4 ms·

Super useful observation that the metrics are the addiction. > Metrics are an addictive drug. Twitter is literally an incremental game. This comes up in the N

Super useful observation that the metrics are the addiction. > Metrics are an addictive drug. Twitter is literally an incremental game. This comes up in the Nir Eyal "Hooked" model, but Eyal spends chapters outlining it, where the author here captures it as the main thing. Watch the number go up, get little hits of dopamine. It's gambling, trading/investing, SEO, literally anything with an 1/f distribution. (http://www.scholarpedia.org/article/1/f_noise http://www.scholarpedia.org/article/1/f_noise) I'd wonder if you could drive engagement on an app by spacing out the arrival of user alerts and feedback so they only reached the user with a delay/period that was hidden, but detectable in the same sense of anticipation music creates. Seems like an unethical experiment, and probably hidden in the annals of FAANG companies.

carrolldunham 6y ago
>I'd wonder if you could drive engagement on an app by spacing out the arrival of user alerts I'm pretty sure twitter does this, for this reason. When you login you have no notifications. Aw. But then after you click around a bit... ding ding - you get the ones that happened while you were logged out. You could say it has to load them in the background. I don't buy it. The sensation is too close to what you described. A pause then a hit
- quotemstr 6y ago
  More likely, the delayed notification thing was originally some kind of bug, and the bug fix tanked metrics, so they reverted the fix. I'm sure there was no conscious and deliberate effort to psychologically manipulate the user, but laser-guided metrics hill-climbing can produce the same thing without any one person deciding to do it.
  - nate_meurer 6y ago
    > I'm sure there was no conscious and deliberate effort to psychologically manipulate the user Totally. Social media corporations would never do that.
    - quotemstr 6y ago
      Seriously, I don't think they would. The individual people who work there are good and decent men and women. That their aggregate activity has the emergent effect of skinner boxing people doesn't make social media company people cackling comic book villains.
      - _jal 6y ago
        You don't have to be a comic-book villain to do bad things. Also, you seem to be asserting the people at FB can operate the machine without being aware of what it is or what it does.
      - nate_meurer 6y ago
        There's a lot of space in between "all employees are good and decent" and "cackling comic book villains". Is it possible that reality is somewhere in between those two extremes?
  - exikyut 6y ago
    Oooh. I would like to subscribe to more second-order applications of hanlon's razor please.
magneticnorth 6y ago
I expect this has been studied, and likely is already being done. Maybe not directly in notification timing (though I wouldn't be surprised), but I expect this is part of what goes into newsfeed rankings - it optimizes to keep the reader scrolling, but also can optimize for rewarding the poster by spacing out their "likes" for optimal dopamine rewards.
- theja 6y ago
  I had worked on an online machine learning method[0] last year that assumes priming effect (similar to the effect you and the GP describe, i.e., the timing effect) and tries to exploit it. With respect to infinite scrolling/ranking, I have also worked on a simplified model (unfortunately does not capture spacing of likes but captures fatigue based on relevancy) and and algorithm to work with it.[1] [0] https://arxiv.org/abs/2006.10356 https://arxiv.org/abs/2006.10356: Learning by Repetition: Stochastic Multi-armed Bandits under Priming Effect [1] https://arxiv.org/abs/1901.07734 https://arxiv.org/abs/1901.07734: Thompson Sampling for a Fatigue-aware Online Recommendation System
RhodoGSA 6y ago
>The widespread occurrence of signals exhibiting such [1/f] behavior suggests that a generic mathematical explanation might exist From the article you linked... That's kind of scary. To be honest, one of the things we have done with ads as the medium of exchange with some IT services is inadvertently dedicated some of the smartest minds of our generation to "Figure out the addiction function". Look what we've done in 26 years, imagine what we can do in 100 years.
ryantgtg 6y ago
In my case, I don’t care for the metrics - the stats have never interested me, be they on my site’s analytics or on social accounts - but I get very addicted to the comments. I only go on Twitter for like 5 minutes a week because it’s very difficult for me not to reply to people, and when I do my mind gets consumed by thoughts about them replying back and how I’ll respond to that, etc. I spent two hours painting a fence last weekend, and for most of that time I was simply daydreaming about imagined responses to a single subtweet. What a waste! I don’t have any social media apps or notifications on my phone, so it’s not like I’m constantly picking it up and seeing or awaiting responses. Instead I’m just wasting my brain power imagining arguments with strangers.

carrolldunham 6y ago

>I'd wonder if you could drive engagement on an app by spacing out the arrival of user alerts I'm pretty sure twitter does this, for this reason. When you login you have no notifications. Aw. But then after you click around a bit... ding ding - you get the ones that happened while you were logged out. You could say it has to load them in the background. I don't buy it. The sensation is too close to what you described. A pause then a hit

quotemstr 6y ago
More likely, the delayed notification thing was originally some kind of bug, and the bug fix tanked metrics, so they reverted the fix. I'm sure there was no conscious and deliberate effort to psychologically manipulate the user, but laser-guided metrics hill-climbing can produce the same thing without any one person deciding to do it.
- nate_meurer 6y ago
  > I'm sure there was no conscious and deliberate effort to psychologically manipulate the user Totally. Social media corporations would never do that.
  - quotemstr 6y ago
    Seriously, I don't think they would. The individual people who work there are good and decent men and women. That their aggregate activity has the emergent effect of skinner boxing people doesn't make social media company people cackling comic book villains.
    - _jal 6y ago
      You don't have to be a comic-book villain to do bad things. Also, you seem to be asserting the people at FB can operate the machine without being aware of what it is or what it does.
    - nate_meurer 6y ago
      There's a lot of space in between "all employees are good and decent" and "cackling comic book villains". Is it possible that reality is somewhere in between those two extremes?
- exikyut 6y ago
  Oooh. I would like to subscribe to more second-order applications of hanlon's razor please.

magneticnorth 6y ago

I expect this has been studied, and likely is already being done. Maybe not directly in notification timing (though I wouldn't be surprised), but I expect this is part of what goes into newsfeed rankings - it optimizes to keep the reader scrolling, but also can optimize for rewarding the poster by spacing out their "likes" for optimal dopamine rewards.

theja 6y ago
I had worked on an online machine learning method[0] last year that assumes priming effect (similar to the effect you and the GP describe, i.e., the timing effect) and tries to exploit it. With respect to infinite scrolling/ranking, I have also worked on a simplified model (unfortunately does not capture spacing of likes but captures fatigue based on relevancy) and and algorithm to work with it.[1] [0] https://arxiv.org/abs/2006.10356 https://arxiv.org/abs/2006.10356: Learning by Repetition: Stochastic Multi-armed Bandits under Priming Effect [1] https://arxiv.org/abs/1901.07734 https://arxiv.org/abs/1901.07734: Thompson Sampling for a Fatigue-aware Online Recommendation System

RhodoGSA 6y ago

>The widespread occurrence of signals exhibiting such [1/f] behavior suggests that a generic mathematical explanation might exist From the article you linked... That's kind of scary. To be honest, one of the things we have done with ads as the medium of exchange with some IT services is inadvertently dedicated some of the smartest minds of our generation to "Figure out the addiction function". Look what we've done in 26 years, imagine what we can do in 100 years.

ryantgtg 6y ago

In my case, I don’t care for the metrics - the stats have never interested me, be they on my site’s analytics or on social accounts - but I get very addicted to the comments. I only go on Twitter for like 5 minutes a week because it’s very difficult for me not to reply to people, and when I do my mind gets consumed by thoughts about them replying back and how I’ll respond to that, etc. I spent two hours painting a fence last weekend, and for most of that time I was simply daydreaming about imagined responses to a single subtweet. What a waste! I don’t have any social media apps or notifications on my phone, so it’s not like I’m constantly picking it up and seeing or awaiting responses. Instead I’m just wasting my brain power imagining arguments with strangers.