How one food delivery merchant saved $300,000 stopping promo abuse

Key outcomes:

• $300,000 saved within 6 months

• Established more cost effective marketing programs long-term

• An average saving of $15 per blocked order

When marketing promotions become a problem

This case study is based on a popular food delivery merchant who was running a referral scheme. This is where an existing customer shares a referral code with a friend and they both get a discount in return.

The scheme was a big hit. The merchant was keeping existing customers happy and gaining lots of new ones. Success all around, right? Well, not quite.

The fraud team suspected something was off. And this very quickly proved to be true. Step-by-step instructions on how to finesse the referral scheme were plastered across social media and forums like Reddit. In this case, there was such a thing as a free lunch!

Worryingly, they didn't know how big the problem was, and how much it was costing their business.

We began working with them to look into it. “It quickly became clear to all that what the marketing team saw as growth wasn’t telling the full story,” says Ruth Crocker, our Director of Investigations.

Referral abuse skews sales figures

The marketing team were celebrating inflated numbers. These were not hundreds of new customers. These were individuals getting a heavily discounted second, third and fourth helping.

Yes, these customers were spending money. But if they never pay full price, it will eat into your profits!

Ruth points out that “around 90% of the referral abuse was being committed by only a few hundred users.” But the losses were much worse than the fraud team feared. We needed to find out how much worse.

Connections are key to identifying referral abuse

The company was already using Ravelin to fight online payment fraud, so we had some information about their customers. But they weren’t sending us any data about their referral schemes. Nothing was being tracked internally, so they were completely in the dark.

They had hundreds of thousands of customers who were ordering food on a daily basis. But looking at these customers individually would be tedious and long-winded. They needed a broader view of what was going on to figure out how to stop it faster.

The key to sussing out referral abuse is looking at the connections between individuals to see who is benefiting from reusing codes.

Fraud team challenge: proving the business case

Referral schemes are one of the most effective tactics for growth. In fact, 92% of consumers trust recommendations from friends over traditional advertising. So it’s no surprise that they’re a marketing team top pick.

Referral schemes are not traditionally thought of as a problem. Fraud teams often have their work cut out for them getting the wider business to sit up and take notice of abuse.

Ruth notes that "fraud prevention solutions are often seen as a blocker to sales and marketing campaigns. It was important that we actually measure and quantify just how big the problem was. Not just for mapping out our solution but for wider buy-in."

Many brands would rather deal with the costs. Even if they end up losing thousands of dollars a month. "We needed to prove that limiting bad actors was essential to long-term revenue growth.”

Our goal was to stop abuse without stopping promotions altogether. As Ruth mentions, “many opportunistic customers are likely to pay full price once they work out their tactics aren’t working anymore.”

We had to prove to the business that they could have their cake and eat it. They would still pull in new customers. And more people would pay full price in the long run.

Using graphs to highlight referral abuse networks

The main method for highlighting the extent of promotion abuse is our graph database.

Ravelin’s graph database allows you to create a connected graph of your customers. This graph includes emails, phone numbers, device IDs and payment methods. Customers that share any of these attributes make up a network.

Our solution was to add a referral code node to track this specific connection between customers. We would be able to tag all the customers that had ever used these codes. We could provide stakeholders with a visual representation of who had used more than one code and how many times a code had been used overall.

Preparation for immediate impact

Because our client had never sent over any data about their referral codes, we didn't have any of them in the network. This meant that it would take some time before there were enough codes to detect abuse.

Ruth explains that “we needed to show the extent of the abuse right away and guarantee immediate impact to get the support of the business. So we implemented a plan to prepare the network.”

We had already added the referral code node to all customers that had recently signed up, which was a good start. To help us along, we asked for a list of all known referral abusers. We then added these nodes to their networks. Finally, we did some additional research and analysis to spot fraudy networks and added referral nodes to all the customers in them.

Setting limits on abuse without losing customers

Once we'd got the ball rolling, we could begin setting thresholds and limits.

Limits and thresholds control how many times a referral code can be used and by how many people within a network.

When we spotted referral abuse, the merchant would inform their customer the code is invalid and allow them to continue without the discount. This meant that they could target abusive behavior without losing a customer.

Result: Over $300K saved in 6 months

In the beginning, around 5% of total referral vouchers were signaled out as abuse. Within the first six months of working together on this, an average of 15% of referral abuse customers were identified and blocked.

The financial implications of this were huge. The merchant saved an average of $15 per block which was between $40-80,000 a month.

We acted on the suspicions of the fraud team to identify the scale of the problem and raise awareness across the business. Demonstrating that this could be addressed without restricting sales was a big win. The marketing team and wider company are now in a position to spend money more effectively.

Ultimately, we were able to work together to lower customer acquisition costs and increase customer lifetime value by addressing promotion abuse.