Professional Development
Cybersecurity / Elections / Media / Politics / Social media

The Washington Post has a Messenger bot tracking how people feel about the election. Why?

The emoji-heavy bot is a new way to engage audiences. Do you feel ? or ? today? Or maybe you feel ??

Hey, Feels. (Photo via Facebook Messenger)

Five more days. Election 2016, which somehow feels like it’s been going on forever, will be over in five more days.

A "Feels Report." (Screenshot)

A “Feels Report.” (Screenshot)


I’m keeping close tabs on the end of the election due to excitement, sure, but also thanks to a Washington Post Facebook Messenger bot that counts down the days for me every morning.
The bot, affectionately called Feels, works like this — every evening it asks users how the election is making them feel. Users can choose from a set of five different (and differently interpretable) emoji responses ranging from a sort of laughing face to a crying face or an angry face (? ? ? ??). Users can also choose to expound upon their emoji choice with words. The next morning, Feels greets users with a jaunty election countdown and a “Feels Report” showing a graph of all responses.
It’s a simple, engaging experience and the anecdotal responses are certainly curious. Still, from a data perspective, anything that happens in Feels isn’t particularly noteworthy. First, there’s all kinds of selection bias. And second, asking people to tell you how they feel might actually serve to influence that feeling (essentially — I’m annoyed that you’re asking me how I feel about the election and thus feeling more annoyed about the election itself as well).
So why is the Post doing this? I called up Joseph Price, senior product manager at the Post, to chat about Feels and the newsroom’s broader bot strategy.
Feels, it happens, is just the latest in a string of experiments. In July there was a Post news bot, complete with headline-type announcements and some Olympics functionality. Then, about a month ago, the company experimented with an SMS bot. And through these experiments, Price said, he’s learned some valuable things.
“I spend all my shower time thinking about bots and what they mean,” he said, laughing.
For example — it’s important to set clear “fence posts” around your bot. Users, Price said, tend to have very high expectations of what can be achieved through a chatbot, but the technology just isn’t there yet. So if, as a creator, you fail to very clearly define what the bot is capable of, people will inevitably be disappointed.
With Feels, the Post is explicitly trying to manage user expectations. And from Price’s perspective it’s working pretty well — he declined to comment on how many users the bot has total, but said that one third of those users interact with the bot every single day.
Feels has certainly seen some interesting changes in the feelings of users. For example, from when the bot launched (after the third debate) through most of last week, the majority of respondents felt ? about the election. But then, after the new email revelations on Friday, the majority of users were suddenly feeling ?. Who know what this final week will bring — but Feels will be around to listen.
Price told Technical.ly that Feels will continue to prompt responses through Election Day and perhaps for one day after — the team will then put together some sort of wrap-up report that will be “playful and interactive.” At the end of the day, all this is about trying to “deliver a delightful experience for our users,” he said.

Companies: The Washington Post
Engagement

Join the conversation!

Find news, events, jobs and people who share your interests on Technical.ly's open community Slack

Trending

DC daily roundup: Tyto Athene's cross-DMV deal; Spirit owner sells to Accenture; meet 2GI's new cohort

DC daily roundup: $10M to streamline govt. contracting; life sciences might dethrone software; Acadia's new $50M

DC daily roundup: the DMV's VC cooldown, SmartSigns for safer driving; Rep. Schiff's AI copyright bill

Will the life sciences dethrone software as the king of technology?

Technically Media