Claims fraud in travel - another form of chargeback

It’s a thought that has probably occurred to some of us at one point: why not buy something we need, use it, and return it? Or, better yet, order a product online and say that it never arrived, allowing the customer to keep the money and the product. Many times, this thought will transcend into action, known as claims fraud, that is costing merchants anywhere between $2.1 to $4.2 billion a year. While this issue is the worst it’s ever been, it’s a problem that goes back decades. In fact, it’s so prevalent that it predates the internet, stretching back into the days of mail order catalogs.

It is estimated that 10% of all order claims are fraudulent, but that doesn’t mean that the damage they cause is limited to that 10%. Because companies don’t know which claims are fraudulent, they have to suspect that all claims could be fraudulent, and that means treating all customers like potential fraudsters. 

The reputational damage that this causes cannot be understated. Alternatively, it means giving up completely on finding the 10% and adopting a “the customer is always right” approach. This is largely what Amazon has done, and because of their success, many companies have adopted the same approach.

The travel industry is hit particularly hard by chargebacks. According to data from our partner Amadeus, chargebacks are up by 50% since pre-pandemic levels. It is estimated that 2-8% of all travelers miss their flights, and a percentage of these people will choose to try to claim money back through the chargeback system. 

For this reason and the general high level of fraud against the industry, airlines, travel agencies and other parts of the sector have to pay extra on card processing to factor in claims fraud and need to devote time, money and manpower to reducing fraud as much as possible. Their customers will also have to experience greater difficulty during the payments process, increasing friction in what can already be a difficult process.

There should be a better way to manage claims fraud, one that doesn’t put undue pressure on legitimate customers, but also doesn’t allow anyone to make a fraudulent claim unchallenged. 

How claims fraud works

The term “claims fraud” is typically used in the insurance industry for an instance of a customer making a fraudulent insurance claim. The term can be used more widely and for a greater range of claims, including those made against eCommerce and travel companies.

Claims fraud is very simple: a customer orders a product, then contacts a company to say that it didn’t arrive or arrived broken. They will refund the customer’s money, leaving the consumer with both the product and the money that they paid for it. Although it is unlikely to be carried out by large criminal organizations, claims fraud is very common and easy to carry out. Like all types of fraud, it is likely to become more prevalent during a financial downturn.

We can see it as an opportunistic crime rather than something that is part of a larger pattern, which makes it difficult to spot using very basic machine learning. It is easy to spot when a customer has a track record of claiming that an item didn’t arrive and asking for a refund, but claims fraud is typically an indulgence rather than a habit.

A person may engage in it because money is tight or because they have a large expense that they couldn’t cover normally. They may even be completely casual about it, making claims on items that they could easily afford and rolling the dice on whether they are to ‘double their money’ by taking a few minutes to make a claim – similar to a 100% mail-in rebate. In this sense, it is very much like a chargeback.

The key difference between claims fraud and chargeback fraud is that merchants will be handling 100% of the refund process rather than working with card schemes as they do with chargebacks. Unless they are operating on an entirely “the customer is always right” policy, a merchant will have to contact their logistics company and request proof of delivery, then return this proof of delivery to the customer. If the customer digs their heels in, then you may be going back and forth between them and the logistics company for days or weeks. If they weren’t lying, then you will have done irreparable damage to your relationship with that customer.

Countering claims fraud

As mentioned above, merchants have traditionally had two options when countering claims fraud. The first is an “active” approach, in which claims have to be investigated, usually by a dedicated staff member or an entire team depending on the size of the company, and this can end with the company losing the claim after devoting time to contesting it or winning and ruining their reputation with their customer. 

The second method is a “passive” approach that involves turning a blind eye to the possibility of fraudulent claims, since any given claim is 90% likely to be legitimate. Few, if any, companies operate a truly passive approach, and even Amazon, who made the approach widespread in e-commerce, will dispute some claims. It’s also not possible for every company to practice this method: a company selling computer parts or high-end fashion can’t write off orders that can cost $1,000 or more, even if it does cause reputational damage.

There is a third way, however, one that allows companies to maintain their relationships while reducing the total number of false claims.

AI and machine learning is already used to find the underlying signals of fraud in chargebacks, which claims fraud is functionally very similar, and it can be easily adapted to the latest fraud trends. Automated systems can also be tied into an AI/ML system – for example, requesting images of deliveries from logistics companies automatically with no human intervention required. This can significantly reduce the amount of work needed to make sure that claims are legitimate, with the added bonus of the vast majority of claimants not being affected in any way by the process. Those that are affected will be selected according to a constantly-improving set of criteria and that selection will happen in real time.

For the travel industry, this means not only a reduction in friction for customers, but an overall reduction in fraud that could lead to the point where the industry is no longer considered to be “risky” and the premium that companies must pay can be reduced. It will require significant investment from the industry, but the rewards more than justify an initial outlay.

Using modern technology, travel merchants can significantly reduce the number of fraudulent claims they receive while maintaining the customer relationships that their businesses are built on.