The ABCs of Onboarding KPIs

Getting the merchant onboarding process right means merchants encounter no obstacles and acquirers get the data they need—fast.

Automation has not only changed the quality of merchant data, it has changed how quickly merchants provide it during the onboarding process. Now, as artificial intelligence tools stream into the acquiring industry, the demand for an expedient and accurate onboarding procedure is intensifying, just as the number of payment facilitators and payment service providers courting merchants, too.

With new tools becoming available and the need for speed intensifying, the criteria for assessing performance in the onboarding experience becomes even more critical.

“Effective merchant onboarding is complex because it must balance the needs of the payments companies with the desires of the merchants,” says Andy Vrabel, general manager of payment ecosystem solutions at LegitScript LLC, a Portland, Ore.-based merchant risk and monitoring services firm.

“Payments companies want thorough checks to ensure a prospective merchant isn’t going to engage in brand-damaging or illegal activity,” continues Vrabel, “while merchants want a quick and easy process so that they can start selling.”

As Matt Bennett, vice president of partnerships and alliances at FICO, puts it, “The primary objective of a merchant-onboarding process is to simplify the workflow and deliver a seamless customer experience.” FICO is a Bozeman, Mont.-based data-analytics provider.

Through the Funnel

Measuring how well that task can be done relies on common key performance indicators, or KPIs. Bennett lists nine KPIs, a few of which include time to activation, approval and abandonment rates, costs, and satisfaction scores. For his part, Vrabel lists speed, the completion rate, user experience, and friction and bottleneck analysis.

“Funnel pull-through rates are one quantitative indicator that helps us understand ease as we look at how many merchants start the process versus how many complete all the way through to taking payments,” says Jessica Young, managing director and head of product at Chase Payment Solutions. “Another key metric is direct client feedback on their onboarding experience, for example as measured in client-facing surveys and research.”

The funnel pull-through rate refers to the percentage of a merchant cohort that ultimately completes the onboarding process, Young says. “For example, if 100 merchants enter the onboarding flow but only 70 make it to the final screen, we would consider that a 70% pull-through rate,” she says.

“We spend time looking at each step of the journey to see drop-off rates on every page of the flow. If a particular screen has a higher drop-off rate, that suggests opportunity to improve that experience to improve overall funnel pull-through,” Young adds.

Chase, like other acquirers, wants to know how long it takes a merchant to complete the process. “Specifically, we want to ensure merchants who onboard are ultimately transacting and that we have put in the right upfront checks and responsible friction in the process to identify and prevent bad actors,” Young says.

Generally, each of these metrics is as important as another, observers say, though time to activation may be most critical. “It provides fast and seamless onboarding, allowing merchants to start generating revenue sooner, which impacts both their success and the acquirer’s growth,” Bennett says.

As Vrabel explains, “No one factor, on its own, is paramount in an effective merchant onboarding process,” he says. “A fast solution is useless if it isn’t accurate, and an accurate solution has limited use if it is rigid and unable to scale. That’s why it’s essential to balance the need for accurate analysis with the industry pressure to vet and approve merchant applications quickly.”

Using accuracy as an example, Vrabel says automated merchant category-code detection, know-your-business checks, and other data can help. The MCC is typically a four-digit number. “Our research suggests that 40% of merchants are miscoded. The most commonly misused MCCs are 5734 and miscellaneous MCCs ending in 99,” Vrabel says.

AI’s Impact

Though automated onboarding services have been available for a while, the introduction of artificial intelligence may be an accelerant and could put more emphasis on onboarding metrics.

“Automation is increasingly vital in merchant onboarding, helping payment acquirers onboard merchants faster while implementing stronger know your customer (KYC) and due-diligence practices,” says Mary Claire Williams, vice president of product management at G2 Risk Solutions Inc., a Burlingame, Calif.-based risk-management provider. “It is why many acquirers have been able to grow merchant portfolios and reduce risk exposures simultaneously.”

It is not hyperbole to say that AI is having a transformative effect, Williams says, “and its impact will only continue to grow as acquirers work to balance speed with risk mitigation.”

The additional utility of AI is manifold. “AI can synthesize billions of data points to produce a nearly instantaneous risk score, in that interest. This comprehensiveness goes far beyond the capabilities of traditional underwriting processes,” Williams says. “Moreover, AI can do this without creating unnecessary friction for legitimate businesses, which is increasingly essential in the highly competitive payment-services landscape.”

LegitScript is one vendor with AI embedded in its merchant-onboarding product suite. Its Xray AI Risk Intelligence platform combines proprietary Website crawling and AI with its universe of risk data of more than 60 million merchants and billions of data points, Vrabel says.

“Our research team views AI as the future of merchant-risk management,” he says. “Manual risk detection, though valuable for certain use cases, is slow, expensive, and unscalable, while traditional automation is inaccurate, inflexible, and limited. AI-driven solutions, however, can do things like instantaneously scan merchant Websites, automatically detect merchant category codes, and learn and adapt as rules change.”

The role of AI, or machine learning, is critical at acquirers like Chase Payment Solutions. “Machine learning is a key capability when it comes to risk and fraud models.  As we gather data on merchant profiles and behavior, the models get smarter and help to identify merchant onboarding applications that require further review,” Young says.

Bennett agrees, adding, “Yes, artificial intelligence can play a transformative role in merchant onboarding by improving speed, accuracy, and user experience while reducing manual efforts.”

Among the roles AI can fill are identity and compliance verification, he says. It can help with automated document recognition, facial recognition and liveness detection, and instant KYC and anti-money laundering compliance checks. It also can help with pricing, intelligent risk assessment, and fraud detection during the onboarding process.

Other uses include powering chatbots and text-based assistance, and automated agreement generation and e-signature demands, Bennett says.

Vetting Vendors

Another key metric is assessing the performance of onboarding services providers. At LegitScript, Vrabel, as a vendor, says third-party providers should “have a demonstrated history of accuracy, accountability, and transparency.”

Bennett suggests evaluating providers on the ease of integration, scalability, adherence to security and compliance standards, and the user experience. Also, the configurability of the platform, its cost effectiveness, customer support and service-level agreements, and reputation and track record, complete his criteria.

AI’s role is not just for the initial onboarding steps, but can factor into ongoing merchant monitoring.

“Beyond onboarding, AI helps acquirers conduct continuous merchant monitoring, critical for maintaining a secure portfolio,” Williams says. “Onboarding is one part of the merchant lifecycle, and payment providers must ensure compliance throughout its entirety. AI tools are never ‘off the clock.’ They constantly ingest information and can detect complicated behavioral patterns and suspicious activity that a merchant may attempt to conceal.”