When thinking about fraud detection, there are a lot of options out there. From geolocation to IP address verification, there are a wide range of solutions with various levels of effectiveness, depending on the channel (mobile or online).
Sometimes it’s nice to have multiple defenses against fraud. It makes sense to want to double-check the results of solutions that aren’t providing a clear picture.
This is often done with manual reviews, which 81% of merchants perform, yet these aren’t foolproof. They are prone to human error, high false positives, and even major fraud losses.
Your anti-fraud program is only as good as the method you use to fight fraud. The better response (beyond manual reviews) to fraud detection that isn’t working is to re-examine the method used to detect fraud.
Performing manual reviews with a faulty fraud detection method is like putting a BandAid (an expensive one, at that) on the problem. The better solution is to attack the root of the problem, in this case, the way transactions are screened.
What are Manual Reviews?
You may or may not have heard of manual reviews used to detect fraud. These are reviews done by real humans instead of computers.
According to a report by Cybersource, around 1 out of 4 orders is manually reviewed. This proportion is larger for small businesses (2 out of 5) and smaller for large businesses (1 out of 10).
Manual reviews involve a team of analysts that check flagged transactions to determine whether they are a real threat or not. If a reviewer says a transaction is a threat, the order is not fulfilled. If it’s not deemed a threat, the order is allowed to continue in the process.
Around 2.3% of orders that are manually reviewed are rejected. This percentage is higher (4.5%) for businesses with ecommerce yearly earnings between $5 to $25 million. This number includes fraudsters and legitimate customers who are not able to make purchases.
There are different ways businesses choose to use manual reviews. Reviews can be done before an order is fulfilled or after a decision has been made to accept or reject an order. In this way, they can either check every order before it is completed, or only the ones that have been picked up by a detection method.
Manual reviews can be used in conjunction with other fraud fighting strategies to help judge the effectiveness of a fraud detection method by its ability to stop bad transactions and let good ones through.
They can also be used to make decisions about transactions that aren’t blatantly risky or blatantly safe. In situations that are less clear-cut, businesses often like to have an actual pair of eyes to assess the details.
The idea behind manual reviews is to cut down on the number of fraudulent transactions by allowing personnel to look at the information available and make an informed decision on how likely it is fraudulent.
Training and Maintaining Personnel
To understand the problem with manual reviews, it’s important to consider the amount of time it would take to review every transaction, or even just transactions considered “risky.” The average time required to review a single order is 5 minutes, with smaller merchants requiring longer (10 minutes).
This first requires having employees ready to jump in and perform the reviews. The manual review process is only as good as the personnel who are trained to do the reviews. If personnel are not trained well, it’s extremely likely significant money will be lost by either failing to deter fraudsters or deterring real customers.
Businesses that employee personnel to do manual reviews must invest in intense training. Training the personnel to manually review transactions is time consuming and expensive. After training, employing personnel in-house to look at every flagged transaction is tedious. Keeping employees well-versed on the ins and outs of a complex and ever-changing fraud landscape is also costly and time consuming. Altogether, these result in the average merchant spending 52% of their fraud management budget on manual review staff.
Manual reviews can also make accepting customer orders extremely slow. They can increase the time between a submitted order and that order being fulfilled. This has the potential to result in customers frustrated by the delay, especially if your business involves “instant” products like digital downloads. In turn, this can result in lost revenue from customers who want a product and want it quick.
When considering the time and monetary cost of manual reviews, it’s also important to realize employees could be spending their time and salary growing the business instead of checking transactions.
Manual Review= High False Positives
By their very definition, manual reviews are at the mercy of human error. Allowing the final decision about a transaction to rest on a reviewer means it’s all about a person’s judgement, which can be inherently faulty. This judgement is entirely dependent on how well an individual is trained and what guidelines they are given about transactions, which vary depending on the business.
The average false positive rate for manual reviews is 10%. This is makes up a significant amount of revenue lost.
This rate often depends on how personnel are trained. If employees are told to reject every transaction above a certain risk threshold, or if it is cheaper to lose a sale than have a fraudulent transaction, there will be high rates of false positives.
This means completely normal customers just looking to make a purchase will be turned away from your business. A false positive doesn’t only affect the sale in question. Customers that have been rejected from a sale are unlikely to come back to your business. They’d rather go to your competitor and get what they’re trying to purchase.
When the choice of whether or not to reject a transaction on the basis of suspicion is left to personnel, there is a lot of potential for error. Manual reviews can result in high false positives, high cost, and excessive time lost that could be spent expanding the business instead of fretting over fraud.
In a fraud detection strategy, manual reviews should be the last line of fraud defense. Checking every transaction is extremely difficult for businesses to invest in. The results of these checks are also unpredictable.
Often, reducing manual reviews by employees means investing in a fraud detection that works. If businesses are investing excessively in manual reviews, it’s time to see if there’s a better solution out there.
Newer fraud detection solutions involving behavioral biometrics allow businesses large and small to cut down on the amount of time spent manually reviewing transactions. This can reduce cost and false positives, and get businesses back to focusing on, well, business. Sounds good, right?