Part of a control environment in any operation with easily moved goods is loss prevention. Amazon warehouses are full of things people want and Amazon has loss prevention roles in the company. From August 2020 to March 2022 Kayricka Wortham was an Operations Manager at an Amazon warehouse in Smyrna, Georgia. In her position, Wortham could approve new vendors and approve vendor payments.
Demetrius Hines was a Loss Prevention Multi-Site Lead at Amazon. He also worked at the Smyrna warehouse and at other company sites. In his position, Hines was:
Wortham led this scheme and would provide fraudulent vendor setup information to her subordinates and then approve the vendor setup. She and others, including Hines, submitted fake invoices to for payment by Amazon in amounts exceeding $10 million. They received payments totaling $9.4 million.
Hines was a roadblock and Wortham overcame this roadblock by recruiting him into the scheme. The typical wisdom is that collusion is really hard to find and stop. This is a great example of why. A key control, Hines, was coopted into the scheme. But they were still caught.
There were also red flags, primarily people living beyond their means.. Wortham was driving a $200k+ Lamborghini Urus and Hines had multiple cars, a $70k Rolex and diamond jewelry. Also an ops manager might have setup info about the occasional vendor, but it seems like this was a large number of vendors.
There are a couple of interesting pieces in here. First, $10 million is a lot to steal in 18 months. Kevin Lee Co stole $4.8 million over 7 years from Holt. Nathan Mueller stole $8 million in 4 years. $10 million in 18 months is a lot.
Second there is no information across multiple articles on how the fraudsters were caught, just boilerplate language about the Secret Service being involved. I’m speculating that a $10 million expense increase over 18 months triggered a deeper look and ultimately a criminal investigation.
How do you catch fraud like this? Collusion was involved to override preventive controls so it gets harder and we many need to fall back on detective controls and analytics. For example, budget analysis, trend analysis, Benford analysis on invoice/payment amounts, and anomaly detection could all be part of the process.