customer choosing to leave, and a fraud-and-chargebacks problem sits right beside that number on the same payment layer: a screening rule tuned for a stranger’s first order, and never re-scored against a returning subscriber’s eleventh one, blocks exactly the customers a subscription brand can least afford to lose. That is the fraud and chargebacks work — replaying whatever rules are already active in Signifyd, Riskified or Stripe Radar against a year of real orders, scoring the false positives instead of assuming there aren’t any, and building the chargeback evidence and card-updater reconciliation that sit downstream of whichever vendor from this list ends up chosen.