Skip to content
Fast Company Executive Board

The Fast Company Executive Board is a private, fee-based network of influential leaders, experts, executives, and entrepreneurs who share their insights with our audience.

“Bot or not” no longer cuts it: New online fraud threats require advanced solutions

Businesses are playing into the hands of cybercriminals if they don’t upgrade their deployments to meet the ever-more sophisticated design of attacks.

Computer hacker

[Images: mohamed/Adobe Stock]

Profile photo of Benjamin Fabre

BY Benjamin Fabre

Your business has just been attacked by cybercriminals, but the usual tools failed to prevent the attack, resulting in significant reputational damage. What was responsible? A simple bot? An AI-powered bot? A human-powered bot? Or was it an actual human manipulating account and payment endpoints? Identifying the type of bot in an attack was once the litmus test for cybersecurity providers when tracing the sources of online fraud. However, that approach is now useless against the increasing sophistication of attacks. 

The stakes are extraordinarily high. In the first half of 2022, Americans lost a record $3.56 billion to online fraud, and the Federal Trade Commission received 800,000 fraud complaints, with 27% of cases incurring a financial loss. These numbers may only harden the resolve of the 87% of consumers who are unwilling to do business with a company if they have concerns about its security measures. 

There is no such thing as a single fraud “recipe.” Hackers are now using a complex set of dedicated tools to design their attacks, and a robust defense starts with anticipating the intention of the visitor—not relying on bot identification. 

THE RED FLAGS OF FRAUD

Some of the red flags of being targeted by fraudsters include high payment declines, failed logins, order cancellations, and bounce rates, as well as server overload. Poor conversion rates that otherwise make little sense are another warning sign that bots are adding items to sales carts to test stolen credit card information.

The statistics are alarming. In a 2022 analysis of 5,200 confirmed breaches globally, Verizon found that over 80% were due to stolen credentials. Remarkably, despite 91% of people being aware of the risks of reusing passwords across accounts, 59% continued to do so anyway. 

It’s not only fraud that is costing businesses, but fraud mitigation itself is a massive expenditure. Yet 30% to 70% of all transactions flagged as fraudulent are actually false positives (when a human request is identified to have come from a bot). Even if a business chooses to accept the transaction, the cost of the mitigation product will outweigh whatever money is received. 

Chargebacks can also run into the hundreds of thousands of dollars per day, so merchants cannot afford to be fatalistic about the inevitability of fraud or oblivious to the real costs of false positives. With up to half of all e-commerce transactions coming from repeat customers, businesses are putting long-term revenue at risk if they remain unprotected.

THE HACKER’S TOOLBOX

Online fraud is the number one concern for IT and security professionals, but who is behind these attacks? It is imperative to understand that every single attack is a work of art designed to exploit very specific vulnerabilities depending on four main constraints: scope, complexity, cost, and ROI.

Scope refers to what is being targeted, whether a website or mobile app, a single organization, a specific country, or an international attack—as well as the timeframe for the fraud. Complexity, however, is a question of targeting large volumes of data or making a surgical hit. In the case of an attack on a specific website, hackers may need to penetrate any protective layers and interact with the site (through filling out forms or engaging in a chat).

Fraudsters evaluate the cost-benefit ratio of an attack depending on whether open-source libraries and access to SAS-based analytics technologies will be required, as well as if residential proxies (which cost significantly more than regular IPs) will be used. 

To maximize their return on investment, they will also determine whether a hit will be large volume but low value, or low volume and high value, using a set of dedicated tools to build the “perfect” attack, including:

  • Cost-efficient basic bots that scrape large volumes of information 
  • AI-powered bots that perfectly mimic human interactions for credential stuffing and scalping
  • Human-powered bots that are used for highly sophisticated attacks, account takeovers, checkout abuse, and “carding”
  • Humans being used as “sweatshop” labor to manually solve CAPTCHAs and to create fake fingerprints of their IP addresses

THE CHANNELS OF ATTACK

Anticipating the intent of an attack begins with determining if the “visitor” has the right to be doing what they are doing. Are they planning to take over your customers’ accounts or go through payment funnels with stolen cards? Are they likely to create fake accounts? These are the immediate parameters you must work within to get ahead of the hackers, regardless of the tactics or channel they are planning to use. 

However, active manipulation by humans represents a loophole in many cybersecurity systems. If you have people creating fake accounts, adopting multiple identities, repeatedly canceling orders, or using payment options from many different countries, businesses must be able to identify them. 

When it comes to bot-detection systems, it’s challenging to measure the real false-positive and false-negative rates (when a bot is identified as a human) in a live environment. With this in mind, it’s important for businesses to use an external company to conduct a pentest to target their main APIs/endpoints. They can also use a bot service to conduct a test themselves—and install another bot vendor to observe detection discrepancies.

PROTECTING YOUR DATA AND REPUTATION

Businesses are playing into the hands of cybercriminals if they don’t upgrade their deployments to meet the ever-more sophisticated design of attacks. Once they have gauged the extent of their vulnerability, the ideal solution will be able to adapt to any new type of threat. 

The benefits are an enhanced user experience, improved reputation, increased conversion rates, and optimized income. But when merchants are unaware of their false-positive rates, and 33% of customers agree they would stop supporting a business that flagged their legitimate transaction as fraudulent, these businesses have no idea how much money (and customer loyalty) they are losing. Ultimately, getting fraud close to zero is not only achievable but the biggest benefit of a full-spectrum online strategy. 


Benjamin Fabre is the co-founder and Chief Executive Officer of DataDome.