Trust and safety on trading platforms

At the year-end conference, our software engineer, Dragan Jović, explained what it’s like to be on the team accountable for fraud prevention on trading platforms, showing us challenges and trends in maintaining trust and safety.

This helpful lecture taught us the challenges and trends in maintaining trust and safety on trading platforms. We learnt that trust and security are fundamental components in business that provide the basis for successful, safe, and long-lasting business relationships.

Trust and safety importance

Safety and trust on trading platforms are of enormous importance. Trust drives user engagement and business growth. Security ensures a positive user experience and protects users from harm. Both include protecting users’ data, securing financial transactions, and implementing fraud prevention measures on online platforms for advertising and sales.

Trust and safety in practice

How do trust and safety on trading platforms look in practice?

Fraud prevention encompasses various tactics such as detecting fraudulent activities during transactions, safeguarding against “phishing” scams targeting users, preventing identity theft, impeding deceptive practices, and combating fraudulent purchase and sale schemes.

It also encompasses the protection of privacy and financial transactions, which includes personal data security, secure cash payments, verification of seller and buyer credibility, and prevention of hacker attacks (this is not necessarily a T&S activity).

Finally, trust is guaranteed by ensuring standards and quality in advertising on the platform, moderation of communication, handling of spam and inappropriate content, fair and timely resolution of disputes between customers and sellers, creating a safe transaction environment, and building mutual trust between users.

Technical approaches and practices

In terms of technical approaches and practices, there are verification systems, automatic fraud detection, and machine learning and AI detection.

Verification systems use multiple ways of verifying the user’s identity. Automatic fraud detection also utilizes different technologies to detect suspicious activities and potential fraud. Lately, machine learning models and AI have vastly improved fraud detection.

Data protection

Secure data handling implies encrypting user data and ensuring compliance with privacy regulations (GDPR). Transparency refers to well-defined and highlighted rules for the utilization and dissemination of user data. User’s disposal of their data implies giving users the right and access to manage all settings related to their data unconditionally. 

Customer support

In customer support, it is essential to provide support to users at the optimal time, most often through direct contact by phone, chat, etc. Expertise is also crucial because customer support can quickly and fairly solve disputes or user problems on the platform.
Finally, fraud trend tracking and training enable support teams to grasp emerging fraud patterns and address various customer issues effectively.

Statutory and legal compliance

Reporting mechanisms enable platform users to easily and quickly report suspicious illegal activity they observe.

Close cooperation with law enforcement agencies in the fight against fraud is an everyday thing.

Upon request, all data, traces, and history of changes left by the fraudster on the platform shall immediately be sent to the police or the court in order to provide necessary information and assistance in investigative actions.

Education of platform users

Platform users are mainly educated by providing simple guidelines, training them on secure buying and selling habits, and raising the user’s awareness of fraud methods and reporting them.

Guidelines for safe business:

• Refers to educating platform users on secure buying and selling practices.
  

Raising awareness about fraud:

• Refers to providing advice on identifying, avoiding, and reporting potential frauds. Additionally, the platform produces informative articles and videos on fraud prevention.

Timeliness and adaptability

Feedback is important. Information and complaints from users can be collected to identify areas for improvement in fraud detection. Then, approaches and strategies in detection can be modified based on monitoring changes in fraud methods (fraud patterns). Furthermore, proactivity and cooperation are essential, and they involve monitoring news in fraud prevention and collaboration with other companies that help fight online fraud.

Building trust on the platform

Reviews and transparency are essential to building platform trust. Encouraging users to share their experiences with other users, giving ratings, and building confidence in the platform is also vital. Furthermore, encouraging users to share experiences on other “trust” platforms (Google, Trustpilot, etc.) enhances credibility. Details of legitimate and successful transactions always contribute to trust.

There is also something called an internal seller score, which is better positioning and promotion of responsible users who bring an excellent rating to the platform. In this way, others are encouraged to behave similarly.

Examples of frauds

Common frauds in car buying include:

• Selling a non-existent vehicle (with advance withdrawal)
• Selling a stolen vehicle for cash
• Selling zombie or clone vehicles (by manipulating the VIN)
• Selling vehicles without proper documentation
• Selling significantly damaged vehicles as functional
• Selling cars with false mileage

In such frauds, the customer usually tries to redirect to some form of external “uncontrolled” communication outside the platform so that they are not caught or prevented from committing fraud. Also, a common type of fraud is the theft of other people’s accounts on the platform (phishing). This allows fraudsters to commit the mentioned frauds more easily for a period of time, i.e., hidden behind stolen accounts of valid individuals or legal persons on the platform until the identity theft is reported or detected.

Fraud indicators

Some of the fraud indicators are:

• Significantly lower vehicle prices compared to the identical existing vehicles on the platform
• A sudden reduction in the price of the car shortly after the publication of the ad with a realistic starting price
• Similarity of advertiser, ad, and vehicle data associated with previous fraud
• Suspicious user behavior and communication
• Attempts to impersonate valid private advertisers and car dealers
• Avoiding communication on the platform
• Add text or contact information to photos

Platforms should be dedicated to monitoring and detecting these and other similar types of fraud, considering the quality of ads and any complaints received, ensuring a safe and reliable environment for all users.

In that case, the ads go to resubmission and are not visible on the site until all the reported defects are removed.

Inspection and detection

Data inspection is performed for each insertion or modification of an ad, change of user data, or new message sent. If detection occurs and customer service confirms it, the fraudster’s user account with all related ads is automatically deleted from the site, and other users who were in contact with the fraudster are immediately informed about it with advice for further action.

All data on fraudsters remains recorded due to legal consequences or the need for ML model training and filters to detect new fraudulent ads.

Secure and reliable buying and selling platforms

Dragan concluded that maintaining trust and safety on trading platforms is challenging, constant, and multidimensional work. By combining modern technical practices and user-friendly approaches, a secure and reliable buying and selling platform can be created.

Explore our blog section for upcoming captivating topics that we’ll be sharing with you soon.