Popunder traffic opens behind the publisher page after a user interaction. Done well, it delivers high-volume, low-cost awareness and cheap conversions for impulse offers. Done badly, it produces bounce rates above 90% and support tickets about "pop-up viruses."
Set the right expectation
Popunder is interruption inventory. Users did not ask to see it, so the landing page has about one second to explain why they should stay. If your offer needs a seven-step explanation, this is not the format — use native or search instead.
Control the flow, not just the bid
Three settings decide whether a popunder campaign is viable:
- Frequency: one popunder per user per day is the ceiling for most offers. Two or more guarantees fatigue and chargebacks.
- Device split: mobile and desktop behave like different channels. Split campaigns so you can cap and bid independently.
- Site quality: exclude domains that produce clicks without conversions after 5,000 impressions. Keep a blocklist per GEO.
Watch the redirect chain
The classic failure mode is a landing page that loads six trackers and a modal before the offer appears. Every extra hop loses users on slow connections. Host the pre-lander on the same domain as the tracker, defer non-critical scripts, and test on a throttled 3G connection — not on your office fiber.
Know which offers fit
Popunder performs best with low-consideration, high-intent offers: utilities, downloads, streaming trials, sweepstakes, dating, and simple mobile games. It struggles with B2B SaaS, high-ticket retail and anything requiring trust before the first click.
A simple test structure
Start with one vertical, three GEOs and two creatives. Cap spend at 3× your target CPA per GEO. If a GEO has not produced a conversion by then, pause it and move the budget. Scale the winners in 30% steps so you can see where efficiency breaks.