6 ms·

Price's Law: Why Only a Few People Generate Half of the Results

57 points by jxub 8y ago

smrk007 8y ago
So by my understanding, if only one person is doing the work, the square root of one is one, meaning they are only doing 50% of the work... Who does the other 50%?!
myth_drannon 8y ago
Hire more people for the law to work.
roma1n 8y ago
That's been debunked by examples where hiring (literal) team players that are not shining on their own enhances a team above the level of an 'all-star' team. So why do we keep repeating this nonsense?
- tomrod 8y ago
  Your question is a good one. The zeitgeist in the US has been for years that we at are all the same and "rockstars" just try harder. In reality, not everyone wants to be that, and not every team needs that.
- wpietri 8y ago
  It'd guess it's because a) it lets us all think we are in the "rockstar" category, which is pleasing, and b) it lets us avoid all sorts of difficult thinking about the system that actually produces the work, and c) justifies unequal rewards, which is entirely useful to the people receiving the rewards. It's the same reason the myth of the 10x engineer is so popular. I've worked with some people who certainly thought they were the 10x engineer. As far as I could tell, they were just showboats who did highly visible work, leaving little things like technical debt for others. Or the "brilliant" engineer who make things that require high cognitive load to understand, instead of doing the extra work to make them clear. [1] Or who shirk the work of supporting colleagues and building strong teams. So basically I think these are the people who optimize for the visible success metric, not the ones creating the most value. [1] Just this week a friend took over responsibility for an internal build system. The previous author had written it in JavaScript. Except that he was really excited about functional languages, so he wrote it in a highly functional style incomprehensible to anybody not used to it. Now this either needs to be mostly rewritten or anybody who wants to work on the build system needs to spend 6 months learning Haskell first. I'm sure this guy looked productive and got to sound brilliant in meetings, but a better productivity analysis would include the significant costs he imposed on others.
  - majewsky 8y ago
    > It's the same reason the myth of the 10x engineer is so popular. I understood why people believe the "10x engineer" stories after meeting 0.1x engineers. (And I don't mean juniors. I mean highly-paid consultants.)
maxxxxx 8y ago
This sounds more like a few people getting credit for the work of many people that work in the background.
petermcneeley 8y ago
These type of results are a consequence of game theory. Given that, I dont know what a popular psychology professor is doing speaking on this subject.
KKKKkkkk1 8y ago
Doing the work is one thing, assigning credit is another. The phenomenon that some people are 10x at doing the work exists but is rare. It is much more common that people are 10x at getting the credit.
- analog31 8y ago
  From a self interest standpoint, if "credit" is how people are rewarded, then getting credit is the work. Also, if the mechanism for assigning credit ensures that only sqrt(n) are rewarded, then it's self-fulfilling.
  - maxxxxx 8y ago
    You see that a lot in companies. Some people understand that getting their name on achievements is what really gets rewarded so they act accordingly.
    - xstartup 8y ago
      I realized this when at a company. The guy recommending the bonus was bit generous (really good at attributing bonus to the person who put in the hard work barring other bias) and the guy handing out the bonus was the one who battled hard to reduce every bonus as the company had hired him to limit the unnecessary bonus to employees. Every employee thought that he is the one fighting management to get them their bonus and got lots of love/respect and the generous guy got nothing.
- tonyedgecombe 8y ago
  Yes, although a sales role is quite easy to measure objectively.
  - maxxxxx 8y ago
    Maybe but sales is also a team effort. The best sales person will fail if he/she isn't backed by a big support system.
mhneu 8y ago
No no, don’t believe Jordan Peterson (main cite in this article). He’s a media troll who takes controversial views on purpose and is at the fringes of psychology. https://www.google.com/amp/s/www.washingtonpost.com/amphtml/opinions/the-profound-sadness-in-jordan-petersons-antidote-to-chaos/2018/05/09/8e1be3a4-53bd-11e8-9c91-7dab596e8252_story.html https://www.google.com/amp/s/www.washingtonpost.com/amphtml/... One of Peterson’s main arguments is about college campuses and politics. But 85% of HARD scientists - who have nothing to do with politics - are liberals (2015 AAAS poll). That argues that smart people like those at universities see through right wing propaganda - they’ve read Orwell and Arendt and Altemeyer and John Dean and “Manufacturing Consent” and understand how propaganda works. Peterson is toxic and definitely not an intellectual model.
- wwweston 8y ago
  This article isn't really about Peterson, it's about Price's Law. There's certainly some discussion to be had about how that law is conceived/applicable, but you seem to be taking the tack that Price's law is a suspect concept because Jordan Peterson is introducing people to it. If we're talking intellectual models, I doubt people are going to be impressed by that approach. Since we're here, though, I'm interested in evidence that Peterson is on the fringes of Psychology (the WaPo link does not make that case; it's not even clear it's a criticism of Peterson, though it arguably damns with faint praise). I think he and his audiences could use more good critics.
- zeth___ 8y ago
  Have you talked to any hard scientists? Reading liberal arts is right up there with reading bondage magazine monthly in terms of popularity and taboo. The definition of liberal is much closer to libertarian. I've never heard people make a defense of pedophilia in public outside of a lab at 8pm when all the normal people have left.
- noobiemcfoob 8y ago
  You're holding up a liberal rag as evidence that Peterson (a vocal conservative) is a troll. He's not a troll so much as someone on the wrong end of identity politics. Sam Harris and others wouldn't associate with him if your claims were true. Working with that 85% of "HARD scientists", yeah, they're mostly pretty liberal, but they're just as likely to spout off horrible conspiracy theories and assume all conservatives are backwater inbreds as anyone else. That bias is supremely real, and it's not because scientists have a magical, rational filter against propaganda.
drostie 8y ago
I liked the intriguing idea of how this article started but it did end on a cop-out, “now that I have suggested this phenomenon rather than understanding what causes it so that we can harness it in practice, I invite you (what must necessarily be the majority of my readership if the pattern holds true) to quit your jobs!” It would be very interesting to work out a probability distribution f(r), 0 < r < ∞ support, such that when sampled N times independently half of the total sum is contained in the top √N terms. Like I would suspect it can be done with long-tailed distributions, though they might need to have an undefined cumulant or so... just that trying to set up that “sum of the top √N” integral sounds very tricky in a way that would foster some clever insight. But of course the more likely explanation would probably have something to do with network effects. If you imagine that each person is a morphism (or a bounded number of morphisms?) in a category for example, perhaps whose objects are indirectly profit centers, then having N profit centers correlates to having N objects that your company relates, which correlates to N^2 employees connecting them together, so if the top 50% of profit resides in some fixed number of these objects, then only the ~N employees out of the ~N^2 who happen to deal with these “cash cows” would make up half of the profit. To detect this you would want to show that rather than linearly scaling with employees, the returns to scale might show a characteristic pattern of profit versus company size.
- mhneu 8y ago
  I suspect a big cause of nonlinear productivity effects like this is the experience and skills required to do the job. In the example cited, sales, the best salespeople have long-built networks of personal connections. That allows them to bring in deals that new people can’t access. It’s very different than retail work where it’s hard to greatly outperform your peers and little training is needed. So agree models like this can be created but my feeling is the key aspects of the model relate to experience (related to your cash cow) rather than scale. At least for creative jobs.
- mc32 8y ago
  Perhaps it's a bit like traffic flow. Once you go beyond capacity, things slow down (productivity) but if you can offload some of the traffic load (to a low volume side street) the main highway becomes productive again. That's to say, while not producing equally as high producers, the slackers do in a way help high producers produce in high volume by offloading some of the more interruptive tasks. Kind of like an executive needing an executive's assistant.
fanzhang 8y ago
Probably depends on the domain. I really don't see it for carrying bricks, but I can totally see it for abstract math research.
wwweston 8y ago
It's probably worth noting that "credit" rather than work is explicitly the relevant concept in play in the formation of Price's Law -- Price seems to have conceived the relationship in the context of looking at academic citations. "Work done" is certainly one component of having a much-cited paper; you are simply not going to have a relatively large number of cited papers if you have not written a relatively large number of papers. But there's the question of what other factors lead papers to awareness, recognition of value, and citation, and those questions likely don't have a linear relationship with "value created." Given the description of some of the dynamics involved as a "preferential attachment process" involving "cumulative advantage" (https://en.wikipedia.org/wiki/Preferential_attachment https://en.wikipedia.org/wiki/Preferential_attachment ), Price seems to have recognized this and probably had a different conception than the author of the piece we're discussing. (Peterson, who the author indicates brought Price's law to his attention, seems to recognize this as well, and sometimes suggests that it's best to assume that people who aren't at the top of some Pareto distribution "P" in contributing value are probably contributing value in some other way that isn't visible to people focusing on P. )
madhadron 8y ago
There's a basic problem with this: Price's law doesn't fit the data. (DOI 10.1016/0306-4573(88)90049-0) It certainly shouldn't be randomly reapplied to things besides scientific publication. Even within scientific publication, the uneven distribution of publications is a result of far more complicated mechanisms than some simple measure of results. Remember, by bibliometric methods, Newton, Feynman, and Perelman are some of the lowest performers ever to work in science.
ouid 8y ago
Price's Law doesn't say what the author of this article seems to think. The marginal utility of hiring additional workers in an organization is smaller for each additional employee. This is both obvious, and completely accounts for the effect on its own.
- droidist2 8y ago
  Oh, kind of like the Mythical Man Month, except there the marginal utility at some point becomes negative for each additional employee assigned to a project.
rdlecler1 8y ago
Taking the idea that the output is the square root of the number of participants, suggests that three companies of 10 people (3 * sqrt(10) ~= 10) do the same work as one company of 100 people (1 * sqrt(100) = 10). Not unreasonable but it does suggest optimal organizational sizes and the emergence of modular design.
- uxhack 8y ago
  A company of 10 will produce an output of 6 if 3 people produce 50% of the work. A company of 100 will produce the output of 20 if 10 people do 50% of the production. So a company of 100 will produce 3 times more than a company of 10.

57 points by jxub 8y ago

So by my understanding, if only one person is doing the work, the square root of one is one, meaning they are only doing 50% of the work... Who does the other 50%?!

myth_drannon 8y ago

Hire more people for the law to work.

roma1n 8y ago

That's been debunked by examples where hiring (literal) team players that are not shining on their own enhances a team above the level of an 'all-star' team. So why do we keep repeating this nonsense?

tomrod 8y ago
Your question is a good one. The zeitgeist in the US has been for years that we at are all the same and "rockstars" just try harder. In reality, not everyone wants to be that, and not every team needs that.
wpietri 8y ago
It'd guess it's because a) it lets us all think we are in the "rockstar" category, which is pleasing, and b) it lets us avoid all sorts of difficult thinking about the system that actually produces the work, and c) justifies unequal rewards, which is entirely useful to the people receiving the rewards. It's the same reason the myth of the 10x engineer is so popular. I've worked with some people who certainly thought they were the 10x engineer. As far as I could tell, they were just showboats who did highly visible work, leaving little things like technical debt for others. Or the "brilliant" engineer who make things that require high cognitive load to understand, instead of doing the extra work to make them clear. [1] Or who shirk the work of supporting colleagues and building strong teams. So basically I think these are the people who optimize for the visible success metric, not the ones creating the most value. [1] Just this week a friend took over responsibility for an internal build system. The previous author had written it in JavaScript. Except that he was really excited about functional languages, so he wrote it in a highly functional style incomprehensible to anybody not used to it. Now this either needs to be mostly rewritten or anybody who wants to work on the build system needs to spend 6 months learning Haskell first. I'm sure this guy looked productive and got to sound brilliant in meetings, but a better productivity analysis would include the significant costs he imposed on others.
- majewsky 8y ago
  > It's the same reason the myth of the 10x engineer is so popular. I understood why people believe the "10x engineer" stories after meeting 0.1x engineers. (And I don't mean juniors. I mean highly-paid consultants.)

maxxxxx 8y ago

This sounds more like a few people getting credit for the work of many people that work in the background.

petermcneeley 8y ago

These type of results are a consequence of game theory. Given that, I dont know what a popular psychology professor is doing speaking on this subject.

KKKKkkkk1 8y ago

Doing the work is one thing, assigning credit is another. The phenomenon that some people are 10x at doing the work exists but is rare. It is much more common that people are 10x at getting the credit.

analog31 8y ago
From a self interest standpoint, if "credit" is how people are rewarded, then getting credit is the work. Also, if the mechanism for assigning credit ensures that only sqrt(n) are rewarded, then it's self-fulfilling.
- maxxxxx 8y ago
  You see that a lot in companies. Some people understand that getting their name on achievements is what really gets rewarded so they act accordingly.
  - xstartup 8y ago
    I realized this when at a company. The guy recommending the bonus was bit generous (really good at attributing bonus to the person who put in the hard work barring other bias) and the guy handing out the bonus was the one who battled hard to reduce every bonus as the company had hired him to limit the unnecessary bonus to employees. Every employee thought that he is the one fighting management to get them their bonus and got lots of love/respect and the generous guy got nothing.
tonyedgecombe 8y ago
Yes, although a sales role is quite easy to measure objectively.
- maxxxxx 8y ago
  Maybe but sales is also a team effort. The best sales person will fail if he/she isn't backed by a big support system.

mhneu 8y ago

No no, don’t believe Jordan Peterson (main cite in this article). He’s a media troll who takes controversial views on purpose and is at the fringes of psychology. https://www.google.com/amp/s/www.washingtonpost.com/amphtml/opinions/the-profound-sadness-in-jordan-petersons-antidote-to-chaos/2018/05/09/8e1be3a4-53bd-11e8-9c91-7dab596e8252_story.html https://www.google.com/amp/s/www.washingtonpost.com/amphtml/... One of Peterson’s main arguments is about college campuses and politics. But 85% of HARD scientists - who have nothing to do with politics - are liberals (2015 AAAS poll). That argues that smart people like those at universities see through right wing propaganda - they’ve read Orwell and Arendt and Altemeyer and John Dean and “Manufacturing Consent” and understand how propaganda works. Peterson is toxic and definitely not an intellectual model.

wwweston 8y ago
This article isn't really about Peterson, it's about Price's Law. There's certainly some discussion to be had about how that law is conceived/applicable, but you seem to be taking the tack that Price's law is a suspect concept because Jordan Peterson is introducing people to it. If we're talking intellectual models, I doubt people are going to be impressed by that approach. Since we're here, though, I'm interested in evidence that Peterson is on the fringes of Psychology (the WaPo link does not make that case; it's not even clear it's a criticism of Peterson, though it arguably damns with faint praise). I think he and his audiences could use more good critics.
zeth___ 8y ago
Have you talked to any hard scientists? Reading liberal arts is right up there with reading bondage magazine monthly in terms of popularity and taboo. The definition of liberal is much closer to libertarian. I've never heard people make a defense of pedophilia in public outside of a lab at 8pm when all the normal people have left.
noobiemcfoob 8y ago
You're holding up a liberal rag as evidence that Peterson (a vocal conservative) is a troll. He's not a troll so much as someone on the wrong end of identity politics. Sam Harris and others wouldn't associate with him if your claims were true. Working with that 85% of "HARD scientists", yeah, they're mostly pretty liberal, but they're just as likely to spout off horrible conspiracy theories and assume all conservatives are backwater inbreds as anyone else. That bias is supremely real, and it's not because scientists have a magical, rational filter against propaganda.

drostie 8y ago

I liked the intriguing idea of how this article started but it did end on a cop-out, “now that I have suggested this phenomenon rather than understanding what causes it so that we can harness it in practice, I invite you (what must necessarily be the majority of my readership if the pattern holds true) to quit your jobs!” It would be very interesting to work out a probability distribution f(r), 0 < r < ∞ support, such that when sampled N times independently half of the total sum is contained in the top √N terms. Like I would suspect it can be done with long-tailed distributions, though they might need to have an undefined cumulant or so... just that trying to set up that “sum of the top √N” integral sounds very tricky in a way that would foster some clever insight. But of course the more likely explanation would probably have something to do with network effects. If you imagine that each person is a morphism (or a bounded number of morphisms?) in a category for example, perhaps whose objects are indirectly profit centers, then having N profit centers correlates to having N objects that your company relates, which correlates to N^2 employees connecting them together, so if the top 50% of profit resides in some fixed number of these objects, then only the ~N employees out of the ~N^2 who happen to deal with these “cash cows” would make up half of the profit. To detect this you would want to show that rather than linearly scaling with employees, the returns to scale might show a characteristic pattern of profit versus company size.

mhneu 8y ago
I suspect a big cause of nonlinear productivity effects like this is the experience and skills required to do the job. In the example cited, sales, the best salespeople have long-built networks of personal connections. That allows them to bring in deals that new people can’t access. It’s very different than retail work where it’s hard to greatly outperform your peers and little training is needed. So agree models like this can be created but my feeling is the key aspects of the model relate to experience (related to your cash cow) rather than scale. At least for creative jobs.
mc32 8y ago
Perhaps it's a bit like traffic flow. Once you go beyond capacity, things slow down (productivity) but if you can offload some of the traffic load (to a low volume side street) the main highway becomes productive again. That's to say, while not producing equally as high producers, the slackers do in a way help high producers produce in high volume by offloading some of the more interruptive tasks. Kind of like an executive needing an executive's assistant.

fanzhang 8y ago

Probably depends on the domain. I really don't see it for carrying bricks, but I can totally see it for abstract math research.

wwweston 8y ago

It's probably worth noting that "credit" rather than work is explicitly the relevant concept in play in the formation of Price's Law -- Price seems to have conceived the relationship in the context of looking at academic citations. "Work done" is certainly one component of having a much-cited paper; you are simply not going to have a relatively large number of cited papers if you have not written a relatively large number of papers. But there's the question of what other factors lead papers to awareness, recognition of value, and citation, and those questions likely don't have a linear relationship with "value created." Given the description of some of the dynamics involved as a "preferential attachment process" involving "cumulative advantage" (https://en.wikipedia.org/wiki/Preferential_attachment https://en.wikipedia.org/wiki/Preferential_attachment ), Price seems to have recognized this and probably had a different conception than the author of the piece we're discussing. (Peterson, who the author indicates brought Price's law to his attention, seems to recognize this as well, and sometimes suggests that it's best to assume that people who aren't at the top of some Pareto distribution "P" in contributing value are probably contributing value in some other way that isn't visible to people focusing on P. )

madhadron 8y ago

There's a basic problem with this: Price's law doesn't fit the data. (DOI 10.1016/0306-4573(88)90049-0) It certainly shouldn't be randomly reapplied to things besides scientific publication. Even within scientific publication, the uneven distribution of publications is a result of far more complicated mechanisms than some simple measure of results. Remember, by bibliometric methods, Newton, Feynman, and Perelman are some of the lowest performers ever to work in science.

ouid 8y ago

Price's Law doesn't say what the author of this article seems to think. The marginal utility of hiring additional workers in an organization is smaller for each additional employee. This is both obvious, and completely accounts for the effect on its own.

droidist2 8y ago
Oh, kind of like the Mythical Man Month, except there the marginal utility at some point becomes negative for each additional employee assigned to a project.

rdlecler1 8y ago

Taking the idea that the output is the square root of the number of participants, suggests that three companies of 10 people (3 * sqrt(10) ~= 10) do the same work as one company of 100 people (1 * sqrt(100) = 10). Not unreasonable but it does suggest optimal organizational sizes and the emergence of modular design.

uxhack 8y ago
A company of 10 will produce an output of 6 if 3 people produce 50% of the work. A company of 100 will produce the output of 20 if 10 people do 50% of the production. So a company of 100 will produce 3 times more than a company of 10.