- From: Don Marti via GitHub <sysbot+gh@w3.org>
- Date: Thu, 17 Oct 2024 21:38:40 +0000
- To: public-patcg@w3.org
@wbl see the comment from @martinthomson above on the stealing problem. In practice, the problem is that stealing conversions scales well and can be automated. The more data points you have on more people, the lower the cost to predict who's likely to buy something. Legit attribution, on the other hand, doesn't scale well. Making a site where people pay attention to the ad, and making an ad that sells, are both hard and hard to predict. So in an attribution system that is weak on fraud, the legit conversions will be swamped by stolen ones -- including both intentional fraud and ML that "learns" to do stealing and doesn't get caught. -- GitHub Notification of comment by dmarti Please view or discuss this issue at https://github.com/patcg/admin/issues/26#issuecomment-2420648178 using your GitHub account -- Sent via github-notify-ml as configured in https://github.com/w3c/github-notify-ml-config
Received on Thursday, 17 October 2024 21:38:41 UTC