To sustain optimal revenue growth and rapport with advertisers, as well as support a positive UX for users on your platform - it’s important to monitor ongoing and as consistently as possible for unusual or fraud-like user behavior. It’s also important to take quick actions on any suspect users. This will deter fraud activity as opposed to encourage it by rewarding invalid users. The user log data history you will capture on users is particularly important to utilize, because this data is not readily accessible for the network or advertisers.
The following are some examples of identifying unusual or fraud-like behavior when analyzing user logs:
- Unusual IP activity - which might include accessing your platform using a proxy IP; or users with an unusual high number of IP connections as they log in; or users connecting from varying geographical locations in a short timeframe.
- Potential multiple/duplicate user accounts of the same user.
- User Agent sameness across suspect accounts
- Unusual behavior on all advertising channels/including Adscend Media Offer Wall - ie: utilizing the log details and completion history in correlation to user behavior on all advertising placements.
General Fraud Control Best Practices:
Use a Proxy Service: We recommend the following proxy services:
1. MaxMind
2. https://getipintel.net/#API
If you do not have the budget for MaxMind, this service is a strong second option.
Block Known Fraudulent IP Ranges: If you have the ability to block IP ranges, this site will provide consistent/common known proxy server IP ranges to block: https://myip.ms/browse/ip_ranges/1/ownerID/38903/ownerID_A/1
Freeze Credits on Suspicious User Activity: Utilize or create the ability to freeze credits to any suspicious users on any single lead payouts (ie: users who have used proxies, flagged for multiple accounts, odd behavior on other advertising channels, etc.). For any suspicious users, payouts should be frozen until you receive payment from the network.
- Based on our current payouts, and typical behavior of fraud users - we recommend to freeze credits on any high payouts with a single lead of $7.50 or greater. If you have the ability to whitelist seemingly valid and trusted users, we then recommend to keep the system optimized for valid users and allow credits to process normally for that group of users.
- We recommend to freeze credits on users if a single user shows $20+ / or an unusually higher amount in credit in terms of what you see on average for every 24 hour period. As mentioned above, if you have the ability to whitelist trusted valid users this is recommended as well.
Note Unusually High Conversion Rates: Unusually high conversion rates can be a red flag for fraud - if you notice an unusually high conversion rate for a particular campaign, it can be helpful to check for other aforementioned suspicious components of the user activity. Some campaigns also might naturally have very high conversion rates as compared to others, which can be normal. If you aren’t sure, your account manager can help determine this.
Block/Ban Non-Compliant Users: It’s important to quickly block/ban users who are non-compliant. This can help you avoid long term revenue charge-back issues, as well as avoid being blocked by top-tier advertisers. A system for disallowing new user registrations on any user info that matches a suspicious or banned account will also help mitigate long term issues.
Implement User Verification Steps: reCAPTCHA, and mobile phone/SMS verification at user sign up can help limit invalid users.
Monitor New Users Closely: If your platform is new, keeping a closer eye on new users is important until you are able to collect more data on their activity. As you perform new user acquisition, fraud users may unfortunately become attracted to your site and it may be difficult at first to determine that the users may behave fraudulently. If you have a high volume of mainly new users, some publishers will temporarily freeze credit for new users in this case until payment is released or until analyzing more of the user’s behavior. More established platforms may also take this same approach when they push mass user acquisition campaigns and onboard a higher number of new users.
EngageMe.TV Fraud Control Best Practices:
Please see below information related to compliance procedure for EngageMe.TV:
- Adscend uses multiple layers of fraud detection (in-house and 3rd party services)
- We actively block IP addresses and user IDs related to fraudulent activity
- In very rare cases, a legitimate IP may become blocked – we support valid users as they inquire by removing the block on their IP if this occurs.
- Proxy/VPN use is not permitted.
- Users who are caught using scripts or bots will result in a permanent user ban as there is a zero tolerance policy for this activity.
- Watching multiple simultaneous videos for credit is not allowed. As the system detects this, the user will become blocked. We encourage publishers to have an open discussion with users about this and get users back on track – in this case, as users agree and become compliant in their activity, they can become unblocked.
- There can be some users who perform sophisticated fraud. To help mitigate potential revenue loss related to fraud on ETV, we recommend to consider freezing points for 48-72 hours on first-time users who are new to your platform. Fraud activity will typically be discovered and reported during this time frame. Some publishers may decide to collect data first and see if ETV fraud reversals do become an issue before implementing this.
- Based on fraud activity that we see, it can also be common for users with different subids to have IP addresses with matches in the string – for example: multiple different active ETV users in your system who’s first two numbers in the IP string are identical, such as multiple user IPs starting with: 50.35.###.## .. or 72.189.##.### .. or 90.142.###.## - If you see this activity, we recommend to freeze credit in this case as well for 48-72 hours.
Questions on Fraud Control?
Reach out to your account manager and he or she would be happy to help answers any questions that you may have on controlling fraud.
Comments
0 comments
Article is closed for comments.