Great point – that's exactly the challenge. Right now we filter for open-market buys and sells only, which already removes a lot of the noise from automatic/planned sales (like 10b5-1 plans). We're exploring additional signal filters, like flagging trades that are unusual relative to an insider's historical pattern, or highlighting cluster buying (multiple insiders buying around the same time). Would love to hear what kind of normalization you'd find most useful.
