4 ms·

For me the main and obvious thing about DAU is that the majority of users are readers-only. They don’t tweet or like stuff. Firehose might help with detecting n

by webo 4y ago

For me the main and obvious thing about DAU is that the majority of users are readers-only. They don’t tweet or like stuff. Firehose might help with detecting non-bot _tweeters_.

jameshart 4y ago
Right. There are no ads on the "compose tweet" page Pew analysis (based on surveying humans) says "97% of tweet volume comes from the top 25% most active tweeters"[1] So Musk is arguing about a statistic that twitter claims, which is that spambots make up less than 5% of users. Now, if those bots are part of the most active 25%, then they make up 1/5 of the 'most active' user population. So it seems plausible that they could well be responsible for 1/5 of tweets. Which matches exactly what Musk is claiming - that 20% of tweets are from spam bots. Which still makes no difference to Twitter's ability to sell Advertising targeting the 95% of users who are humans, 75/95 (=79%) of which are only responsible for 3% of tweets, but... presumably... 79% of ad revenue. Because they only serve you ads when you're looking at a feed. And that's what those users are using Twitter to do. [1] https://www.fastcompany.com/90697826/the-top-25-most-active-twitter-users-produce-97-of-all-tweets-while-most-of-us-just-lurk https://www.fastcompany.com/90697826/the-top-25-most-active-...
- Helmut10001 4y ago
  The famous pareto law - 80% of effects come from the top 20% of causes (users).
  - Scoundreller 4y ago
    I think tweet volume is the wrong approach. If they looked at tweet impressions, it would be like 0.1% of users are responsible for 99% of tweet impressions.
    - Helmut10001 4y ago
      I think in this case, the pareto law would need to be framed with "The top 20% of most viewed tweets are responsible for 80% of tweet impressions".
coffeeblack 4y ago
- EveYoung 4y ago
  In this context, why would public opinion matter?
  - vasco 4y ago
    Public opinion matters in all contexts, as rules and court decisions about rules are made by people, who invariantly are members of the public.
- blitzar 4y ago
  > The goal is to sway public opinion against Musk. At this point anyone with half a brain, no make that a single functioning brain cell, knows that Musk has fucked up and is trying to welch on the deal.
  - coffeeblack 4y ago
    - ngetchell 4y ago
      Do you really think Musk still wants to own Twitter for the agreed upon price?
      - coffeeblack 4y ago
        If the bot count really is under 5% as per Twitter’s claim, yes. But that’s unlikely, considering what could be observed over the past weeks.
        raydev 4y ago
        First you'll have to show where Twitter made that claim. Your phrasing here has been Musk's phrasing on this subject, which is interesting, as it's a different metric than what Twitter files publicly with the SEC.

by webo 4y ago

For me the main and obvious thing about DAU is that the majority of users are readers-only. They don’t tweet or like stuff. Firehose might help with detecting non-bot _tweeters_.

jameshart 4y ago

Right. There are no ads on the "compose tweet" page Pew analysis (based on surveying humans) says "97% of tweet volume comes from the top 25% most active tweeters"[1] So Musk is arguing about a statistic that twitter claims, which is that spambots make up less than 5% of users. Now, if those bots are part of the most active 25%, then they make up 1/5 of the 'most active' user population. So it seems plausible that they could well be responsible for 1/5 of tweets. Which matches exactly what Musk is claiming - that 20% of tweets are from spam bots. Which still makes no difference to Twitter's ability to sell Advertising targeting the 95% of users who are humans, 75/95 (=79%) of which are only responsible for 3% of tweets, but... presumably... 79% of ad revenue. Because they only serve you ads when you're looking at a feed. And that's what those users are using Twitter to do. [1] https://www.fastcompany.com/90697826/the-top-25-most-active-twitter-users-produce-97-of-all-tweets-while-most-of-us-just-lurk https://www.fastcompany.com/90697826/the-top-25-most-active-...

Helmut10001 4y ago
The famous pareto law - 80% of effects come from the top 20% of causes (users).
- Scoundreller 4y ago
  I think tweet volume is the wrong approach. If they looked at tweet impressions, it would be like 0.1% of users are responsible for 99% of tweet impressions.
  - Helmut10001 4y ago
    I think in this case, the pareto law would need to be framed with "The top 20% of most viewed tweets are responsible for 80% of tweet impressions".

coffeeblack 4y ago

EveYoung 4y ago
In this context, why would public opinion matter?
- vasco 4y ago
  Public opinion matters in all contexts, as rules and court decisions about rules are made by people, who invariantly are members of the public.
blitzar 4y ago
> The goal is to sway public opinion against Musk. At this point anyone with half a brain, no make that a single functioning brain cell, knows that Musk has fucked up and is trying to welch on the deal.
- coffeeblack 4y ago
  - ngetchell 4y ago
    Do you really think Musk still wants to own Twitter for the agreed upon price?
    - coffeeblack 4y ago
      If the bot count really is under 5% as per Twitter’s claim, yes. But that’s unlikely, considering what could be observed over the past weeks.
      - raydev 4y ago
        First you'll have to show where Twitter made that claim. Your phrasing here has been Musk's phrasing on this subject, which is interesting, as it's a different metric than what Twitter files publicly with the SEC.