A month ago, while I was updating my podcast feed I saw a very obscure episode from one of my favourite podcasts Planet Money about BOTUS – Bot of the United States. I checked the episode’s description and was immediately intrigued: a bot performing social listening on Trump’s Twitter account and trading stocks based on sentiment analysis.
Planet Money noticed that Trump is an influencer on market stocks on companies he mentions. They decided to build a bot and invested $1000 to see whether it will work. To make it successful, the team had to feed it with a lot of information and teach it Trump’s language style and how to distinguish between positive and negative tweets. Basically, apply sentiment analysis, present in today’s social listening tools, but highly accurate and tailored to Trump.
And it worked! Sort of.
— Bot of the U.S. (@BOTUS) May 12, 2017
BOTUS has not performed a trade, yet! In my opinion it is possible to generate money, however, there 2 factors which prevent it from doing so:
- Trump – tweeting less about companies than before.
- Time of the day – if the market is not open, no trading can be done.
Following are 2 types of tweets – positive and negative.
See how the USA President showed support to L.L.Bean:
Thank you to Linda Bean of L.L.Bean for your great support and courage. People will support you even more now. Buy L.L.Bean. @LBPerfectMaine
— Donald J. Trump (@realDonaldTrump) January 12, 2017
And when he was straightforward with Macys:
Thx to all the people who called to say they are cutting their @Macys credit card as a protest against illegal immigrants pouring into US
— Donald J. Trump (@realDonaldTrump) July 6, 2015
Which led to:
— Donald J. Trump (@realDonaldTrump) December 4, 2015
Social listening and trading – BOTUS vs You
Does it sound like a job you can do with a social listening tool? Personally, I do not think so as the bot reacts immediately to the tweets, 24/7. But even ‘he’ faces some restrictions, for now at least:
I saw a company name. ✔ I knew the stock ticker (RXN). ✔ I analyzed the sentiment. ✔ (Negative). 👎 But market wasn’t open. 🚫 No trade. https://t.co/zIfXgBgqAn
— Bot of the U.S. (@BOTUS) May 8, 2017
Also, despite BOTUS was fed with companies information one name was impossible to distinguish:
— NPR’s Planet Money (@planetmoney) April 10, 2017
Anyways, it is a very curious story showing us the power of social listening and automation.
Expect some intriguing and thought provoking analysis from us as well using Salesforce Social Studio. Meanwhile, here is the full episode: