I ran observations with video editors to see how they work and how they create captions. Combined with interviews this built up a picture of their pain-points and revealed how automation could help them best. Although creating captions and subtitles is time-consuming this research revealed the creative control journaists wanted to have over their video content.
Mapping workflows across systems
To help explain the complicated tasks journalists have to complete, I mapped out their workflows. This helped developers to understand how the tool needed to fit in with the already complex ecosystem of tools and identified how best to speed up the process, while ensuring editors had as much creative input as before.
Sketching ideas with users
Once I had identified some opportunities to help editors, I ensured they were involved as much as possible in the design process. Sketching the tools and it's interactions with them made sure we didn't make any assumptions about how they work and helped them to consider new possibilities.
Involving developers in these early sessions also helped generate ideas and ensure the team were aligned.
Prototyping complex interactions
To test detailed interactions, I used Axure to create interactive prototypes. Eventually the prototypes lost their ability to drive insights as we got closer to building the tool, but were useful to specify complex interactions to remote developers in slack without resorting to extensive documentation.
The solution aimed to make captioning as simple as creating a transcript in a word doc. We relied on the known behaviours of text editors which allowed captioning to become a much more flexible, creative process, allowing editors to compose, edit captions and publish in many formats.
We also aimed to give them the power of a video editor timeline and text editor in-one, to make the most uncluttered captioning interface available. Knowing the exact needs of the users made it possible to strip out all but the required functionality and make for a simple, intuitive interface.
The tool is now used by BBC teams and speeds up subtitling of longer videos when the speed of getting an accessible video online quickly is important.