1

EBOOK

And what you can do about it

3 Warning Signs that Your ID Document Verification Process is Hurting Your Bottom Line

2

Table of Contents

3 Introduction

5 Warning Sign 1: Applicants are waiting minutes, hours, or days to be approved

7 Warning Sign 2: Large numbers of legitimate customers are being falsely rejected

11 Warning Sign 3: Synthetic identities, scammers, counterfeiters, or underage users are infiltrating your ecosystem

14 ID document verification solution checklist

In the aftermath of the digital acceleration brought on by the pandemic, businesses from banks and telehealth to gaming and two-sided marketplaces are aiming to meet the demand driven by consumers with a clear preference for using digital channels – even for risky transactions.

Companies are also looking for ways to offset revenue losses from current economic challenges. By automating onboarding and customer service processes for new digital products and models, businesses are reducing operational costs and reversing supply chain issues, while improving user experience.

According to McKinsey, digital adoption has taken a quantum leap. Globally, 38% of companies say their current strategic posture is to create a competitive advantage around their online offerings, while another 19% report they are refocusing their entire business around digital models. On average, organizations say that 50% to 60% of their product portfolios have been digitalized, up from 25% to 41% before the pandemic.

Meanwhile, this elevated focus on digital channels also highlights the need to verify the identities of people who apply for or need access to digital financial products and consumer services. Companies bound by regulatory requirements must conduct stringent Know Your Customer (KYC) or age verification checks, while companies with a core business of bringing strangers together need to reinforce an environment of trust and safety.

EBOOK

If you’re implementing a roadmap to digitalize more products, models, and high-risk transactions online, you’re not alone

38%

Percentage of global

companies that say their

current strategic posture

is to create a competitive

advantage around their

online offerings.

3

Ensuring accurate identity verification is critical to good business outcomes Ensuring accurate identity verification while getting legitimate customers into the ecosystem as quickly as possible is vital to business success. In particular, fully-automated, government-issued ID document verification is being used to quickly onboard and re-authenticate customers. Regulated industries use it in their workflows for step-up verification when outcomes from KYC or fraud checks are inconclusive, while others apply it at the top of the funnel to autofill applications for easier sign-up, which also cuts down on manual entry errors.

Unfortunately, many companies rely on fully or partially manual reviews or subpar technology for ID verification. These manual reviews, in addition to delaying onboarding or introducing friction, often also introduce accuracy errors with false approvals, false rejections, or both. Instead of quickly approving applicants to support growth, these methods can slow down onboarding or subject innocent people to possible harm by giving fraudsters or criminals access to users on the platform.

Simultaneously, artificial intelligence has enabled fraudsters to scale their operations for creating sophisticated forged IDs, deep fakes, 3D masks, and more. The output is so convincing that human eyes in a manual review process often fail at detecting these fake artifacts, resulting in false approvals and ultimately fraud losses.

Further, many compliance and risk professionals believe that their current KYC program is adequate. However, these programs are often falling short of requirements, as evidenced by the more than $55 billion in anti-money laundering fines that were doled out in 2022 — a 50% surge over the prior year.

Here are three warning signs that indicate your documentary checks are potentially risking poor user experience, your company’s bottom line, the safety and security of other users — or all of the above. If these warning signs sound familiar, we’ll walk you through countermeasures for creating a best practices approach.

EBOOK

Companies bound by

regulatory requirements

must conduct stringent KYC

or age verification checks.

4

5

EBOOK

Warning sign 1 Applicants are waiting minutes, hours, or days to be approved

If your ID verification decisions take minutes, hours, or days, then your process is definitely costing you customers. Response times higher than fifteen seconds are typically using a partially or fully manual review for document verification.

When an applicant has to go through an arduous process that requires snapping numerous ID and selfie photos, completing complicated movements and tests to prove “liveness,” only to be escalated to a lengthy review, there’s a high possibility that they will lose interest or go elsewhere.

How this hurts your bottom line Identity proofing that involves a selfie with biometrics to match the ID and detect liveness is an extremely strong verification tool, but friction-filled approaches are cumbersome and error-prone. Applicants get frustrated with a poor capture experience or active liveness requirement that is often finicky, bandwidth-eating, and slow—and then to be put into limbo on a decision, which can take hours or days when volume spikes.

Human reviewers are also susceptible to human problems. They have bias, unconscious or otherwise, and get tired or lose focus, especially when reviewing a high volume of IDs and selfies that are increasingly difficult to to identify as fake due to AI generation enhancements. The decision on what to accept or reject is left up to the individual with standards that might not be current or fit your requirements.

At best, these friction-laden processes create a poor user experience; at worst, they cause customers to drop-off and go to the competition, which equates to revenue loss. And, that doesn’t consider the cost of manual reviews and fraud write-offs.

Automation aids precise,

fast decisioning and

almost entirely

eliminates the need for

manual reviews.

6

EBOOK

The Solution

The benefits of using an automated identity verification solution instead of manual ones leads to higher customer satisfaction, improved scalability, operational cost efficiencies, and an overall better way to prevent fraud and get good customers to revenue faster. Automation aids precise, fast decisioning and almost entirely eliminates the need for manual reviews.

• Collect a high-quality photo of a government issued ID — without blur or glare

• Conduct document validity checks quickly and efficiently with minimal resubmits

• Match a selfie photo to the image on the ID

• Use the same selfie to perform a liveness check to detect spoofing attacks

............

Automation aids precise,

fast decisioning and almost

entirely eliminates the need

for manual reviews.

Automated ID verification with guided image capture increases user satisfaction In the digital era, consumers have high expectations for fast turnaround. Automated ID verification with guided image capture is a key component to a superior user experience.

Automated document verification guides prospects through the image capture process in real time via a mobile app that reduces errors and ensures a high completion rate on the first try. The most robust document verification solution will:

7

EBOOK

Warning sign 2 Large numbers of legitimate customers are being falsely rejected

If you’re suffering from low conversion related to ID verification checks, then you may be falsely rejecting a high percentage of good customers and seriously hurting your organization’s bottom line. A false rejection rate (FRR) is the percentage of legitimate customers that are incorrectly rejected during ID verification. High-quality document verification solutions typically yield a less than 10% false rejection rate upon post-hoc review.

Subpar image capture and poor data extraction processes are frequent triggers behind rejecting real users. Here are four primary causes:

• Left on their own, users may inadvertently take poor quality document and selfie images, which often are distorted, blurry, and poorly lit.

• In the age of typing on tiny smartphone keys, typos in submitted PII are more prevalent than ever causing incorrect input that will not match with extracted ID data.

• Optical character recognition, or OCR-only solutions present inherent limitations for extracting data from text-based fields when faced with backgrounds that may include holograms, watermarks, and glossy surfaces. If accurate data isn’t extracted, it will be rejected when compared to user input.

• API-based integrations typically don’t incorporate automated, client-side quality checks, often yielding partial or low resolution images.

High-quality document

verification solutions

typically yields a less than

10% false rejection rate upon

post hoc review, or right

after an event such as

a transaction takes place.

8

EBOOK

How this hurts your bottom line

A recent study from Innovatrics showed that 63% of financial services applications are abandoned when the process is inadequate when you consider these high abandonment rates in combination with high false rejection rates (FRR) within a system, the result to your bottom line will be significant.

If we assume you have 5,000,000 applications per year, a 63% abandonment rate with a 10% false rejection rate will result in approximately 30% of your applications (1,535,000) actually going through your system. If we assume each application brings in $50 of revenue, that’s an annual loss of over $173 million dollars to your organization.

By implementing simple, incremental enhancements to the user experience and system accuracy, you can achieve significantly more revenue

Providing users with a superior UX that reduces the application abandonment rate by just 5% to 58% and leveraging a more accurate system that reduces the false rejection rate from 10% to 5%, your business could process 420,000 more applications each year.

A recent study showed that

63% of financial services

applications are abandoned

when the process is

inadequate.

139452 86

26395 7579

83645 25

A P P L I C AT

I O N

13945286 263957579 8364525 5585806 26395

13945286 263957579 8364525 5585806 26395 3250938474

13945286 263957579 8364525 5585806 26395 3250938474

139 261 804 932 455 969

9

EBOOK

What could a lift in revenue look like?

Total Applications

Drop-off Due to Poor Experience

Remaining Applicants

False Rejections

Number of Lost Customers

Number of Remaining Customers

Revenue ($50/Application)

Lost Revenue ($50/Application)

Lift in Applications Processed

Lift in Revenue

5,000,000

3,150,000

1,850,000

315,000

3,465,000

1,535,000

$76,750,000

$173,250,000

5,000,000

2,900,000

2,100,000

145,000

3,045,000

1,955,000

$97,750,000

$152,250,000

↑420,000

↑21,000,000

63%

10%

58%

5%

Assume 63% Drop-off and 10% FRR

Assume 58% Drop-off and 5% FRR

A 5% decrease in abandonment and a 5% decrease in False Acceptance Rate (FAR) could result in an increase of $21M or 27% in annual revenue for your business.

13945286 263957579 8364525 5585806 26395 3250938474

13945286 263957579 8364525

13 9 4 52 8 6

26 39 57 57 9

8 36 4 52 5

55 8 58 0 6

26 39 5

32 50 9 38 4 74

13 94 52 86

26 39 57 57 9

83 64 52 5

55 85 80 6

26 39 5

32 50 93 84 74

10

EBOOK

Powerful image capture and machine learning forensics and data extraction with best-in-class accuracy

The best way to stop rejecting good customers (and to boost revenue while reducing customer acquisition costs) is by relying on a smartphone-based image capture app for document verification. This app should provide state-of-the-art features and real-time coaching — in verbal form for aiding accessibility — so it empowers users to snap well-lit, full- frame, high-resolution document photos on the first try.

Further, OCR data extraction must be combined with barcode and machine readable zone (MRZ) components along with robust machine learning- and computer vision-powered forensics and data integrity checks to attain the highest degree of coverage and accuracy. A powerful forensics and data extraction engine conducts dozens of validity checks on minute details, such as face and orientation detection, edge detection and cropping, and colorspace analysis to detect fake or tampered IDs.

Once relevant data is extracted, a correlation is made between the front and back of the ID along with the PII data provided by the user. For typos and cases of slightly mismatched input, robust ID verification solutions incorporate “fuzzy matching” to quickly correlate and resolve small differences; for example, an input nickname (“Bill”) with a formal name (“William”) on a government-issued ID.

Keep in mind that the human eye cannot decode a barcode; detecting information mismatches between encoded data and data printed on an ID is not practical without technology. With a rise in the use of AI to aid fraudsters in churning out indeterminable numbers of fake IDs, this is an important consideration.

By focusing on best-in-class image capture and data extraction associated with the ID verification process, you gain the highest level of accuracy and can almost entirely eliminate rejecting good users.

The solution

13945286

26395757 9

8364525

5585806

26395 32509384

7

11

EBOOK

Warning sign 3 Synthetic identities, scammers, counterfeiters, or underage users are infiltrating your ecosystem

AI technology is making it easier than ever for criminals to commit identity fraud. Other challenges arise when ill-intentioned users slip by ineffective controls at onboarding. These profiles show up in your ecosystem as synthetic identities, scammers, counterfeiters, or underage users.

The Department of Justice (DOJ) calls synthetic identity fraud the fastest growing financial crime of our time. Synthetic identities are created when fraudsters combine real and fake PII to form a feigned identity. Socure estimates that incidents of synthetic identity fraud will grow to $5 billion by 2024.

Consumers lost $8.8 billion in 2022 to imposter, investment, job-related, and other types of scams, according to the Federal Trade Commission. Incidents of investment scams — which includes increasing cases of cryptocurrency pig butchering, or tricking users into investing money in seemingly legitimate businesses on social media or dating websites -- more than doubled.

Federal, state, and local law enforcement note a growing trend of counterfeited products, involving automotive parts, electronics, prescription drugs, cosmetics, and other consumer products. eMarketer.com estimates that 42% of counterfeited products are distributed and sold through online marketplaces.

Further, underage users are slipping into social media sites, dating apps, cannabis delivery systems, and drinking establishments and reportedly comprise a healthy portion of total patrons. Alcohol purchases and consumption are prohibited in the U.S. under the age of 21. However, a number of states have enacted legislation or are moving toward limiting underage access to certain other categories of commodities and services, or banning use altogether. For establishments and services that have such restrictions, ignoring age verification requirements can result in license suspension, fines, or other punitive measures. Some businesses choose to think of these types of fraud as just a cost of doing business. But this doesn’t have to be the case; high fraud losses or bad-intention use cases are a broader warning sign that you’re ignoring critical security gaps.

High fraud losses or

bad-intention use cases

are a broader warning

sign that you’re ignoring

critical security gaps.

12

How this hurts your business By entrusting your onboarding identity checks to manual processes or subpar legacy systems, you may be unknowingly allowing fraudsters and dangerous or underage users into your ecosystem. This not only risks losses to your bottom line, but potentially subjects your organization to regulatory action and fines, attacks the name and value of your brand, and opens the door to financially or physically harming your existing customers, as well as other businesses and the general public.

For example, synthetic identity fraud has been called a “victimless” crime, but it is far from that. The victims include actual people — often children — whose Social Security numbers are stolen to create synthetic identities and who later learn that their credit is ruined. It also includes victims whose lives are devastated when public funds are stolen by synthetic identities that block legitimate claims or from crimes funded through synthetic fraud, including human trafficking and terrorist financing.

Pig butchering victims are losing their entire life savings to unscrupulous criminals who stalk victims online. And, physical harm has been inflicted by one user on another on numerous social media, marketplace, and dating sites, including instances of death. Allowing underage users into your ecosystem could result in damaging mediaexposure and out-of-compliance fines.

13945286 263957579 8364525 5585806 26395 3250938474

DDA

Synthetic identity fraud has

been called a “victimless”

crime, but it’s far from that.

13

EBOOK

Robust ID document verification with layered predictive fraud models

Robust document verification and selfie checks establish the authenticity of a government-issued ID and ensure the presented identity is the same person depicted on the ID. Yet, fraudsters can still slip detection due to unfilled security gaps.

To be sure it’s safe to allow a person into your ecosystem, work with a provider that supports flexible workflows. This should include incorporating layered predictive fraud models to determine the riskiness of the PII elements submitted by the user.

Proactively protecting the user experience while keeping bad actors out requires a defense-in-depth approach, including:

The solution

The best solutions rely on hundreds of offline and online data sources and incorporate real-world feedback data, along with outcomes from a database of known good and bad identities, to gain the highest assurance on the riskiness of the individual. When coupled with robust ID and biometric verification, you gain a complete view of your customer and a high confidence level in the person with whom you are doing business.

• Forensic data analysis using the rich data derived from machine readable components for fake ID detection.

• Fraud detection models including fake headshot detection, image alert list, selfie-to-ID matching, and age discrepancy check.

• Additional downstream data sources.

14

EBOOK

Fast, accurate ID verification at onboarding is critical to building digital trust and sets a high standard for the entire customer journey. For satisfying user experiences that bolster your bottom line with high conversion rates, while keeping out fraudsters, it’s important to know how to identify and assess best-in-class ID document verification solutions.

Here’s a checklist of the requirements:

ID document verification solution checklist

CH EC K L I S T

▶ Consider which ID document use cases you need to support related to your current and future product sets, and ensure the document verification solution can handle those transactions in a fully-automated fashion.

▶ A vendor should be investing in a robust machine learning- and computer vision-powered forensics and data extraction engine with the capability to conduct dozens of document validity checks.

▶ Opt for a solution that provides state-of-the-art image capture technology with real-time coaching and accessibility features.

▶ Assess which fraud models, explicit and implicit risk signals, and workflows you need and select a vendor that can orchestrate them all.

▶ Machine learning algorithms must be trained on vast, diverse datasets which include offline and online sources, real-world feedback data, and a database of known good and bad identities.

▶ Run a proof of concept test with your preferred solution provider to assess performance + operational improvements.

▶ Use a selection process that optimizes value, where cost consideration not only includes dollars spent on a given solution, but also the cost of compliance fines, lost business due to poor conversion, and the dollar amount of fraud losses.

▶ Select a vendor committed to ongoing user experience, accuracy, and speed improvements.

15

DocV uniquely leverages

predictive fraud models from

the Socure ID+ comprehensive

identity graph, including rich

device, phone, behavioral,

address, + geolocation signals.

EBOOK

It enables increased customer acquisition while stamping out online identity fraud by marrying an unmatched forensics

engine and data extraction capabilities with a frictionless, front-end image capture app that provides an automated

user experience. Additionally, DocV uniquely leverages predictive fraud models from the Socure ID+ comprehensive

identity graph, including rich device, phone, behavioral, address, and geolocation signals, to provide risk decisions on

the individual connected to the physical credential. With a response time under 2 seconds and true accept rates of 93%

compared to the industry norm of just 64%, Predictive DocV sets the highest industry standards for speed, accuracy, fraud

reduction, and user experience, while accelerating expansion into new markets and geographies.

Socure’s Predictive Document Verification (DocV) 3.0 satisfies all these requirements and more

? ?

13945286 263957579 8364525 5585806 26395 3250938474

EBOOK

See how Socure can help you accelerate digital identity trust.

About Socure Socure is the leading platform for digital identity verification and trust. Its predictive analytics platform applies artificial intelligence and machine learning techniques with trusted online/offline data intelligence from physical government issued documents as well as email, phone, address, IP, device, velocity, date of birth, SSN, and the broader internet to verify identities in real time. The company has more than 1,800 customers across the financial services, government, gaming, healthcare,telecom, and e-commerce industries, including four of the top five banks, 13 of the top 15 card issuers, the top three MSBs, the top payroll provider, the top credit bureau, the top online gaming operator, the top Buy Now, Pay Later (BNPL) providers, and over 250 of the largest fintechs. Marquee customers include Chime, SoFi, Robinhood, Gusto, Public, Stash, DraftKings, State of California, and Florida’s Homeowner Assistance Fund. Socure customers have become investors in the company including Citi Ventures, Wells Fargo Strategic Capital, Capital One Ventures, MVB Bank, and Synchrony. Additional investors include Accel, T. Rowe Price, Bain Capital Ventures, Tiger Global, Commerce Ventures, Scale Venture Partners, Sorenson, Flint Capital, Two Sigma Ventures, and others. 16 EBK_Doc V-3 Warning Signs 5125077696 © 2023 Socure, inc. All rights reserved.

LEARN MORE

https://www.socure.com/products/document-verification

2 Introduction 5 Warning Sign 1: Applicants are waiting minutes, hours, or days to be approved 7 Warning Sign 2: Large numbers of legitimate customers are being falsely rejected 11 Warning Sign 3: Synthetic identities, scammers, cou