Do You Need Fraud Prevention? (Signifyd vs Riskified vs Doing Nothing)

Here's the uncomfortable truth most fraud-tool sales decks won't lead with: the order you just blocked because it "looked risky" was probably a real customer. And that mistake costs you more than the fraud you're scared of. I've sat with founders who installed a fraud tool the week after a brutal chargeback, convinced they were finally protected. Six months later their approval rate had quietly dropped, repeat customers were complaining their cards got declined for no reason, and nobody connected the dots. The tool was "working." Revenue wasn't. So before you go shopping for Signifyd or Riskified, let's figure out whether you actually need a dedicated tool at all, and what it should really do for you.

The story the numbers tell

Fraud is real and it's growing. Global ecommerce fraud losses hit roughly $48 billion in 2025, and the all-in cost of a single chargeback averages around $110 once you count the lost goods, shipping, the dispute fee, and the staff time to fight it.[1] Processor chargeback fees alone run $20 to $50 per dispute, and global chargeback volume is forecast to keep climbing past 337 million transactions.[2] First-party fraud, the "I didn't order that" lie from a real customer, is now the single biggest fraud type, about 36% of reported fraud.[2] So yes, the threat is genuine. But here's the stat that should reframe the whole conversation. False declines, legitimate orders that get wrongly blocked, cost retailers an estimated $443 billion a year globally. That's roughly nine times more than actual fraud losses.[3] In the US alone, merchants lose about $118 billion to good orders they refused, and 39% of those rejected customers never come back.[3] Read that again. The bigger leak in your business probably isn't the fraudster. It's the customer you're treating like one.

Chart: false declines vs ecommerce fraud loss, global, USD billions/year. Sources: Signifyd (false declines), Statista/industry (fraud loss). Built with QuickChart. That's the real job of a modern fraud tool. It's not just "block the bad guys." It's approve more good orders with confidence while shifting the chargeback risk off your books. The good tools are an approval engine first and a shield second.

When doing nothing actually wins

Let me say the thing the vendors won't. If you're doing under a few hundred orders a month, your fraud rate is low, and your average order value is small, you probably don't need to pay anyone yet. Your platform's built-in fraud analysis (Shopify, Stripe Radar, BigCommerce) plus a bit of manual review on the weird-looking orders will carry you a long way for free. The trigger to upgrade isn't fear. It's math. Add up your monthly chargeback losses, your dispute fees, the hours your team spends reviewing orders, and your best guess at good orders you're rejecting. When that total reliably exceeds what a guaranteed tool would charge, you've crossed the line. For most stores that's somewhere north of a 0.5% to 1% chargeback rate, high AOV, or a category that attracts fraud (electronics, gift cards, luxury, anything resellable).

Photo: card payment in progress. Source: Unsplash (rupixen), free to use under the Unsplash License.

How the chargeback-guarantee model works

This is the part worth understanding before you sign anything. Both Signifyd and Riskified run a guarantee model. They use machine learning to score each order, then on the orders they approve they take on the financial liability. If an approved order turns into a fraud chargeback, they reimburse you. You pay a small percentage of each approved order, typically in the 0.3% to 1% range depending on your vertical, volume, and ticket size.[4] The clever bit is the incentive alignment. They only get paid on orders they approve, so they're motivated to approve as many good orders as possible rather than just blocking everything to look safe. That's the opposite incentive of a tool that charges you a flat fee to decline aggressively. A "native or manual" setup doesn't shift any liability. You eat every chargeback. The trade is no per-order fee versus all the risk and labor staying on you.

Signifyd vs Riskified vs doing nothing

How I'd actually evaluate this

Don't start with the vendor demo. Start with your own numbers. Pull your last 12 months of chargebacks, dispute fees, manual-review hours, and your approval rate. If you can't see your false-decline rate, that's your first red flag, because it's likely the biggest cost on the list and you're flying blind. Then, when you talk to Signifyd and Riskified, ask the questions that actually matter. What's the projected approval-rate lift, not just the fraud reduction? What exactly does the guarantee cover (fraud only, or non-fraud chargebacks too)? How fast do reimbursements land and how messy is reconciliation? Signifyd leans toward broader coverage including some non-fraud chargebacks; Riskified gets repeat praise for clean, easy-to-reconcile billing. Both will run an analysis of your store and quote a custom rate, so make them show their work on the approval-rate math. And genuinely consider "keep doing nothing for now" as a real option. If a guarantee tool would cost you 0.7% of revenue to save 0.4% in losses and labor, it's a bad trade today. Revisit it when you scale or move into a riskier category. If you're piecing together fraud scoring, payments, and international checkout on one stack rather than bolting on five tools, an all-in-one like Sellarix is worth a look, because the fewer disconnected systems making approve/decline calls, the fewer good customers fall through the cracks.

The takeaway

Fraud prevention isn't really about stopping fraud. It's about approving more good orders without getting wrecked by the bad ones. A guarantee tool earns its fee when your losses plus your false-declines plus your review labor clearly outweigh the per-order cost, and not a moment before. Until then, your platform's built-in tools and a little manual review are doing more than you think. So here's my question for you: do you actually know your false-decline rate? Because if you don't, you might be quietly losing nine times more to your own caution than to any criminal.

Sources

• Signifyd, Chargeback guarantee: a complete guide. https://www.signifyd.com/blog/chargeback-guarantee-a-complete-guide-to-modern-fraud-protection/
• Signifyd, Payment Approval Percentage (false-decline data). https://www.signifyd.com/blog/payment-approval-percentage/
• Signifyd, Pricing and Plans. https://www.signifyd.com/pricing/
• Signifyd, Complete Chargeback Protection. https://www.signifyd.com/complete-chargeback-protection/
• Chargeflow, Chargeback Statistics: Trends, Costs & Solutions (2025). https://www.chargeflow.io/blog/chargeback-statistics-trends-costs-solutions
• Mastercard, What's the true cost of a chargeback in 2025. https://www.mastercard.com/us/en/news-and-trends/Insights/2025/what-s-the-true-cost-of-a-chargeback-in-2025.html
• Riskified, Signifyd vs Riskified comparison. https://www.riskified.com/lp/signifyd-vs-riskified/