Riskified Ltd. (RSKD)

Riskified is a software company that solves a persistent, expensive problem for online retailers and marketplaces: determining which transactions are genuine and which are fraud, chargebacks waiting to happen, or abuse. The company, incorporated in Israel but operating globally with its largest market in the United States, sells risk-intelligence software on a subscription and transaction basis to e-commerce merchants, payment processors, and marketplaces. Its customers range from mid-market retailers to some of the largest online shopping platforms in the world. The core business is connecting billions of data points about a shopper’s behavior, device, network, and purchase history to answer a single high-stakes question: should this order go through? Riskified trades on the Nasdaq under the ticker RSKD.

The problem Riskified attacks is not hypothetical. Online fraud costs merchants billions annually. Chargebacks — when a customer disputes a charge with their card issuer and the merchant loses both the merchandise and the sale price — can break small retailers and erode the margins of large ones. Friendly fraud, where a customer buys something, receives it, then claims they never authorized it, is endemic. Payment processors, credit networks, and retailers have long-standing methods to catch obvious fraud, but these methods also produce false positives, blocking legitimate orders and driving customers away. Riskified was founded in 2010 to do better, building models that could distinguish real fraud from innocent purchases with far greater accuracy than legacy rule-based systems.

The company’s software works by gathering behavioral and contextual intelligence on every transaction. It examines the customer’s device — its operating system, browser, geolocation history — and compares it against known-good patterns from past legitimate orders. It analyzes the email address, phone number, shipping and billing addresses, and the products being ordered, checking for red flags and inconsistencies. It applies machine learning models trained on years of fraud and legitimate-purchase data, constantly retuning itself as fraud tactics evolve. When a shopper attempts to check out, Riskified’s API returns a risk score and a recommendation in milliseconds. The merchant can then decide to approve the order, challenge the buyer with additional verification, or decline it.

Riskified’s business model is a hybrid of subscription and transaction fees. Merchants pay a base monthly or annual fee for access to the platform, and then pay a small transaction fee for each order analyzed and approved through the system. This creates an incentive alignment: Riskified’s revenue grows as its customers’ sales grow. The transaction-fee component also means the company benefits directly from its customers’ success — if a merchant’s fraud rate drops and it can confidently approve more orders, Riskified earns more. Unlike a pure subscription model where the provider is paid the same whether the customer thrives or struggles, Riskified shares in the upside of better fraud prevention.

The company’s customers face competing pressures that Riskified must navigate. On one hand, merchants want to minimize fraud losses and chargebacks, which can be ruinous. On the other hand, they want to approve as many legitimate orders as possible — declining a real customer is lost revenue and a bad user experience. Blocking too many transactions with excessive friction, like forcing every shopper through identity verification, drives people to competitors. Riskified’s machine learning engines are designed to maximize the area between these two imperatives, approving good orders while stopping bad ones, and doing so while improving over time as the models ingest more data.

The company’s expansion has depended on moving into adjacent problems. Beyond fraud detection, Riskified has built modules for chargeback management, helping customers understand which disputes are worth contesting and how to present evidence to the card networks. It has expanded into payment optimization, identifying when an order should be sent to one payment method versus another. It has ventured into identity verification and buyer authentication, using behavioral signals and document checks to confirm a customer is who they claim to be. Each expansion broadens the value proposition and deepens the integration between Riskified and its customers’ transaction flows.

The competitive landscape includes both established players and newer entrants. Legacy fraud-detection tools from payment networks and processors, built on rule-based systems and historical patterns, still dominate by sheer ubiquity but are increasingly seen as crude and prone to false positives. Rivals like Stripe Radar, PayPal’s internal tools, and purpose-built fraud startups offer machine-learning alternatives, and Riskified competes on the sophistication of its models, the breadth of its data, and the depth of integration its customers have built. The company’s stickiness comes from the fact that changing fraud providers is not simple — a merchant has to retrain its team, integrate new APIs, and potentially accept a period of transition where the new provider’s models are learning and less accurate.

Riskified is not alone in selling to e-commerce merchants, but it is one of the few companies at scale focused exclusively on risk. That focus was a strategic choice, and it is both a strength and a vulnerability. The strength is that the company has built extraordinary depth in fraud science and chargeback management, hiring experienced fraud fighters and data scientists who understand the problem from first principles. The vulnerability is that the company’s growth is tightly coupled to e-commerce growth, the health of its customer base, and the evolution of fraud itself. During pandemic lockdowns, e-commerce boomed and fraud changed character; in the years after, as retail shifted back into balance, transaction volumes moderated and fraud pressure eased, affecting Riskified’s customer demand.

The company’s path to profitability has been uneven. As a SaaS vendor, Riskified’s unit economics depend on acquiring customers, onboarding them without burning capital, and retaining them long enough to exceed the cost of acquisition. The company has invested heavily in sales, marketing, and product development, running at a loss for years to build market share. By 2024, it had moved toward profitability, but margin expansion remains a focus as the company matures and operational efficiency becomes as important as growth.

Investors in Riskified are backing a company operating at the intersection of e-commerce and fintech, in a space where fraud is perpetual and evolving. The company’s success depends on keeping its machine-learning models ahead of fraud tactics, maintaining customer retention as competition intensifies, and scaling the platform efficiently. The fraud-prevention business is recession-resistant in some ways — merchants care about losses even in downturns — but vulnerable to e-commerce saturation and the consolidation of the retail landscape around a smaller number of massive players that might build their own risk capabilities instead of outsourcing.