For better and worse, we humans are not driven strictly by logic and self-serving interests. While decisions based on empathy, social norms and love allow us to have a functioning society, they often fail us when it comes to managing a business. In business, it’s important that we try to overcome psychological biases and instead strive to make the best decisions based on data and facts.
The work of behavioral economists is devoted, among other things, to identifying and explaining the psychological biases that lead us to make suboptimal decisions. In this post, I’ll introduce a couple of well known behavioral economics terms and will illustrate their application in the field of eCommerce fraud management.
What Is Loss Aversion and How Does It Affect Your Decisions?
A psychological bias first demonstrated by Nobel laureates Daniel Kahneman and Amos Tversky, Loss Aversion refers to individuals’ strong preference to avoiding losses over acquiring gains. A nice example of loss aversion and its implications in the real world is a field experiment conducted by Colin Camerer, a professor of Behavioral Finance and Economics at the California Institute of Technology.
In the experiment, Camerer examined the motivation of NYC taxi cab drivers. With highly volatile and unpredictable daily revenues, drivers set themselves a daily target. Camerer found that on bad days, when rides were scarce and drivers’ average hourly income decreased, drivers worked longer hours in order to reach their daily revenue target. On good days, however, when hourly wages were relatively high, the drivers stopped working early, once they had hit the target. The cab drivers stopped working even though they could potentially make much more money on these days.
The most logical, cost-effective behavior would be to work less hours on poor days and longer hours on the good days. But this is where loss aversion comes in. The drivers were more afraid of losing money than they were excited about the prospect of making additional money. In other words, the psychological impact of a dollar lost is greater than the impact of a dollar earned.
How Loss Aversion Applies to ECommerce Fraud Management
In the world of fraud management, I often see how loss aversion comes into play when executives try to calculate the cost of fraud to their organization. Retailers tend to focus on chargeback rates, and pay little if any attention to other costs associated with fraud management – such as lost sales revenue and poor customer experience.
The logical way to assess the issue would be to calculate the losses but also measure the potential gains. However, although the value of declined sales often outweighs the chargeback-related losses, retailers underestimate the potential gains from approving more orders as those are less palpable than actual losses suffered in the past. Instead, the focus remains on the cost of chargebacks as these are dollars the company made and subsequently lost (and for which fraud managers are considered responsible).
For example, I recently spoke with a well known electronics retailer based in North America. This retailer has several brick and mortar stores as well as a fast growing online channel. When I first reached out, the retailer was reluctant to evaluate the potential benefit of Riskified’ service, stating that they were content with their current situation, as their chargeback rate was low, at 0.2%. While I agreed that this was an impressive number, my interest was in understanding the “price” at which this low chargeback rate was achieved.
When the retailer actually dug in and looked at metrics other than chargebacks it became clear how they were keeping chargebacks to a minimum. The merchant was declining a whopping 11% of all incoming orders at the gateway level (with AVS filters and other rules), and their fraud management team went on to decline another 2.6% of all orders (14% of orders routed to manual review were rejected). In other words, this electronics retailer was so focused on maintaining low chargeback rates that it had adopted an incredibly risk-averse fraud policy – that is most probably resulting in plenty of legitimate orders being declined along with the fraudulent transactions.
What Is the Endowment Effect?
The Endowment Effect, a hypothesis commonly attributed to world renowned Professor of Behavioral Science and Economics Richard Thaler, is closely tied to loss aversion. The hypothesis suggests that individuals attribute higher value to certain items solely because these items are in their possession. Most real-world implications of the endowment effect involve identifying the different pricing points people allocate to specific items, and examining if and how this price fluctuates based on ownership.
The endowment effect was widely researched through field experiments, many of which are quite cool, even to those of us who are not economists. One of the most interesting and well known experiments was conducted by Professor Dan Ariely of Duke University. In the experiment, Ariely examined students’ willingness to pay for tickets to Duke’s men NCAA basketball team’s final four game. The university allocated a certain amount of tickets at a discounted price to be sold to Duke students, yet as demand heavily outweighed the supply of tickets, potential buyers had to queue in front of the box office for days in order to score a ticket to the game.
Once all the tickets allocated to students were sold, Ariely approached the disappointed students who didn’t manage to purchase a ticket and asked them how much they were willing to pay to buy a ticket off those who had successfully obtained a ticket. He then contacted the lucky students who had bought the student-priced tickets and asked them to name the lowest price at which they were willing to sell their tickets.
Amazingly, Ariely could not match a single buyer and seller pair to complete a transaction. The highest price offered to buy a ticket was no where near the lowest sum for which ticket holders were willing to sell. On average, buyers offered to pay $170 while sellers asked to receive $1,400 per ticket. Given the fact that earlier that same day both the sellers and the buyers were equally enthusiastic about the game, and were prepared to spend the same amount of money to buy tickets, these results were astounding. What had changed? Ownership. Simply possessing the tickets spiked their perceived value by more than 800% in the mind of the ticket holders!
How the Endowment Effect Applies to CNP Fraud Management?
The endowment effect is not limited to academic research. In fact, I encounter it on a daily basis when speaking with eCommerce executives. I find that merchants with an established fraud management process and team are less willing to explore the true cost of fraud for their organization, while merchants who don’t yet have an existing process are more open to exploring and adopting new technologies and processes.
Case in point, the other day I was speaking with a sports equipment eCommerce merchant that sells mostly to the US market. The retailer carries high quality brands and is enjoying impressive year-on-year growth. A couple of years ago, when the store was first hit with significant fraud, the e-tailer implemented a tool that helped keep their chargeback rate under control. The executives with whom I spoke were happy with their status and considered themselves to be ‘in a good place’ when it came to managing online fraud.
However, when I ask for specific numbers it became clear that the situation is far from ideal. Eighteen months into implementing this fraud prevention solution, they had a chargeback rate of 0.55% and were declining 3% of all incoming orders. While this is not an awful situation, by rejecting 3 of every 100 orders placed on the site, this merchant is most probably leaving a lot of money on the table.
Naturally, I don’t expect anyone to simply get rid of an existing process whenever a new opportunity presents itself, but I do expect them to know the true costs of their fraud management operations and to have the ability to assess the ROI of implementing alternative solutions.
In business in general, and specifically in fraud management, psychological bias can hold us back and prevent us from achieving optimal results. The first step to overcoming any problem is awareness, and in that sense I hope this article has been helpful. Online retailers looking to boost their performance must try to conduct a real ROI analysis and ask themselves the following questions – “How much does my current fraud prevention process “cost” me, and is there a more efficient alternative?”, and “How much money do I stand to make by adopting a better process or solution and how can I ensure I actually gain?”.
In partnership with the MRC, a non-profit that helps merchant combat fraud, Riskified created a free and interactive calculator that helps provide a full picture of the cost of fraud in your organization. The tool is still in beta – but if you’re interested to try it out – let me know!