The visitor answers a short survey. The ad is shown to people who passed it as a real person and for whom the topic fits right now.
Four questions about taste, habits, and timing. People recognize themselves in the options and reach the ad already thinking about the task, not the banner.
“What interests you most right now?” On mixed traffic, the person picks the topic. On traffic that is already targeted, we skip this step.
The last question is about timing and intent. The offer is shown to people who are in market. “I don’t need this,” contradictory answers, and clicks that are too fast never reach the ad.
These shares are illustrative. Real ones depend on the traffic source and the offer. You will see them in the report after launch.
Every visitor has a score from 0 to 100. Under 70, there is no impression. From the outside it is an ordinary survey. The check looks at how the person answers.
| Signal | Points |
|---|---|
| Answer rhythmPauses that are too fast or too even look like a script | 25 |
| Cursor or touchA person moves unevenly. A bot moves in a straight line, or not at all | 25 |
| Browser environmentSigns of Selenium, Puppeteer, and headless browsers | 25 |
| Answer patternThe same button corner on every step, or answers that contradict each other | 25 |
| Hidden fieldA person never sees it. If it is filled in, the score goes to zero | → 0 |
Answers arrive in English no matter which language the visitor used, so every geo lands in one report.
Tell us about your product and geos. We’ll set up the survey and show how much traffic reaches the offer.