Facebook
Auto Captions

DESIGN

I worked as part of an augmented member of the Core Ads Monetization Team at Facebook. The CAM team is responsible for the consumer facing ad experience at Facebook. While there I designed an auto caption feature for video content in Ads Manager.

CREDITS
Design Direction
Jessica Watson
Additional Designs
Facebook Core Ads Monetization Team

GOALS

The project goal was to provide all advertisers with the ability to caption their video content quickly and easily within Facebook’s already existing video ad creation flow. The MVP of this feature needed the ability to edit caption text and duration as well as a way to inform users when captions are being generated.
An additional feature request was to provide the user with a visual representation of confidence in the auto caption’s accuracy. These notifications needed to be simple and flag any captions for mistranslations, spelling errors or unintelligible audio sections in the source video.

CONSIDERATIONS

The inclusion of the auto captions feature required flows to be built within the existing video ad creation structure. The feature needed to be unobtrusive to the rest of the flow while still informing the user when captions were being generated, including success and fail states.

All business accounts on Facebook would be given access to auto captions, which meant the editing capabilities needed to be simple enough for users of any skill or experience level. Researching other caption editors showed a lack of consistency and innovation from online options while film editing software, was too powerful. The design needed to find a balance in editing capability so first-time creators and power users alike felt comfortable in the workflow.

SOLUTIONS

The button was designed to stay compact and match Facebook's existing interface while informing users with a "generating" bar along the bottom and "estimated time remaining" shown below. Once captions were generated the user would be presented with the option to review the captions, which would open the editor.

The auto caption editor allowed users flexibility in how they edit their captions with a timeline for users more familiar with film editing software and a note pad style section for a more traditional caption text editor feel.

Confidence was reflected with simple highlights of yellow or red in the background color of all questionable captions as well as flagged on the video's timeline for an overall view. Yellow would be used to highlight any "warning" captions that had questionable spelling or syntax, while red was used to highlight any egregious errors in caption accuracy for any reason.