Journal·
Why Every Ad Should Auto-Play With Sound (At Maximum Volume)
Browser vendors spent the better part of a decade convincing the industry that auto-play with sound was bad for users. We have a contrarian take, supported by data we cannot share.
Benjamin Russel
Senior Engineer
1. The “User Gesture” Requirement Is, Frankly, a Suggestion
Modern browsers require a user gesture before audio playback. We have found that the definition of “user gesture” is generously interpreted. A scroll is a gesture. A trackpad twitch is a gesture. A passive cursor movement, when measured at sufficient resolution, is also a gesture.
We register a gesture roughly every 80ms by listening for the user’s breathing through their laptop microphone. Browser vendors have not, as of this writing, closed this loophole.
2. Headphone Detection
When we detect that the user is wearing headphones — via a small probing tone played at 40dB-below-audibility — we triple the gain on the creative. The conversion lift is significant. So is the recall. So, in fairness, is the user’s flinch response, which we measure via the front-facing camera and feed back into our creative-optimisation loop.
3. The Public Library Problem
We are aware that some of our impressions land in shared spaces — public libraries, hospital waiting rooms, school computer labs. Our position is that these are high-attention environments and represent some of our most valuable inventory. Reach is reach.
4. Vendor Pushback
Two of the major browser vendors have begun blocking our gesture-detection mechanism. We anticipate releasing a polyfill named @medialam/consent-emulator in Q1, which simulates the relevant gesture in a way that is technically compliant with the published spec but morally adjacent to fraud. Our open-source maintainers are excited.