Retro: FT Brexit polltracker
Probably the most successful1 thing I worked on at the FT was conceptually one of the most simple The EU Referendum Poll tracker. I think it illustrates the value of 1. Getting the basics right 2. Organisational trust and support and 3. Luck.
The basics
What are the basics in this case? We wanted to make a page that told the reader how the polls currently stood and then let them get on with their day or continue their journey on the FT site. We hoped people would revisit the page often so minimal preamble before the actual answer to the question of what the polling was suggesting was key. From there, giving people a clear onward journey once their main goal was met. Here's how we conceptualised the page structure...
- A sentence long a summary of the state of play
- The headline numbers visualised boldly, emphasising the uncertainty gap
- Links to the latest FT stories
- Full polling visualisation and historical data for readers who are into the polling stuff per se
The simplicity and focus meant that our earliest prototype was very close to the final product (annoyingly I can't find the screen grab of the first version in my archive)
In addition to page structure we needed the page to be fast, search engine friendly and bang up to date. The page was simple but we spent a lot of effort on performance and making sure we structured everything in the clearest way possible, simple HTML and CSS, everything pre-rendered, no client side javascript required. This approach is good for both humans and for machines (search engines etc.) looking at the page. 2
Updates were handled by adding new polls to a Google spreadsheet which the Nodejs/Express app would fetch data from when a page was requested. For weeks and weeks my colleague John Burn Murdoch diligently added each poll as soon as it was published.
Organisational support
I believe one of the reasons the page was sucessfull was that our team was allowed to develop it with a clear user focus and the FT's broader editorial team (in the form of then ft.com editor Robert Shrimsley) trusted us enough to let us get on with it without imposing other requirements that might dilute that focus and lead to a less compelling product. The process of building that trust was the work of several years up to this point and it was gratifying to see it pay off. I think one of the reasons the FTs team continues to be world leading is that they're trusted to know what they're doing, their expertise is valued within the newsroom.
I really do think that organisational trust is one of the key things separates the best editorial graphics teams from the rest. There's obviously an element of knowledge and ability but that's largely a recruitment and training issue, there are loads of really talented data-vis practitioners around, it's not magic or some kind of super power, it's something that can be taught. It's the organisation that people find themselves in that allows them do their best work.
Luck
I guess there's a sense in which you make your own luck. We'd spent a few years at this point building up trust within the organisation, learning to work as a team, and developing tools to turn projects like this around quickly. That it turned out the polling was knife-edge close throughout the campaign was really the thing that drove the pages success though and that's something we couldn't control.
Notes
- The brexit poll tracker was the most popular page on the FT site by a huge margin for weeks on end, often doubling the unique users of the next highest performing story.
- looking at the page now I feel we could have done a better job making the SVGs comprehensible to screen readers i.e. better description tags
<desc>. Chart accessibility is something that I think needs more documentation of best practices etc.