Using Interactive Fraud Proofs to Detect Fraudulent Activity

How we can fight ad fraud as a web3 native Data Management Platform (DMP)

One of the benefits of storing performance data on-chain is the ability to detect anomalies in campaign data across advertisers.

In a Web 2 DSP managed service, there are teams devoted to squashing fraudulent bots on the SSP side of the equation. However, these services are only available after a certain threshold of spend.

With our Web 3 ad server infrastructure, we are able to use interactive fraud proofs to detect anomalies in performance data (eg: impressions, viewability, clicks, unique clicks, video views, conversions). The interactive fraud proof processing will be off-chain, ensuring that the metadata that is available on-chain protects the integrity of the advertiser's business performance data by anonymizing the advertiser with a client ID and campaign ID hash.

Cross-auditing against SSP ad server data

A cross audit with the Server Side Platform's (SSP) ad server data will ensure the closest match to what was actually served in the media campaign. Campaign performance data at a granular, more detailed view enables the SSP and DSP data scientists to tie out performance data audit at a very detailed view (eg: by device ID, ADID, IP address, etc).

Cross-auditing with the advertiser's website analytics

This, combined with the advertisers website behavioral data (eg: unique visitors, page views, average time on site, visitor by country, visitor by IP address, visitor by Device ID), ensures that Myosin's campaign data is audited and as close to accurate as possible before it is minted on-chain on a Layer-2 blockchain like Polygon.

Here are some examples of audit findings we have detected from previous ad performance audits against SSPs:

• A surge in traffic from the same IP address over a time period

• A steady stream of clicks from the same device ID

• Activity from IP addresses that fall outside of the campaign geo-targeting