On the night of 23 June, Financial Times journalists David Blood and Petros Gioumpasis headed to London’s Millbank Tower, where Leave.EU campaigners were gathered to await the arrival of Nigel Farage. Farage would later stand in front of a crowd there and declare that “dawn is breaking on an independent United Kingdom”, after it became clear that Britain had voted Leave in the Brexit referendum.
David and Petros were there to conduct the FT’s first experiment with livestreaming 360-degree video. We had never done one before and wanted to figure out both the logistics of what we needed to do, and to answer the editorial question of when live 360-degree videos are better than traditional videos. The experiment involved simply trying to livestream the exact moment when it became clear which side had won.
It was not an experiment we spoke a lot about internally afterwards, nor shared with the wider journalism community, for one simple reason – we failed to achieve what we set out to do, and it is embarrassing to talk about failures.
The problem is not a lack of willingness to experiment, but the difficult step of showing enough vulnerability to share and learn from our failures as well as successesRobin Kwong, Financial Times
Making the most of a failed experiment
The situation was so chaotic and the press scrum so packed, that all we managed to film was the back of many reporters’ heads from a corner of the room where Farage gave his victory speech. Viewers saw even less. The Wi-Fi connection was so bad that the screen was just pixelated blocks of grey.
But while we failed to produce a watchable live stream, the experiment was a success – we learnt eight specific lessons for next time we will want to attempt a similar project:
Have a two-person crew fully dedicated to the live stream. One can act as a ‘scout’ for the best location while the other handles the equipment.
What makes live video especially difficult: good camera placement is key and it is hard to predict where the action will occur in live situations.
It is hard to set up the camera in press scrums because the camera has to be on a tripod and be physically connected to the laptop.
It’s hard to judge when to tell people to tune in. Too early and no action is happening, too late and most people will miss the key moment. This format favours longer-lasting streams.
Wi-Fi bandwidth matters. A lot.
If the surrounding area is busy with activity (even if it just the press milling about), 360-degree video is a lot more compelling than traditional video, because the viewer can explore and has a sense of presence in the place you are filming.
But the flip side of that is that 360-degree video is a lot less compelling when not viewed in real-time. Recorded 360-degree video is a lot more like traditional video in terms of what makes it compelling to watch.
It is probably best, in most cases, to keep the live chat disabled, given the general quality of commentary on YouTube (unless you want to add moderating the live chat to your already long to-do list).
I finally spoke about this experiment in a session I facilitated at SRCCON in Portland, Oregon, on 29 July, because I think talking about our failed experiments is an important part of getting more out of doing them.
My suspicion is that there are many such experiments happening in newsrooms all across the world, but no one talks about the ones that ‘failed’. Among the digital news community, the problem is not a lack of willingness to experiment, but the difficult step of showing enough vulnerability to share and learn from our failures as well as successes.
What do good newsroom experiments have in common?
During the session, which was attended by about 40 people, we also discussed common traits of good experiments (see some of these illustrated with cat GIFs in my slides):
They aim to solve a problem.
They have an element of risk, for example 'this might not work'.
Failure = we didn't learn anything from doing it.
They are incremental: build up knowledge or expertise.
They are talked about and discussed afterwards to share knowledge.
Attendees then split into small groups and used a few simple prompts to design their own experiments to conduct back at their respective newsrooms (more detail on the prompts can be found in the session notes).
The prompts, which boil down to four easy steps to design your own experiment, were:
What do you notice? What intrigues you? What problems need solving?
What are some possible ways of solving that problem? What could you do differently?
How will you know if this new way is better or not?
How will you let others know about what you've learnt?
Each person spent about ten minutes being interviewed by the rest of the group on those four questions.
I encouraged attendees (or anyone interested) to share the results of their experiments with me either by email, or by tweeting with the #thismightnotwork hashtag.
This might not work, but I am monitoring that hashtag and looking forward to hearing about failed (and successful) experiments over the coming months.
Robin Kwong is special projects editor at the Financial Times, where he runs newsroom experiments and helps coordinate large-scale coverage. Previously, Robin was an interactive data journalist at the FT. He can be found on Twitter @RobinKwong and on his blog, robinkwong.com.
Free daily newsletter
- Tip: Advice for finding the story in your data set
- Tip: Take note of this advice for measuring the success of your Facebook Lives
- App for journalists: Google Street View, for capturing 360-degree photos
- The last ten years and the next – Preparing for what's to come in journalism
- NZZ is developing an app that gives readers personalised news without creating a filter bubble