Unfortunately, any time in marketing when you offer valuable incentives in exchange for an action, you risk attracting undesirable behavior from those looking to unfairly profit from the offer. And the more generous the rewards and incentives you offer, the more likely you are to attract such behavior.
And so, if you don’t take appropriate action, then you can find that your marketing campaigns cost you a lot of money in return for little or no measurable benefit.
And in this respect, referral programs are no different. In the case of referral marketing, the behavior you need to consider protecting yourself from includes:
- Valueless transactions – where a reward and/or incentive is paid out and the referred-in friend later returns the purchase or simply doesn’t go on to complete the purchase, for example by not paying a bill or premium.
- Self referral, where a customer seeks to refer him- or herself using the same or a different name or email, typically to get a discount on the next purchase and/or the referral reward.
- Referrals without any intention of completing the transaction on behalf of the referred-in friend. In such a case, often the referrer and the friend do exist and use their actual identities, but work together to get a reward and/or incentive while having no intention of completing the transaction or by simply taking advantage of the low minimum commitment required in order to earn a more valuable reward.
- For example, if your program offers a reward of $100 for creating a savings account with a minimum deposit of US$1, you shouldn’t be surprised to get lots of accounts funded with exactly US$1, as you’re offering US$99 free money!
- Alternatively, where a valuable reward is paid out for an action that doesn’t have a cost to the referred-in friend, as Uber found out with its consumer referral program, this led to lots of self referral simply to get an extra free ride.
- ‘Get a Referral Code’ sites and affiliates where a new customer can get a referral code from someone completely unknown to the referrer. In this case you end up paying rewards and incentives for a transaction that would probably have happened anyway and/or even double paying by paying an affiliate for that transaction as well.
- Abuse by agents, brokers and employees creating fake referrals to close sales and earn commissions or by paying out for fake referrals.
- Wider Fraud and Money Laundering – this includes more organised fraudulent activity such as creating fake referrals to earn money on a large scale, using stolen identities or credit cards to get products and rewards, and even money laundering, where illegally obtained money is used to purchase something that isn’t wanted just to obtain cash or cash equivalent rewards that can be used legally.
As you can see, these range from normal consumer behavior, such as returning goods that were disappointing, to valuable customers trying to get extra discounts for repeat purchases, or people simply trying to take advantage of your naivety to get ‘free money’, and up to something altogether more nefarious. What they all have in common though, is that unless you take steps to mitgate against all of these actions, you risk paying out valuable rewards without receiving any benefit in return.
Designing your Referral Program to reduce Gaming and Fraud
Here are some of the things you can do with the design of your referral program to reduce gaming and fraud:
- Make sure your referral terms and conditions are bullet proof:
- State clearly on your referral program that you will not pay out for self referral or for the referral of existing customers etc.
- Consider one-side vs two-sided rewards
- Consider the value of the Rewards and Incentives you offer
- Consider the type of rewards you offer
- Consider a closed program
- Consider Referral or Reward Limits
- Consider only paying out Rewards after X confirmed referrals
- Investigate large referrers
- Check to see if you’re paying commissions elsewhere for the same new customer
- Real time verification of the referred-in friend
- Control access to your referral platform
- Block bots and suspicious emails
- Platform based abuse detection
The first and most obvious thing you can do is reserve the right in your terms and conditions to not pay out rewards and incentives for any unwanted behavior. This is important, as you may find that without doing this, that some types of unwanted behavior will entitle a referrer and/or friend to a reward or incentive.
While your terms and conditions provide you with the legal right to refuse to pay out rewards and incentives, the reality is that few referrers or friends will read these. So it’s a good idea to put a simple statement up front on the program stating that, for example, ‘this offer is not open to existing customers and that self referral is not permitted’ etc.. This may discourage some gaming.
While, on the whole, we see that two-sided rewards tend to work better to generate referrals, if you’re particularly concerned about gaming, you might decide not to offer an incentive to the referred-in friend, for example. At least you could test the difference between offering one-sided and two-side rewards.
While higher value rewards tend to increase referrals, they can also increase the incentive to game the system. So you should also think carefully about the value of the rewards and incentives you offer compared to the value of the action you are looking to drive. Being too generous will likely encourage gaming, as in the example of the US$100 reward for opening a savings account with a minimum US$1 deposit mentioned above. So the sweet spot is to offer a reward that is just enough to encourage referrals but not so generous as to generate unwanted behavior.
Surprise, surprise, the type of rewards most likely to encourage gaming or fraud are cash rewards or cash equivalents. So you can consider offering other rewards that are less susceptible to this kind of behavior, such as gift cards or vouchers that require a minimum spend and can’t be stacked? Alternatively, you could offer non-monetary fan based prizes, like Tesla did with its factory tours.
Again, all things being equal, an open program that allows anyone to refer your brand will drive more referrals. But you can consider making your program only available to existing customers, or a subset of existing customers, by putting it in a customer area that can only be accessed with a customer email and password and/or requires a customer identifier. The assumption being that genuine customers who love your brand will be less likely to engage in abusive behavior than random people who stumble across your program and are incentivized to share just to get a reward.
One way to limit the extent of gaming and fraud is to limit the number of referrals or the total rewards that any one person can earn over any given period of time, or even over the lifetime of the program. No doubt you’ll have seen referral programs advertising how you can ‘earn up to US$550 a year’. While these limits can reduce the number of total referrals overall, they do limit the amount any one bad actor can earn from gaming the system.
Hear from Asya Kuznetsova Product & Growth at Wise why the brand chose to protect value by only paying out after 3 confirmed referrals:
How to Ensure You Only Get Valuable Referrals | Wise
Robin Bresnark: "The referrer gets their reward after three referrals, rather than for every one, which is unusual."
Asya Kuznetsova: "We want to provide the cheapest transfers possible and that means that we have to maintain our price really low. We mostly want to grow our loyal customer base."
Asya Kuznetsova: "We don't want you to have this as a primary incentive 'Oh I want to earn some cash out of Wise'. We want you to genuinely recommend your products to others."
Asya Kuznetsova: "Let's say invite three friends comes in really handy, because again if it's one friend, you're usually 'Oh yeah, I would invite this person and earn £20 out of them' and that's it."
Asya Kuznetsova: "It's a long journey, and I would recommend Wise to others around me, not only 1 or 2 people. That sets a very interesting path for loyalty and advocacy. And people who are simply interested in earning fast cash, they wouldn't enter this. So that's a tool that we use to work specifically with people who love our product and genuinely want to recommend to people around them."
Another way to reduce the incentives to game the system can be to only pay out rewards after, say, 3 confirmed referrals, as the Wise referral program does. Again while this may demotivate genuine referrers who feel that they can only generate one or two referrers, it may reduce the temptation to game the system.
“It’s a long journey, and I would recommend Wise to [many] others around me, not only 1 or 2 people. That sets a very interesting path for loyalty and advocacy. And [for] people who are simply interested in earning fast cash, they wouldn’t enter this. So that’s a tool that we use to work specifically with people who love our product and genuinely want to recommend to people around them.”
Asya Kuznetsova Product & Growth – Wise
Hear from David Hixon, Executive Director – Head of Product & Lifecycle Marketing at Ally Bank as to why you need to investigate large referrers to see if they’re recommending you for the right reasons:
Not All Referrers Are Equal | Ally Bank
David Hixon: "If I looked at the distribution of the number of referrals of all of the customers that have taken part in the program, the lion share of them is in the one and two range.
We have customers that have hundreds and hundreds of referrals. And it it's super fascinating to go and understand what that really means. I'm looking now our number one person has 661 referrals.
There's a big difference between somebody who has 10 referrals and somebody who has one referral, or two referrals.
Like the 'onesie twoosy' guys are probably the best ones you have, because that's the conversation over dinner with your with your high school buddy where you say, 'Hey, man. I don't know where you bank, but I'm at Ally and they have great rates and a great product and a great brand. You should go check it out.'
That's a more organic referral that is more likely to turn into the type of customer that we like versus what we're finding is these guys that have hundreds and hundreds of referrals are basically influencers.
You've got folks out there that are influencers around financial services and their influence is around finding the right the right bank and the best bank, based off of product set and feature set and digital capabilities and rates and things like that.
And if that person is telling a good Ally story, and then including their referral code, we love it because that person's coming to us because they love Ally Bank.
Conversely, there are influencers out there that are that have an audience of people that are trying to find a way to take advantage of the offers that are in market, just across the board."
See the full interview here
Where you don’t have caps on the number of referrals or total rewards, if a referrer generates a large number of referrals within a short period of time, your referral program should generate an alert to investigate this. While there may be perfectly good reasons for this, such as an influencer genuinely recommending you as the best product or a motivated brand advocate signing up all the parents in a school, it could be a ‘Find a Referral Code site’ or an influencer posting a way to ‘get free money’. Investigating this and blocking the referral codes and/or the IP address generating the bad referrals can stem the problem.
“What we’re finding is these guys that have hundreds and hundreds of referrals are basically influencers. You’ve got folks out there that are influencers around financial services and their influence is around finding the right the right bank and the best bank, based off of product set and feature set and digital capabilities and rates and things like that. And if that person is telling a good Ally story, and then including their referral code, we love it because that person’s coming to us because they love Ally Bank. Conversely, there are influencers out there that are that have an audience of people that are trying to find a way to take advantage of the offers that are in market, just across the board.”
David Hixon, Executive Director – Head of Product & Lifecycle Marketing – Ally Bank
Hopefully your back end systems will let you see if you’re also paying a commission or fee for the same new customer elsewhere. This could be where an affiliate is passing on a referral code but also trying to claim an affiliate commission. We saw an example, after replacing an in-house program for a large North American telecoms brand, of where a brand’s own agents were registering their clients as referred-in by other clients, to effectively get the new customer a discount and close more deals and earn more commissions for the agents. This leads to double-paying and a regular audit of your customer can help identify this.
Where you cannot delay the payment of rewards and incentives until you can carry out checks, you can implement a real time real time verification of friend eligibility with a Friend Segmentation webhook and API to see if the email address is already in the database.
As mentioned above, one of the principal causes of fraud is the actions of bad actors who have access to your referral program, such as employees, agents and intermediaries. For this reason, it’s crucial to ensure that you limit access to your referral platform back end to trusted and identified individuals.
Your referral platform should be able to identify and block bots as well as suspicious emails, such as known fraudulent emails, known joke emails, temporary emails and gibberish.
Many leading referral platforms offer a variety of real time checks to try and identify obvious or potential cases of gaming or abuse based on a series of identified variables such as cookies, IP addresses, time between referrals etc. and by assigning scores to each variable. As explained below, the aim of these is to identify the most suspicious cases by accumulated scores and prioritize these for review by your team. However, these are not a 100% protection against gaming and fraud and, as explained below, are susceptible to the false positive problem.
The Only Way to Completely Avoid Gaming and Abuse and Paying for Valueless Transactions
While it can be very difficult to prevent well organized and determined fraud, and you should not expect your referral program to provide full protection against it, all of the solutions listed above can help limit the risks of gaming, abuse or fraud, often at the cost of reducing the total number of referrals you’ll get, or by risking creating false positives. But in reality, the only way to avoid problems associated with gaming and abuse is to delay the payment of any rewards and incentives until you’re able to fully determine the value of the referral.
In other words, you only pay out rewards and incentives post completion after certain conditions have been met, such as the new customer not cancelling within the statutory cooling off period and having made a payment. By delaying the payment of rewards and incentives, this also gives you time to check the new customer against your database to see if the customer is new and do any KYC or other anti-fraud checks you would typically do with any transaction.
This of course requires not only that you can delay reward payments but also that you have systems in place that enable you to make these checks and that you’re able to reconcile the referrals with the information from your back end and then communicate the outcome of these checks to your referral system. Ideally you’ll be able to automate this reconciliation, but you may be able to upload CSVs or even manually approve or reject as required. If you work with an enterprise referral platform like Buyapowa, you will be able to transmit this information to the Buyapowa platform and fully protect your margins by only paying for valid referrals, and even pay different rewards according to the value of each referral.
Gaming the System and the False Positive Problem
One of the problems of trying to use any automated system for fraud detection based on scoring is the problem of false positives, where falsely denying a valid referral can lead to customer dissatisfaction, and even leading to customers posting in social media that your program is a fraud.
To understand the problem, consider this example:
Imagine my name is Freddie Smith.
- I’m named after my uncle Fred Smith.
- I invited my uncle to my house for Sunday lunch and said: ‘Uncle, you need to get that great mobile phone contract I was telling you all about. Let’s get you up and running now, and I can get you a great deal.”
- So, I open my laptop, send my referral link to my uncle via email, close my account and get my uncle to open his email on my laptop and complete the transaction.
- All in a couple of minutes.
In any kind of scoring based system this looks very like gaming:
- Fred Smith refers Fred Smith
- With the same browser cookie and the same IP address
- Within 5 minutes
- With similar gmails fredmith1979 and fredmsith1997
Denying this because of rules would create two unhappy customers. That’s why most scoring systems recommend that referrals that are scored and listed as potential fraud or gaming are reviewed by your team and accepted or rejected, and often have a rule that if a referral is not rejected within a set time period, it’s automatically accepted. This either requires that your team regularly inspect and manually approve and reject, or that you set up an automation connected to your back end systems to do this.
Is there a tolerable level of gaming?
Given that many of the tactics suggested above, such as making your referral program closed, or limiting the number of referrals any one person can make in any given period, will tend to reduce the overall number of referrals you will get, and the risks associated with false positives, you might ask whether there’s an acceptable level of gaming you’re willing to accept?
This will likely depend on your business model and the nature and value of the rewards you’re paying out. For example, many fast fashion businesses typically see more than 50% of all purchases returned. So they may view the fact that the referred-in friend created an account as being more valuable than the initial purchase, assuming that person will go on to make future purchases. Also the incentive is likely to be a discount off first purchase, and when the product is returned, the effect of that is negated. The inconveniences are mostly the logistics of the return and what to do with a returned dress that appears to have been worn at a party.
Similarly, where a regular customer self refers to get a small discount, a brand may view that as a small price to pay to keep a good customer.
However, where the brand is paying out high value rewards in cash or cash equivalents, as is often the case in the telecommunications, insurance, finance, energy and gaming industries, these arguments are not valid. Here a brand needs to be sure that it is getting value from the referrals. As mentioned above, the best way to do this is to delay the payment of rewards until the new customer has been validated and has cleared hurdles, like paying a phone bill. Then using APIs to automatically reconcile, accept and pay rewards or reject referrals is the best way forward. Also in these circumstances, there’s little benefit in many of the program design changes suggested above, as you can let a wide number people refer your brand confident that you’ll only pay for valid transactions.
What next?
If you’re considering setting up a referral program but are concerned about potential abuse, or if you already have a program and are seeing abuse and wonder how to deal with it, then get in touch and we’d be happy to talk through the options with you.