Facebook · Video Off-Platform
I improved the web experience for people who visited Facebook for video, then encouraged them to open the app or log in when they wanted to continue.
Problem
Video was a major content surface, but logged-out visits were shaped around access controls. That made it difficult for unfamiliar people to understand Facebook’s value before being asked to log in or open the app.
Context
The visit was not one page or one prompt. It was a sequence of layers, and each layer had a different job before the next one could work.
Solution and ownership
I worked across the logged-out page and its conversion moments, deciding what the experience should help someone do first and which next step made sense on mobile and desktop web.
People arrived for a specific video. The page needed to make that visit worthwhile before Facebook asked them to change context, authenticate, or install anything.
Keep the video and its immediate context available instead of turning the visit into an access gate.
Explain what logging in or opening the app adds so the action has clear value to the visitor.
I audited existing prompts, proposed experiment directions, and developed patterns for app entry, saved-account login, and reauthentication. The goal was not one universal prompt; it was a useful path for the visitor’s current context.
Saved-account entry and reauthentication gave returning visitors a lower-friction way back without forcing the strongest ask first.
Opening the app remained the most ambitious outcome, so it belonged where continuing there offered a clear benefit.
How I framed the decision
Two ladders were shaping the same decision. The incrementality ladder defined what Facebook wanted someone to do. The aggressiveness ladder defined how forcefully the interface asked. Combining them made the tradeoffs visible in one model.
Where it landed
The work expanded success beyond app opens. Web login became a conversion metric too because it gave returning visitors a lower-friction bridge into Facebook. The combined framework made both the outcome and force of each ask explicit.
App opens remained the strongest acquisition outcome, while web login became a tracked conversion rather than a secondary interaction.
Numeric results pending verified evidence.
The combined incrementality and aggressiveness model gave teams a reusable way to compare prompts without collapsing every visit into “ask” or “do not ask.”
Reflection
People should understand what opening the app or logging in unlocks. The prompt works when the benefit is concrete, not when it only names the action.
The same ask can help or interrupt depending on when it appears. I learned to place it after the visitor had received enough value for the next step to make sense.