Jumio

Their "maybes" are costing you money. It's time for "yes" or "no" answers.

Online Identity Verification:
The Growth of Automation & Indecision

The growth of the online identity verification market has introduced a number of automated solutions that rely on optical character recognition (OCR) and basic machine learning to determine if an ID is authentic or the consumer’s digital identity is legitimate.

But, these automated solutions have a number of shortcomings:

  • Poor verification accuracy (i.e., too many false positives)
  • Return too many “suspect” decisions (aka “maybes”)
  • More maybes lead to more manual reviews
  • More manual reviews lead to more frustration and abandonment

Translation: Automated solutions cause more legitimate customers to be erroneously denied, triggering greater frustration, customer dissatisfaction and online abandonment.

Jumio

What’s the Harm in Maybes?

Jumio

A “caution” is the identity vendor’s admission that they don’t have a clue as to whether the customer is who they claim to be. This means that the solution provider could not verify the government-issued ID (e.g., the passport, driver’s license, etc.) or the identity by comparing the selfie to the picture on the ID document of the new account applicant.

Every “caution” or “maybe” requires manual review by a team of analysts to determine whether they are a real threat or not. But, there are real costs to manual review.

Real costs to your business. Real frustration by your users. Real costs to your brand.

Do the Math

Obviously, the more manual reviews, the more staffing required to review those flagged transactions. But, far more damaging than the incremental employee costs, is the opportunity costs of losing legitimate customers who bail on the online application process if they’re subject to long delays. That’s why we’ve developed this online calculator so you can calculate the dramatic impact of increasing the number of manual reviews.

The math is undeniable.

Digital businesses need to minimize the number of manual reviews and try to convert as many customers online as possible (while keeping out the fraudsters). The best way to reduce the number of “maybes” is to shift the decisioning burden back to your identity verification vendor and ensure they deliver a “yes” or “no” decision. No more maybes!

The Cost of Manual Review Calculator

Planning Assumptions

Current number of online identity verifications performed each day
Percent of manual reviews as a result of verification "maybes"
Incremental number of daily manual reviews caused by "maybes"
30
Average time spent on each manual review (in minutes)
Total time required for incremental reviews (in hours per day)
5

Incremental Staffing Costs

Average salary of manual reviewer
Incremental staffing cost for incremental manual review (per year)
$28,125

Estimated Opportunity Costs (from customer abandonment):

Estimated abandonment rate
Estimated customer's lifetime value
Number of customers who abandon each year
780
Estimated annual lost revenue from abandonment
$780,000
Total estimated cost of manual review per year
$808,125

Measuring Quality

If you’re going to hoist more of the responsibility on the verification vendor, then you need to have confidence that they’re delivering the right answer. That is, the identity verification decision returned is accurate and defensible. But, how do you measure accuracy? It really comes down to two key quality metrics:

Ability to identify fraudulent ID documents or identities

Ability to convert legitimate customers

So, how do you know if your identity verification solution is getting it right?

Jumio recommends performing a controlled test to ensure that the right decisions are being made and that the quality of the ID and identity verifications is thoroughly vetted. It’s surprising how few companies actually perform this step and live with the consequences of poor quality. It’s increasingly vital to compare—not only from a verification accuracy perspective, but also to see how many “maybes” get returned by the participating vendors.

The Ingredients of Quality

Achieving verification accuracy is no mystery. It’s the byproduct of a number of technologies, processes and experience. Look for solutions with:

Jumio

Definitive Answers

Don’t settle for “maybes,” risk scores and “cautions”—get a definitive “yes” or “no” that tells whether they should accept or reject an individual identity verification.

Jumio

User Course Correction

Better verification solutions will return additional information about why an ID or selfie was rejected so users can try again.

Jumio

Face-matching Accuracy

Image quality and a person’s appearance change over time, so test and compare the quality of the face-matching algorithm to ensure your vendor is getting it right.

Jumio

Machine Learning

Machine learning at scale can play a large role in reducing manual reviews and making identity verification more effective.

Jumio

Data Tagging

Data tagging helps train machine learning algorithms since humans can quickly determine if an ID document is scuffed, hole punched or captured in bad lighting, and still render a decision.

Jumio

Auditing

Look for solutions that audit 10-20 percent of transactions to ensure that their verification engine is correctly flagging fraudulent IDs and approving your good customers in a timely manner.

Jumio

Liveness Detection

Better identity solutions leverage biometric facial recognition and liveness detection to combat fraudsters who attempt to fake their identity with static images.

Jumio

Geographic Coverage

Can your chosen provider support the countries, languages and ID types of your user community?

Jumio

Geographic Coverage

Can your chosen provider support the countries, languages and ID types of your user community?

Jumio

ID Version Support

In some states, there may be as many as 15 versions of a particular driver’s license, each with their own unique set of security features.

Jumio

Omnichannel Support

Just supporting mobile (SDK) is table stakes in today’s identity verification market. Better solutions support mobile (SDK and web), desktop, API-based and native mobile implementations.

Jumio

Blur & Glare Detection

Not all cameras are created equal. If you allow webcams for image capture, you need a way to alert users when an image isn’t clear so they can retake the picture.

Jumio

Blur & Glare Detection

Not all cameras are created equal. If you allow webcams for image capture, you need a way to alert users when an image isn’t clear so they can retake the picture.

Jumio

Blur & Glare Detection

Not all cameras are created equal. If you allow webcams for image capture, you need a way to alert users when an image isn’t clear so they can retake the picture.

Jumio

Black & White Detection

All legitimate IDs are issued in color, so it’s important to be able to detect black and white images.

Jumio

Address Verification

Online document verification allows businesses to quickly extract data from supporting documents captured by smartphone, providing an extra layer of identity assurance.

We guarantee our quality. And so much more.

At Jumio, we stand by our word and our quality. Jumio has invested in the technologies, processes, and people to deliver the highest verification accuracy in the industry. And it’s this confidence that emboldened us to offer this challenge.

Jumio

If you perform a competitive test and the other online verification solution is able to deliver more accurate results than those produced by Jumio’s Netverify, we will pay you $10,000. Terms and conditions apply, please see the Official Rules for details. NOTE: Your organization is still eligible for the $10,000 reward even if you decide to sign with another identity solution provider.

Jumio