11 ms·

Reminds me of the research NPR did before closing their comments section. >In July, NPR.org recorded nearly 33 million unique users, and 491,000 comments. But

Reminds me of the research NPR did before closing their comments section. >In July, NPR.org recorded nearly 33 million unique users, and 491,000 comments. But those comments came from just 19,400 commenters, Montgomery said. That's 0.06 percent of users who are commenting, a number that has stayed steady through 2016. https://www.npr.org/sections/publiceditor/2016/08/17/489516952 https://www.npr.org/sections/publiceditor/2016/08/17/4895169...

dugmartin 6y ago
That pretty well matches the 90-9-1 rule: https://www.nngroup.com/articles/participation-inequality/ https://www.nngroup.com/articles/participation-inequality/
- ciarannolan 6y ago
  I wonder if dang or others could comment on whether this is true of HN too.
  - Cactus2018 6y ago
    Dang comment about the 1% rule from 7 months ago: https://news.ycombinator.com/item?id=22622983 https://news.ycombinator.com/item?id=22622983 > The number of accounts that have posted to HN this year, divided by the number of IP addresses that have accessed HN, is 0.008. How close that is to the '1% rule' ratio depends on which is the bigger factor: users with more than one IP or IPs with more than one user. We don't know. If the former is bigger, then 0.008 is a lower bound. > Here's another way. The number of accounts that have posted this year, divided by the number of accounts that have viewed HN while logged in, is 0.36. That doesn't tell us much, but we can estimate the ratio of logged-in users to total users this way: logged-in page views divided by total page views. That ratio is 0.23. We can multiply those two to estimate the ratio of posters to total: > So the two ways of estimating produce 0.8% and 8% respectively. Both ways are bogus in that they assume things we don't know and mix units that aren't the same, but they're the two I came up with and I don't remember how I did it before. It's interesting that they're almost exactly an order of magnitude apart. That makes it tempting to say the number is probably in between, but that's another cognitive bias talking.
    - ciarannolan 6y ago
      Very interesting, thanks.
  - avianlyric 6y ago
    I think dang has commented in a number of other threads basically saying that HN pretty much follows this rule as well. Hopefully someone will be able to find an example because my Googlefoo is failing me.
    - Cactus2018 6y ago
      I found a thread with the search term hackernews comments to visitor rate
- dugmartin 6y ago
  Re-reading that article makes me realize I've been using that rule a bit differently (wrongly?) over the last 10 or so years. I always assumed it meant 90% only read, 9% interact/comment and 1% create.
- glitcher 6y ago
  I have to say that the term "lurkers" to describe the 90% who don't participate rubs me a little the wrong way. I know this point is somewhat trivial, but it has a negative connotation that seems to promote the assumption that the only "correct" way to use social media is for all to participate. On the flip side, I wonder if everyone on social media were to comment on most items they had interest in, if it wouldn't completely drown out all of the extreme viewpoints. It may actually make social media extremely boring, as compared to the current standard. One can dream...
TheOtherHobbes 6y ago
Is this unusual for online forums? It's been a rule of thumb for decades that 1-5% of users are active posters and everyone else is what used to be called a "lurker" - there to read and be informed/outraged/entertained. I'd expect the same for politics. Of course politics affects everyone personally in all kinds of ways, but most of the population just isn't that interested.
- skrebbel 6y ago
  1-5% is very different from 0.06% though.
  - HPsquared 6y ago
    The 0.06% is the share of total NPR visitors, not those who read the comments sections. If we limit our sample to people who view the comments sections, it's probably closer to 1%.
    - skrebbel 6y ago
      Good point! You convinced me, it's not an impressive statistic (except maybe wrt a "nobody reads comments" argument)
base698 6y ago
Also knows as Pareto Principal and Matthew Effect https://en.m.wikipedia.org/wiki/Pareto_principle https://en.m.wikipedia.org/wiki/Pareto_principle https://en.m.wikipedia.org/wiki/Matthew_effect https://en.m.wikipedia.org/wiki/Matthew_effect

dugmartin 6y ago

That pretty well matches the 90-9-1 rule: https://www.nngroup.com/articles/participation-inequality/ https://www.nngroup.com/articles/participation-inequality/

ciarannolan 6y ago
I wonder if dang or others could comment on whether this is true of HN too.
- Cactus2018 6y ago
  Dang comment about the 1% rule from 7 months ago: https://news.ycombinator.com/item?id=22622983 https://news.ycombinator.com/item?id=22622983 > The number of accounts that have posted to HN this year, divided by the number of IP addresses that have accessed HN, is 0.008. How close that is to the '1% rule' ratio depends on which is the bigger factor: users with more than one IP or IPs with more than one user. We don't know. If the former is bigger, then 0.008 is a lower bound. > Here's another way. The number of accounts that have posted this year, divided by the number of accounts that have viewed HN while logged in, is 0.36. That doesn't tell us much, but we can estimate the ratio of logged-in users to total users this way: logged-in page views divided by total page views. That ratio is 0.23. We can multiply those two to estimate the ratio of posters to total: > So the two ways of estimating produce 0.8% and 8% respectively. Both ways are bogus in that they assume things we don't know and mix units that aren't the same, but they're the two I came up with and I don't remember how I did it before. It's interesting that they're almost exactly an order of magnitude apart. That makes it tempting to say the number is probably in between, but that's another cognitive bias talking.
  - ciarannolan 6y ago
    Very interesting, thanks.
- avianlyric 6y ago
  I think dang has commented in a number of other threads basically saying that HN pretty much follows this rule as well. Hopefully someone will be able to find an example because my Googlefoo is failing me.
  - Cactus2018 6y ago
    I found a thread with the search term hackernews comments to visitor rate
dugmartin 6y ago
Re-reading that article makes me realize I've been using that rule a bit differently (wrongly?) over the last 10 or so years. I always assumed it meant 90% only read, 9% interact/comment and 1% create.
glitcher 6y ago
I have to say that the term "lurkers" to describe the 90% who don't participate rubs me a little the wrong way. I know this point is somewhat trivial, but it has a negative connotation that seems to promote the assumption that the only "correct" way to use social media is for all to participate. On the flip side, I wonder if everyone on social media were to comment on most items they had interest in, if it wouldn't completely drown out all of the extreme viewpoints. It may actually make social media extremely boring, as compared to the current standard. One can dream...

TheOtherHobbes 6y ago

Is this unusual for online forums? It's been a rule of thumb for decades that 1-5% of users are active posters and everyone else is what used to be called a "lurker" - there to read and be informed/outraged/entertained. I'd expect the same for politics. Of course politics affects everyone personally in all kinds of ways, but most of the population just isn't that interested.

skrebbel 6y ago
1-5% is very different from 0.06% though.
- HPsquared 6y ago
  The 0.06% is the share of total NPR visitors, not those who read the comments sections. If we limit our sample to people who view the comments sections, it's probably closer to 1%.
  - skrebbel 6y ago
    Good point! You convinced me, it's not an impressive statistic (except maybe wrt a "nobody reads comments" argument)

base698 6y ago

Also knows as Pareto Principal and Matthew Effect https://en.m.wikipedia.org/wiki/Pareto_principle https://en.m.wikipedia.org/wiki/Pareto_principle https://en.m.wikipedia.org/wiki/Matthew_effect https://en.m.wikipedia.org/wiki/Matthew_effect