I built a multi-touch attribution engine that detects ChatGPT and "Dark" traffic(zyro.world)
zyro.world
I built a multi-touch attribution engine that detects ChatGPT and "Dark" traffic
https://www.Zyro.world/
2 comments
OP here. Just to get ahead of the 'why not just use GA4?' questions:
GA4 creates a black box around 'Direct' traffic and struggles with the new wave of AI referrals. We found that for many SaaS/E-com sites, 15–20% of traffic is now coming from LLMs (ChatGPT, Perplexity) or 'Dark Social' (Discord, Slack) which GA4 misattributes.
We built this to give raw access to that data and allow you to act on it (by changing the site experience) rather than just looking at a dashboard 24 hours later.
Happy to answer any technical qs about the intent scoring or our crawler!
GA4 creates a black box around 'Direct' traffic and struggles with the new wave of AI referrals. We found that for many SaaS/E-com sites, 15–20% of traffic is now coming from LLMs (ChatGPT, Perplexity) or 'Dark Social' (Discord, Slack) which GA4 misattributes.
We built this to give raw access to that data and allow you to act on it (by changing the site experience) rather than just looking at a dashboard 24 hours later.
Happy to answer any technical qs about the intent scoring or our crawler!
I built Zyro because I got tired of Google Analytics dumping 40% of my traffic into "Direct" and claiming credit for every sale.
Most analytics tools rely on simple referrers. We built a traffic brain that parses 50+ specific traffic signatures, specifically targeting the "blind spots" in modern tracking:
It also does standard A/B testing, but with a twist: you can trigger experiments based on intent signals (like "read shipping policy" or "rage clicked") rather than just URL routing.
The backend runs on a custom SQL schema designed for high-throughput event logging. I’d love you to roast my attribution logic or ask how we fingerprint the AI traffic.