Josh Anderson

Information Architect

This is a Successful Tweet

Warning: the subject matter of this blog is intense. It’s about my experience with finding footage of real-life violence on Twitter.

I used to have Twitter and Facebook accounts. After much time and effort I was able to delete them. True story: I quit alcohol before I managed to quit social media.

But sometimes I come back to Twitter. I can click open a new tab, type nothing more than “t” into my Google Chrome search bar, and a list of my bookmarked or otherwise frequented Twitter user page URLs will unfurl. I click a favorite user. Someone whom I can trust to act as a sort of news feed for myself.

Nowadays when I’m on Twitter, I’m never signed in. I arrive incognito, which lets me see Twitter’s sidebar through the eyes of a stranger to Twitter. I swing through the links on that sidebar. I think Twitter calls them “trends.”

Turns out I still haven’t been able to quit Twitter even though I managed to close my account. I don’t think it would be outrageous to refer to Twitter as a behavioral addiction I have.

Today on Twitter, I saw this viral Tweet:

A Tweet reading "watch this shit shit escalated fast as shit." 3.2K replies, 39.7K retweets, 31.8K likes. Sensative content warning.

With an intro like that, I just had to see what it could be. Click.

There’s a snowy, suburban drive onscreen. Reminds me of the kind of place I grew up in.

A Tweet reading "watch this shit shit escalated fast as shit." 3.2K replies, 39.7K retweets, 31.8K likes. First frame of video.

Well, I shouldn’t mislead. Immediately I knew what this video was. I had seen the headlines.

News stores about 3 dead in show shoveling dispute

So I’ll be clear: I did know what I was getting myself into.

Twitter conversation about who is responsible for what they watch

I understand both sides; I really do. But I take responsibility.

As soon as this video began, I expected that I was going to see video footage of the shootings. Every part of the interface assured me: “This is a good Tweet.” I agreed and kept watching. I could have stopped the video and walked away at any point. I could have. I probably should have. But I didn’t.

Judging from the view count to the right of the play button, at least 1,500,000 others took the same user journey as me.

Successful twitter metrics
Every part of the interface assured me: “This is a good Tweet.”

Two minutes and twenty seconds later, I had watched high definition murder footage. Well, I guess the video wasn’t that great, but somehow the clear sound was enough to imprint this into my memory as “high definition.”

I think that’s what made this video get under my skin so much, to bother me enough to feel the need to write this blog post: I didn’t expect the content to be this “high quality.”

Of course, what did I think it was going to be, right? Maybe I didn’t expect to hear so clearly the final words that she ever did.

I scrolled down to the replies and found some cleverly re-purposed memes.

Lisa Simpson shocked meme

It’s actually sort of a funny, isn’t it? It’s messed up, I know. But still: 10.6K hearts makes sense.

"If 'I tell you what' was a person"

Eh, that one’s a little weaker. 724 hears seems appropriate.

Other people took the time to sprinkle on one last slur over the woman’s dead body.

"She argued like a Karen til her last breath. R. I. P."

To my relief, there did seem to be some level-headed people in the mix who put an appropriate framing around the scene.

"How is yall in the comments laughing yo are people really this desensitized"
"nah y'all gotta stop sharing videos of people gettin' killed. this ain't content."
"People are fucking horrible."

Am I a prude now, for agreeing with their sentiment? I guess some people would call me a prude. I’m okay with that.

Later, after this footage ends, the man with the rifle kills himself, leaving 3 people dead total. I think subconsciously part of the reason I stayed glued until the end was to see how much of the murder-suicide made it onscreen.

The video concluded. Twitter got what it wanted from our transaction. I felt numb.

Successful Twitter metrics

39,700 retweets? That’s impressive. And over three thousand comments? Wow, that’s incredible.

Twitter looks at numbers that high and sees:

  • Electronic hallways of conversation blossoming from what must have been a striking, thought-provoking two-minute seed of an idea
  • Friends, family, and peers staying connected despite pandemic social distancing
  • Democracy in action
  • The marketplace of ideas

People are engaging with content. Clicks are up. Time on the page is up. This is a successful Tweet. We want more of this. The world would be better with more of this.

The medium is the message. The content of the Tweet is irrelevant. It is the metrics that matter: the comments, the retweets, the likes, the engagement. No can argue this Tweet isn’t engaging.

Does it matter what the people are saying in the comments? Only if they’re using #hashtags. Then we can capture information on what users who engaged with this Tweet might want to engage with next.

If pink bubbly hearts are up, advertisers click “buy.” The people have spoken: the people “like” this.

This is a successful Tweet.

This is inhumane.

I had a hard time focusing for the next hour or so. That frustrated me—I had wanted to be more productive. Well, I did this to myself, I guess.

I don’t think Jack Dorsey et al. deliberately designed a trauma & outrage information superhighway. Not all designs are deliberate. Actually, you could probably say that most aren’t.

An information architect looks at the context, content, and users present in the scene—in the other words, the elements of the experience that a quantitative social media algorithm cannot break down into clean 0s and 1s and sort into a database; the metrics that not only are not, but—try as they may—cannot be Twitter’s success metrics.

When I look at the context, content, and users of this Tweet, I see disaster and tragedy. I see human beings in pain. An algorithm cannot see what you and I see. Maybe one day it will. Then it would recognize that is isn’t content that I want to engage with. Then it wouldn’t show it to me. Hopefully.

Link to the Tweet:

I don’t recommend that you watch it. In fact, I strongly recommend that you don’t watch it. But let no one call it fake.


Recent Posts


Upcoming Speaking Dates

STC Rochester Webinar
June 27, 12-2PM EST
“Prompt Engineering”

More info coming soon.

LavaCon 2023
October 14-17, 2023
San Diego, CA
“Microcontent Pipeline to Chatbots and Voice Assistants”

Use microcontent to put your team in the driver’s seat to the next level of content publishing for intelligent chatbots. By now reality has revealed that ChatGPT is not ready to write our product documentation for us. With no control over what the publicly-trained models scrape from our websites and blogs and how it assimilates our documentation with other less-authoritative sources, we’re left to wonder where we go next. If we want reliable bots, we’ll need to train our own models and deploy them to our staff and our customers. Let us show you what we’ve learned so far.


My Spotify Playlists

Create a website or blog at

%d bloggers like this: