Types of anti-fraud operations models for your business

Fighting fraud online in eCommerce cannot be done without third-party providers. There are several different types of antifraud third-party providers. The choice depends on your business maturity and internal competency.

One type is full outsourced antifraud with “insurance for chargebacks”. That is a situation where you pay your antifraud provider a premium rate for processing payments. In case there is a chargeback, it is covered from the “insurance” pool by the provider.

Another type is hybrid, where you get a discounted rate for processing payments, and you are responsible for a part of the chargebacks. Fraud detection and prevention are done by the provider, but you decide which orders to approve or decline.

The most advanced companies build their own internal fraud teams and use antifraud providers only for specific tasks, like scoring, data enrichment, monitoring, or other machine learning.

There are many options in between these three, and the best way to see what suits you is to try out a few and see what works best for your company.

Fully outsourced model

In the fully outsourced model, the eCommerce company pays the antifraud provider a premium rate for processing payments. In case of a chargeback, it is covered from the “insurance” pool by the provider.

The main advantages of this model are:

  • You don’t need to worry about chargebacks, as they are covered by the provider.
  • You don’t need to have an internal fraud team.

The main disadvantages of this model are:

  • You are at the mercy of the provider, as they have the ultimate decision on whether to approve or decline an order.
  • The provider has the incentive to reject any slightly suspicious case, as they make more money from the premium that was not used on covering chargebacks.
  • You don’t have visibility into the fraud detection process, so it’s difficult to improve it.

Hybrid model

In the hybrid model, you get a discounted rate for processing payments, and you are responsible for a part of the chargebacks. Fraud detection and prevention are done by the provider, but you decide which orders to approve or decline.

The main advantages of this model are:

  • You have more control over which orders are approved or declined.
  • You have visibility into the fraud detection process so it’s easier to improve it.

The main disadvantages of this model are:

  • You still need to have an internal fraud team.
  • You need to be able to integrate with the provider’s systems.

Internal fraud team

The most advanced companies build their own internal fraud teams and use antifraud providers only for specific tasks, like scoring, data enrichment, monitoring, or other machine learning.

The main advantages of this model are:

  • You have complete control over the fraud detection process.
  • You can customize the fraud detection process to your specific needs.
  • You don’t need to pay for a service, just for the tools that you use.

The main disadvantages of this model are:

  • You need to have an internal fraud team.
  • You need to be able to use the tools that you’re paying for.
  • You need to keep up with the latest fraud trends.

Progression

A usual progression is to start with the fully outsourced model. Start collecting data, start slowly building internal competencies, and understand the fraud landscape.

As the company grows, it moves to the hybrid model. With an understanding of the fraud landscape, it starts to understand case management and usage of data. Enough data is collected to optimize rules provided by the provider.

And finally, builds its internal fraud team. The fraud team does not consist only of case management fraud fighting operations, but also of data analytics and machine learning experts.

Once a company builds its own team, the focus is on improving rules, adding rules, finding data sources for enrichment, and improving scoring and detection.

Drivers of progression

The main drivers of progression are costs and customer experience.

  • With the fully outsourced model, a company pays a premium for every order. As the company processes more orders, the costs go up. At some point, it becomes cheaper to build an internal team.
  • With the fully outsourced model, the company is at the mercy of the provider. Any change in the provider’s rules can lead to a change in customer experience. As the company grows, it wants to have more control over the customer experience.

What holds back the progression

The main reasons for not moving from the fully outsourced or hybrid model are lack of competency, scarcity of talent, operational complexities, and management costs.

  • Lack of competency: most companies don’t have the internal competency to build a fraud team. They don’t know where to start, what to do, and how to do it.
  • Scarcity of talent: it’s hard to find people with the right skill set. Even if you find them, they are very expensive.
  • Operational complexities: it’s hard to set up the right tools and processes.
  • Management costs: it’s expensive to manage a fraud team. You need to invest in training, coaching, and tools.

Where to start

There is no “one size fits all” solution when it comes to choosing the way to handle antifraud. The best way is to align based on the cost considerations and internal competencies that you have.

If your company does not rely on fraud fighting as a core competency (case of payment companies, where it is one of the main competencies), the best is to slowly progress from fully outsourced to fully managed.

If your company relies on fraud fighting as a core competency, then it is better to start with a fully managed fraud team from the beginning.

When to build own team

The most important considerations when choosing a model are:

  • The size of your company: the bigger the company, the more likely it is to need an internal fraud team.
  • The fraud landscape: the more complex the fraud landscape, the more likely it is to need an internal fraud team.
  • The customer experience: the more important the customer experience, the more likely it is to need an internal fraud team.
  • The cost of customer acquisition: the higher the cost of customer acquisition, the more likely it is to need an internal fraud team.
  • The cost of goods sold: the higher the cost of goods sold, the more likely it is to need an internal fraud team.
  • The volume of transactions: the higher the volume of transactions, the more likely it is to need an internal fraud team.
  • The customer lifetime value: the higher the customer lifetime value, the more likely it is to need an internal fraud team.

Manage antifraud rules smarter.
Use a decision engine.