• Evan Ryan

Twitter's Bot Problem

Alright Seconds and Cents, Episode Six. We're going to talk about Twitter bots quite a bit today. I'm thinking a lot about Twitter bots here. And we're going to talk about attention and kind of the attention economy overall. It's something that I think about a lot with Lede AI as we write sports articles, and we capture attention for our customers who are newspapers, I think about it a lot. As I'm publishing the podcast. I'm publishing on Twitter, I'm publishing on Facebook, I'm publishing on the blog, and I think about it a lot just in general, in terms of well, what is the attention economy really, and who is deciding what the attention economy pays attention to? And that kind of falls into the Twitter mob and the bot problem. Because I think that the attention economy overall is driven by Twitter. I also think that Twitter is kind of the best example of the attention economy and that you're limited to 280 characters. And so the first thing here about attention is I think that attention right now as it stands. I think it's overall a race to the bottom, especially on Twitter, especially on social media where most people are providing a lot of their attention to brands and they're providing a lot of that attention to content creators. I think it's a race to the bottom because I think a lot of times a short attention span derives content creation that's for a short attention span, which makes it which makes the attention span shorter and then it drives the content creation of the shorter, etc, etc. And I know this to be true overall because of several examples of which I'll only give a few but the first is I was looking into best practices for Pinterest video ads the other day, and I think the Pinterest advertising units are just unbelievable. And I was looking at Pinterest video ads, and they said that the average highest converting Pinterest video ad is at between six and 15 seconds. And that is something that's really interesting because you'd think on Pinterest while people are going there to be inspired right and Pinterest is connecting inspires with the people who Want to be inspired. And so maybe they're going to be willing to be inspired over a longer period of time. That's simply not the case. It's still only between six and 15 seconds. I think about Twitter quite a bit in that Twitter is only 280 characters and they actually raised it, they doubled it from 140. But it's only 280 characters. And basically, you can read a tweet and under five seconds, understand what it is and keep scrolling. And I think about attention quite a bit in Instagram ads. And one day I kind of did a scroll through Instagram ads, or through Instagram and I wanted to count how many ads they show me and they showed me an ad every four to five posts. And that was a remarkable thing for me because if you catch me scrolling through Instagram on a normal day, I have absolutely no idea what ads they've shown me yet Instagram is counting that as an impression, which is a topic for another day. But I think overall here attention is a race to the bottom and that at a large scale short attention spans deprive contract creation. That's for short attention span, which makes it shorter and shorter and shorter and shorter. And you can see this through click Beatty headlines. A recent headline about COVID-19 was that a teenager, a perfectly healthy teenager, was killed by COVID-19. And it was kind of an outrage headline. And then when you read the actual article, well, the teenager had type one diabetes, which puts him in the most at risk demographic for COVID-19. And I thought, well, you know, they wanted to get the clicks because you know, the clicks matter for advertising revenue, they wanted to get the kind of shares on Facebook, the shares on Twitter, that's gonna drive more clicks, it's driving more organic traffic. And to be quite frank, people can read 20 words more than they can read much faster than they can read 2000 and that kind of brings me to the second point, which is, I think there are quite a few, one second Decisions here that are driving attention. And I think the repost on Facebook of that article or a retweet on Twitter of some tweet about canceled culture or somebody is over part hashtag somebody who's over party or whatever it is, those ones that just decisions are driving the algorithm because they're saying to the algorithm, well, we have increased engagement on this, clearly, the increased engagement is keeping people on our platform longer, whether that's Facebook or Instagram or Twitter. And so we're going to continue to promote this to more and more people's followers. Well, what happens with that increased engagement is now all of a sudden, you're promoting something that you didn't even necessarily mean to promote, and you're sending it to more people that you didn't even necessarily mean to send it to. But in one second, a one second split decision, you decided I'm going to hit like, I'm going to hit retweet, I'm going to hit share, I'm going to hit whatever and then that game, the algorithm and to promoting whatever this thing was, that was probably a fairly short attention span, it was probably not the full story, not the full picture. And now it's creating a warped view in somebody's mind over what's going on here. And so I'm thinking quite a bit about these one seconds decisions. And in that I think these decisions are really driving a lot of the way that people view society right now. And I think people view society as a series of one second decisions, but when you have a conversation with somebody about how they feel about any particular topic, what I'm finding is that typically, most people are in agreement. And most people are in agreement, but they're only in agreement after you get 15 to 20 seconds into the conversation and you're able to explain your viewpoint. And most of the time, people are attacking the same problem just with different solutions. And so I think a lot about attention here and I think about it in terms of the Twitter mob because I think people mostly talk about the Twitter mob. I think the Twitter mob is getting a lot of steam right now as it seems to be. canceling a new person every single day. But I'm thinking more importantly, in terms of the bot problem. And I think Twitter has an enormous bot problem. And that when you look at somebody's Twitter account, they might have between 40, or actually it will say, in between zero and 150 followers. And if you look at their Twitter handle, it's gonna be like maybe a word or two, and then a bunch of letters and numbers all in a row to create kind of a unique identifier. And it's clearly a bot. It's clearly a bot that is there to just kind of stir up controversy. But it does that very, very well because it only has to do 280 characters from a technical perspective, building a Twitter bot, that only engages with kind of enraging content. And building a Twitter bot that is kind of keeping people on Twitter a lot longer and is helping divide the opinion is very easy to do. It can be done in just a couple of days. And then once it's done in a couple of days, that can be recreated over and over and over again. And, you know, they talk a lot about people, Twitter talks a lot, or Twitter has kind of a big controversy every several months where they will delete the account, or they will suspend the account of somebody who had a 20,000 to 30,000 or 40,000 person following, but they said something that broke their terms of service or violated the rules. And so Twitter suspends or bans the account. Everybody gets mad for a couple of weeks up to a month, and then they kind of forget about it and they go on with their lives. And for that person who just got banned or suspended. That person has to start over at zero and they have to get their entire following again. But with a bot a bot doesn't necessarily care about the following. all they're doing is engaging with comments, they're retweeting, they're trying to play the algorithm play the Twitter trends, and they're trying to really just stir up controversy that's going to drive engagement that's going to divide people's thinking. It's going to Find that thought process. And with a bot because they're not trying to create a following, you can create another bot in a heartbeat. Once you have the code to do at once, you can do it as many times as you want to. It's one of the really scary things about thoughts in general is that you can knock out one of the bots, but you didn't knock out the source. Whereas if I had broken Twitter's Terms of Service, they could have banned my account and now I either am banned for life, or I have to start over at zero. And with a bot that kind of incentive that negative incentive to run away from pain doesn't really apply and it doesn't really matter. And I'm seeing that there are a ton of Twitter bots right now. I think I mean, 10s of thousands of tweets per hour are bots. I think the bots are a serious problem for Twitter and that Twitter is kind of creating a global conversation. They are allowing people to share their opinions, share their short form opinions, and then allowing people to share information But I think that the bots are really doing something here and that they are playing a huge role in the Twitter mob. They're playing a huge role and beating Twitter's algorithm in terms of what drives attention. What drives what I see when I go onto my timeline, and I think the Twitter bot problem is going to be a really serious issue for for Twitter as it becomes more and more high profile, especially during the election season where bots in the 2016 election, were crushing it on Facebook, and they're very much able to sway the public opinion on Facebook and then Mark Zuckerberg how to go in front of Congress. I can see that absolutely being a possibility in the 2020 election, in that the social cons the social ability. The social capital of these bots, is that they're showing up and replies they're showing up in comments. They are clearly a bot, but Twitter is hiding the second half of the user handle. And a lot of people are just reading to see the reading the comments to see what people say, but they're not actually verifying the sources. And so now you have everybody being a journalist, but nobody being verified. So we'll see what happens over the next coming months we'll see what happens over the 2020 election cycle and what happens afterward. But I would not be surprised at all if these million 10 million or 100 million or however many bots exist on Twitter, really end up being a PR issue for Twitter once people are able to determine who is a bot and who is not. And people are able to kind of cancel the bot and start not read over the bots.

The above is an AI transcription of Evan’s Seconds and Cents podcast.