What is the best way to catch discount code and refer-a-friend abuse where one person creates dozens of accounts?

Audit the redemptions once the campaign has run for a while. Group them by device, payment card and delivery address, and sort the groups by size. Repeated use of those details can expose duplicate claims, but shared details do not always identify the same person. Hold referral rewards until the referred customer's first order is delivered and past the return window, so the audit happens before you pay.

An audit works because creating accounts is easy and creating everything else is not. An attacker can produce email addresses without limit. Devices, payment cards and addresses where parcels can be received are scarce, so they get reused. The audit looks for that reuse across the whole set of redemptions.

Begin with an export. For each redemption, collect the account, the email address, the device identifier, the payment method fingerprint from your payment provider, the delivery address, the IP address, the referral link used and the order outcome.

Normalise before comparing. Reduce email addresses to their underlying mailbox by removing dots and plus-suffixes where the provider ignores them. Standardise delivery addresses, so that variations in flat number, abbreviation and added characters resolve to one address. Use the payment provider's card fingerprint, which identifies a card across accounts without exposing its number. Virtual cards weaken this signal, since one person can generate many, so do not rely on it alone.

Then connect the accounts. Treat any two accounts that share a device, a card fingerprint or a normalised address as linked, and follow the links to form groups. An account that shares a device with one account and an address with another joins all three together. Rank the groups by number of accounts and by the discount or reward value they received.

Inspect the largest groups by hand. Confirm that they are what they appear to be, since a student residence or an office can produce a large innocent group by address alone. Record what distinguishes the abusive groups, such as accounts created on the same day or orders containing the same item.

Put a number on it. The total discount and reward value taken by confirmed groups tells you what the problem costs and what a control is worth.

Turn the findings into rules for the next campaign: a limit on redemptions per device and per address, and a hold on rewards for accounts in a group above a certain size. Choose proportionate actions for the current one. Cancel unshipped orders that break the terms, withhold unpaid rewards, and exclude the group from the offer while leaving the accounts open for normal purchases.

How Kairal helps

Session linking connects recorded devices, IPs and sessions to the user IDs your integration supplies. Compare those visits with your own account and order records before acting; a device match does not establish that two people are the same person.

Session linking