Skip to content

Journal

·5 min read

Bounce rate lies when agents shop for you

Playwright adding a pair of shoes is not a bounce. It is also not a conversion you should mix with Safari. Actor filters keep both numbers honest.

ActorsGoals

A human hand and a robot hand reaching toward the same cup
Same checkout. Different actor. Do not average them.

Headless agents now fill carts, compare SKUs, and abandon the page in 800 milliseconds. If you count them as humans, bounce rate climbs and conversion rate dissolves. If you throw them away entirely, you miss a channel that is already buying on someone’s behalf.

Vector 5 tags those sessions ai_agent. They appear in the live feed and the AI traffic report. They do not sit in the default human bounce or the default human goal rate. You can still open the filter and look at them — including which paths they complete.

A headless mannequin facing a security camera in a dark studio
Headless is not a crawler. It is automation with a task.

Goals stay filterable

Name an event, pass optional props, then ask: humans only, humans via AI, agents, or everyone. A “purchase” completed by crawl4ai is a fact. It is not the same fact as a purchase completed by a person who arrived from Perplexity.

A blended conversion rate is how you talk yourself out of a working page.