Case Studies

Large eCommerce Retailer Safeguards their Call Center

On track to save company nearly $10M in fraud losses

The Challenge

This retailer is one of the world’s largest brick-and-mortar organizations with one of the fastest growing and most dynamic e-commerce organizations. With a large portion of the U.S. population living in close proximity to its stores, the retailer is primed to combine physical locations with e-commerce business to offer a level of convenience never before seen. The retailer has also continued to pursue ways to bring technology into retail to offer customers a seamless shopping experience across its stores and e-commerce websites.

The rise and prevalence of social engineering in the e-commerce and retail industry was a major concern. LexisNexis® Risk Solutions study1 found that U.S. e-commerce merchants have been hit with the highest fraud costs, with nearly half(47%) related to replacing/redistributing lost goods2. Pindrop® Voice Intelligence and Security Report, 2023 also reinforced this threat. The report states that retail has a fraud rate of 1 in 99 calls, which is seven times higher than the next highest industry (banking) at 1 in 749 calls. Retail is vulnerable to various fraud risks, including return fraud, e-commerce fraud, and debit/ credit card fraud.

The Solution

Utilizing the Negative VoiceID solution at this retailer, Pindrop has identified several prolific and repeat fraudsters who were using multiple caller IDs to attack the call center. For example, a fraudster “Salao Khan” made 5 separate phone calls from 5 different ANIs within 3 weeks using 4 different names. “Salao Khan” was trying to socially engineer the agent into processing fraudulent returns for missing items. By matching this fraudster’s voiceID across the phone calls, Pindrop produced a Negative VoiceID which allowed the retailer to stop the fraudster from succeeding with the transaction and further investigate the velocity of accounts that the fraudster had already nested (see table on next page for additional examples.

In just the first few weeks of utilizing the new feature, Pindrop has identified 86 distinct
fraudsters that have placed 8,906 calls from 6,049 unique devices ANIs, putting the retailer in a position to close 1000’s of accounts and prevent a significant amount of fraud losses associated with the orders placed by these accounts3.

For e-commerce companies, an astonishing 14% of returns have been found to be fraudulent4. Estimates from the National Retailers Federation reveal that e-commerce return fraud may cost US retailers $23 billion annually5. The retailer was specifically concerned about this trend of bad actors extracting fraudulent concessions (refunds, replacements, adjustments, promos, gift cards) by deploying social engineering tactics in the call centers.

The situation before Pindrop

The retailer faced a growing volume of fraud in the contact center. 1 in every 60 calls that came into the contact center were fraudulent. Many of the fraudsters utilized a “Concessions Abuse as a Service” model that targets large e-commerce merchants with robust customer service who are more willing to risk a financial loss in exchange for customer satisfaction. The fraudsters were targeting this particular retailer due to its large e-commerce volume and customer-friendly practices and were going after accounts that closely resembled those of loyal customers. As a result, the retailer faced the challenge of losing millions of dollars and putting the brand at risk.

How does Pindrop do it?

To help the retailer solve this fraud problem, Pindrop utilizes a multi-factor platform that
includes the voice, device, behavior, and several data risk models of the caller that produce a risk score on every call in real-time. Producing a multi-factor risk score in real-time to the retailer allows them to route calls from the IVR to a group of specific contact center agents that they want to service that call. The agents have all the tooling and trained skills necessary to review all of the account information needed to accurately disposition the calls as fraudulent.

In addition to the multi-factor risk platform, Pindrop developed and operationalized a new feature called “Negative VoiceID” to identify when a single fraud voice is attacking multiple times over a given period of time. Pindrop Protect continues to utilize voice analysis and matching, comparing the voice features of the current caller to a set of negative voice profiles trained based on confirmed fraud cases which influences the risk score associated with the current call. However, the Negative Voice ID match is a risk factor that is useful for identifying a fraudster when tactics are used to change or mask the calling device, location, or phone number, but the same fraudster is speaking.

What’s next

The retailer wants to use Negative VoiceID not just to identify and investigate fraud incidents but to also block these accounts from being opened and utilized for fraud. In addition to identifying these fraudulent accounts and the volumes of fraudulent returns that are processed, the retailer also plans to block these calls from making it to any customer care agents so as to eliminate or significantly reduce the call center agent handle time.

The retailer is planning to expand Protect and the Negative VoiceID feature to other lines of business and geographical locations to protect their business from these prolific fraudsters and to enhance genuine customer experience which will allow the retailer to achieve their targeted goal of reducing fraud losses by 3x.

Sources Cited

1. https://risk.lexisnexis.com/insights-resources/research/us-ca-true-cost-of-fraud-study

2. https://orbograph.com/lexisnexis-report-cost-of-fraud-is-on-the-rise-since-the-pandemic-set-in/
3. Pindrop analysis of fraud calls at large US retailers for 4 weeks in 2023
4. ECR Community Buy Online, Return in Store, The Challenges and Opportunities of Product Returns in a Multichannel Environment, 2019
5. National Retailers Federation survey on cost of retail returns, 2022
6. Results from a pilot study conducted by top US retailer comparing Pindrop’s fraud detection rates against two other vendors 
7. Pindrop Voice Intelligence and Security Report, 2023
8. Forecasted annualized savings based on Pindrop’s fraud detection rate applied to total call volume and average industry fraud losses

Voice security is
not a luxury—it’s
a necessity

Take the first step toward a safer, more secure future
for your business.