Signifyd

Service Overview

Signifyd is a US commerce protection provider headquartered in San Jose, with additional offices in Denver, New York, Mexico City, Sao Paulo, Belfast, London and Budapest. It positions itself as an end-to-end commerce protection platform, using its Commerce Network to lift conversion, automate customer experience and remove fraud and consumer abuse risk.

The part that must be read precisely is the deliverable. Most risk providers hand back a score or a recommendation and leave the merchant to decide whether a human should look. Signifyd hands back a binary outcome: ship, or do not ship - backed by a 100% financial guarantee against fraud and consumer abuse losses. That single choice defines the entire business model.

There are 4 product lines: Session Intelligence reads behavioural signals at the session layer, Account Integrity handles account integrity and takeover, Purchase Optimization manages pass rates in the buying step, and Returns Assurance covers the returns stage. Together they map onto pre-checkout, account, purchase decision and after-sale, which is what "end to end" means here.

On history: co-founders Mike Liberty and Raj Ramanand came out of PayPal and started with order scoring, then moved to guaranteed fraud protection, then expanded into policy abuse, returns fraud and abuse, unauthorised resale, promo abuse, optimisation under European SCA requirements, and products aimed at financial services. That trajectory tracks the industry's shift - fraud moved from stolen cards to real identities exploiting rule gaps, so risk control moved from transaction features to policy abuse.

Industry context it cites gives a sense of scale: e-commerce was 20% of global retail in 2024 and is projected at 23% by 2028, while global e-commerce fraud losses were 44 billion USD in 2024 and are projected at 107 billion USD by 2029 (Juniper Research). Its own 2024 observations were sales up 7% year over year, high-risk orders up 20% and consumer abuse up 23% - high-risk orders and abuse growing faster than sales, which has a direct implication for merchants. Its site also states that Fast Company named it one of the 10 most innovative AI companies, and that Digital Commerce 360 has ranked it a leading provider of payment security and fraud prevention for large e-commerce brands for 4 consecutive years.

Key Strengths

• What it actually replaces is the manual review queue: not another rule engine but the question "does this order need a human", turned into "ship or do not ship, and we cover the loss". For e-commerce with a review team, an entire labour line can be removed rather than adding a tool to configure.
• The guarantee converts exposure into contract terms: a 100% financial guarantee against fraud and consumer abuse moves chargeback losses, card-network fines and related accounting work off the merchant's P&L. For thin-margin, chargeback-heavy categories that matters more than model precision.
• ROI comes from two places: the review labour saved, and the good customers who would have been declined. The second is routinely underestimated - false declines never appear in reporting, yet they are real lost revenue that gets more expensive as acquisition cost rises.
• Network data spans 100+ countries: the hardest part of cross-border business is the lack of a local behaviour baseline, because an address format or delivery path carries entirely different risk meaning in different markets. Multi-country network data reduces the "only safe option is to decline" penalty in unfamiliar markets.
• Abuse is covered, not just fraud: policy abuse, returns fraud, unauthorised resale and promo abuse never appear in chargeback reports; they sit in after-sales cost and marketing budget and are usually written off as a cost of doing business. Bringing them into the guarantee gives those losses a quantifiable home.
• There is a dedicated answer for European SCA: 3DS and SCA are mandatory in Europe but also suppress conversion. Offering optimisation means it treats exemptions and challenge strategy as a conversion variable rather than a compliance switch.

Product Ecosystem

Session Intelligence

Collects behavioural signals at the session layer before checkout to judge whether the purchase intent is genuine. Separating human behaviour from automated scripts happens here, forming the input for later decisions.

Account Integrity

Covers registration, login and account changes, identifying account takeover and bulk registration. It shares the same network data as payment fraud, closing the seam where the account side looks clean while the transaction side goes wrong.

Purchase Optimization

The conversion-facing module. It aims to let good orders that would have been declined through, turning false declines from an accepted cost of risk control into an optimisable revenue line.

Returns Assurance

Covers returns fraud and abuse - empty-box returns, buy-and-swap, late returns, policy abuse. Most risk systems have no entry point for these losses at all.

Guarantee and Decision Interface

Integrated by API into order systems, returning ship or do-not-ship and recording guarantee attribution for later dispute matching and claims.

Limitations

• The guarantee has eligibility gates and exclusions: category, ticket size, shipping geography and delivery method all affect whether and how much is covered. High-risk categories or emerging-market shipments often need separate assessment, and whatever sits outside stays with the merchant. Sign eligibility terms and the exclusion list against your actual order mix before signing.
• You hand over final authority: the guarantee presumes you follow the decision. Orders that support or risk staff release for relationship reasons generally fall outside cover. For experience-first teams this is a genuine workflow conflict.
• The two error directions are asymmetric: approving a fraudulent order costs goods plus chargeback fees and card-network fines; declining a good customer costs this order plus lifetime value. The guarantee covers the financial consequence of the first, not the operational consequence - dispute-rate monitoring and account health are still affected.
• Data sharing and cross-border compliance cost: order, session and consumer behaviour data must be transferred. Cross-border operations trigger GDPR and personal data export assessments, with a separate path where export applies, and legal and security review often outlast integration.
• The model depends on scale and data quality: network data value correlates with merchant size and audience overlap. Narrow verticals and very low-volume merchants receive tighter underwriting and cannot realise the premium.
• No coverage of regulated obligations: AML transaction monitoring, onboarding review and sanctions screening are out of scope. Licensed institutions still need a separate system and end up running two risk stacks.
• Citable industry data reflects the vendor's framing: growth and abuse figures it cites support its market case and belong to vendor narrative. Use them as background for budgeting, not as expected returns for your own business.

Use Cases

• Mid-to-large e-commerce stores (★★★★★): high order volume and review cost make the combination of guarantees and automated decisions easily measurable, with the clearest ROI.
• High-ticket and brand e-commerce (★★★★☆): large per-dispute amounts and expensive manual review per order make outsourcing decision and guarantee together clearly economic.
• Categories sensitive to returns and abuse (★★★★☆): apparel, consumer electronics and beauty have high return rates and strong abuse incentives, where returns assurance and abuse prevention cover losses conventional risk tools never see.
• Cross-border multi-market sellers (★★★☆☆): network data across 100+ countries reduces conservative declining in unfamiliar markets, but underwriting terms differ by market and must be confirmed one by one.
• Mainland China operations (★★☆☆☆): usable when a mainland entity runs cross-border business, but transferring order and consumer data abroad must follow the Personal Information Protection Law and the applicable export rules, so compliance assessment comes first. Purely domestic business does not need overseas guaranteed services and has domestic equivalents without an export step.
• Very low-volume new stores (★☆☆☆☆): least suitable. Guarantee eligibility plus cold-start network data means tighter terms and higher unit cost, so returns do not cover the premium.
• Teams that must keep manual final approval (☆☆☆☆☆): least suitable. Manual overrides of ship decisions void the guarantee, which means buying protection you cannot use.

Pricing

Main cost drivers are reviewed order volume, GMV, ticket-size distribution, shipping countries and regions, guarantee scope, whether manual investigation is included and the integration approach. Signifyd does not publish standard rates; headings and final rates follow the current official quote and contract terms.

FAQ

• Does the "100% guarantee" cover everything?
  No. It covers fraud and consumer abuse losses that occur inside the contractual scope after you have followed its decision. Category, ticket size, shipping geography, delivery method and execution consistency are all conditions. The right pre-signature step is obtaining the eligibility terms and exclusion list and checking them against your real order mix.
• Will it conflict with our own review team?
  There will be friction. The guarantee requires following the decision, and orders released manually generally fall outside it. The usual pattern is handing approval to the system and limiting humans to the referrals it sends. If the business genuinely needs a manual override right, accept that those orders are uninsured and document the risk owner internally.
• How does this compare with a scoring-only provider?
  Decide whether you lack people or risk capacity. Scoring-only is cheaper and more flexible, but you staff the review team and carry every chargeback. Guaranteed is more expensive and more constrained, but outsources both loss and review labour. Many merchants start with scoring, then use real dispute and false-decline data to judge whether the premium pays.
• Is it only about fraud, or do refund disputes count?
  More than fraud. Returns Assurance covers returns fraud and policy abuse, and promo abuse and unauthorised resale sit within abuse prevention. Recurring-billing disputes and non-delivery service disputes may have separate rules, so confirm that the dispute types you actually see fall inside the guarantee semantics.
• Can a mainland China company use it for cross-border business?
  It can be evaluated, but data compliance has to be solved first. Transferring order and consumer data abroad involves the Personal Information Protection Law and the applicable export rules, so confirm whether a security assessment, standard contract or certification path applies. Purely domestic business does not need this service and has no export question.

Source: original 16IDC content. Product capabilities, supported countries and fee definitions follow the official documentation and your contract.