Help Center

| Submit or View Help Requests | Developer Docs |
|

View desktop instructions
View mobile app instructions
Advocate Fraud Protection Explained

Potential fraud is a consideration for every brand with a referral program. Inevitably, someone will attempt to test the system by referring themselves.

Your Advocate program has an optional, automated Fraud Protection feature that can be used together with manual actions like blocking email domains and specific participants and IP addresses to mitigate the effects of fraudulent behavior. However, keep in mind that the strongest defense against referral program fraud is effective program design.

Referrals are only placed into a pending state after you turn on Fraud Protection for your Advocate account.

  1. In the left navigation bar, select ellipsis-v-solid.svg [Menu] → Settings.

  2. From the Advocate Settings section, select Security.

  3. Select toggle-on-solid.svg [Toggle on] Automatic fraud detection.

Fraud scoring explained

When Fraud Protection is turned on, Advocate will evaluate and assign a fraud score to all new referrals.

The fraud score ranges from 0 (no risk detected) to 100 (very high risk). It is calculated by comparing each new referral against previous referrals to look for similarities. The overall score assigned to a referral determines whether it’s automatically approved or denied, or placed into a pending state for manual review.

Factor

Influence on score

Example Referral Information

Same email address

High

robinbanks@example.com and robinbanks+test@example.com

Same name (exact match only)

High

Robin Banks and Robin Banks

Same IP address

Medium

192.158.1.38 and 192.158.1.38

Similar email address

Medium

robinbanks@example.com and RBanks@example.com

Similar name

Low
Only used to support email address and IP address detectors.

ROBIN BANKS and Robby Banks

Fraud Protection settings

Advocate supports three Fraud Protection levels: Relaxed, Moderate, or Strict. The level you select applies to all of your referral programs. Legacy referrals won’t be evaluated.

Relaxed

The Relaxed setting is best for brands who want to streamline their referral program and reduce the likelihood of denying legitimate referrals. This level allows more referrals while accepting the possibility of some self-referrals.

Using the Relaxed setting, most referrals are automatically approved. If a referral is highly likely to be fraudulent, then it will be automatically denied.

Fraud Score

Action

0 to 84

Automatically approved

85 to 100

Automatically denied

Moderate

The Moderate setting is appropriate for brands with capacity for hands-on review of their referrals. This level provides a middle ground, allowing manual review of potential self-referrals for a more accurate referral program.

Fraud Score

Action

0 to 69

Automatically approved

70 to 84

Manual review required

85 to 100

Automatically denied

Tip

If you turn off fraud protection while any referrals are in a pending state, these referrals will remain pending until manually reviewed.

Strict

The Strict setting is recommended for brands prioritizing highly secure programs. This level may occasionally flag legitimate referrals, but it significantly reduces the likelihood of self referrals.

Fraud Score

Action

0 to 49

Automatically approved

50 to 100

Automatically denied

Participant experience for instant access widgets

Typically, programs using instant access widgets immediately grant a reward to the referred friend. When Fraud Protection is turned on and Advocate detects a suspicious referral attempt, the referral is either denied or placed into a pending state, depending on its fraud score and your settings. Customer advocates see a message within the widget letting them know the referral was either denied or flagged for manual review. For pending referrals, no reward is issued unless you approve the referral.

Instant access widgets for Advocate show a warning message when someone attempts to self-refer

Did you find it helpful? Yes No

Send feedback
Sorry we couldn't be helpful. Help us improve this article with your feedback.