3 ms·

But the question is why? Shouldn’t there be the same race to single digit margins in software as well?

simmerup 4y ago
I think you're right. It's my suspicion that eventually software engineers will be paid the same as copy writers and scientists. But not yet.
- insightcheck 4y ago
  I was initially going to disagree that copywriters get paid in the same range as scientists, but the numbers in the US are roughly similar from the Bureau of Labor Statistics and Indeed. -Indeed average copywriter annual pay [0] is $54,998 USD, versus a staff scientist's [1] at $69,091 USD -BLS median copywriter pay for 2021 [2] was $69,510 USD, versus a medical scientist's [3] at $95,310 USD I'm surprised that the median annual salary difference is only roughly $15,000 to $25,000 USD in favor of working as a scientist, considering the additional years doing a Master's/PhD, whereas copywriter jobs often don't need even a Bachelor's. In addition, I would've guessed that the median copywriter salary would have been far lower, as my perception was that there is a far greater supply of candidates for copywriter jobs than scientist jobs in the US (maybe this is offset by fewer scientist positions?). Alternatively, maybe the numbers are missing important contextual information in some way. [0] https://www.indeed.com/career/copywriter/salaries https://www.indeed.com/career/copywriter/salaries [1] https://www.indeed.com/career/staff-scientist/salaries https://www.indeed.com/career/staff-scientist/salaries [2] https://www.bls.gov/ooh/media-and-communication/writers-and-authors.htm https://www.bls.gov/ooh/media-and-communication/writers-and-... [3] https://www.bls.gov/ooh/life-physical-and-social-science/medical-scientists.htm https://www.bls.gov/ooh/life-physical-and-social-science/med...
  - simmerup 4y ago
    The scientists I know see it as almost a vocation. They want the prestige of having their names in journals and the feeling that they're helping advance human knowledge. I can't really blame them... I class it as a bit like development in the way they undervalue their skills so that they can get it. I guess the other side of it is that the people paying scientists really don't know if their grant money will pay off so they want to hedge their bets on costs (eg salaries)
  - t-3 4y ago
    > In addition, I would've guessed that the median copywriter salary would have been far lower, as my perception was that there is a far greater supply of candidates for copywriter jobs than scientist jobs in the US (maybe this is offset by fewer scientist positions?) Probably, but given that ~20% of Americans are functionally illiterate, while ~40% have a college degree, the supplies might not be too different. Reading and writing are also skills which are neither prestigious nor especially lucrative, while scientists of any kind are esteemed pretty highly.
insightcheck 4y ago
The distribution cost of selling software is far lower than selling a physical product. For example, to manufacture and sell cars, you currently have to worry about supply chain considerations and chip shortages. You don't need to worry about this to sell software to an interested purchaser.
atwood22 4y ago
The marginal cost of software is basically zero. The marginal cost of hardware is high.
ccorcos 4y ago
Marginal cost of an additional user is zero. And instant global distribution with the internet. And then there’s the freemium B2B SaaS business model (Slack, Notion) that is just insanely lucrative…
majormajor 4y ago
A different answer than everyone else so far: Leverage. Super-simple example: a single developer focused just on internal tools for a 10 person company. Let's say that this company has a full plate, everyone is busy 100% of the time. They can't really bring on new clients. They could hire a new person to do some of the work, but let's say that that new person will cost almost as much as a new client would bring in. Their margins are very tight, they wouldn't be able to expand fast at all, and losing a client would be very perilous. (This is how a lot of small companies work. People getting laid off because a client went away is something I've seen firsthand.) Let's say over the course of a year an engineer could build out enough basic, CRUD-app-style automation to free up 30% of everyone else's time. Now you can get multiple new clients for the cost of one new person! And if you can further automate and reduce the drudgery, even more! (This is also a partial explanation for why the outsourcing, salaries-will-go-to-zero expectations of the early-2000s haven't come to pass: it requires tight communication and a good understand of people's workflows, which is much harder to do across time zone and language barriers. Hardware, on the other hand, needs a firmer spec and less ability to iterate rapidly with rapid communication and feedback.) Of course the real world is more complicated and some of that automation comes from SaaS tools, etc (a less-direct way of paying those engineers, and something that could be more outsourceable in the long run) but a surprising amount of companies still need hands-on internal expertise to glue everything together. Less code tends to result in more code, because now that you and your competitors have gotten more efficient too, you have to get even more efficient to move the needle again. The thing about business processes is that they will always evolve as old companies die, new ones start, and consumers (both individuals and other companies) want different things.
est31 4y ago
There will be, but right now software is eating the world (and establishing entire new markets) faster than developer talent enters the market. So prices for developers go up. Eventually this will end, either by the demand curve slowing down, or developer supply meeting demand. Whether it happens within the next 5 or 10, or even 20 years, who knows.
- dougabug 4y ago
  This saturation of demand for software talent has been predicted for literally decades. In 1992, Ed Yourdon published “Decline and Fall of the American Programmer,” only to reverse himself four years later with the publication of “Rise and Resurrection of the American Programmer.” If anything, the rate at which capable software professionals can create value has dramatically increased over time. Software is not a zero sum game, cumulative advances make it possible to solve an expanding frontier of problems with increasingly effective and efficient solutions.
  - est31 4y ago
    Frontier is a good term because just as the US frontier age eventually ended, the software frontier expansion is also eventually going to stop.
    - dougabug 4y ago
      I don’t mean frontier as in “the American Western Frontier.” I meant frontier as in, “the frontiers of human knowledge.” It’s not clear to me that the latter is limited in the way that the former is. Creation of knowledge products both facilitates the further production of new knowledge products and expands the scope of new knowledge products which can be created. This process is like a chain reaction where the potential energy unleashed may ultimately exceed any practical scale modern economics has meaningfully considered. Software costs almost nothing to deploy, even on a global scale, versus its potential value delivered; and it can be created or improved with almost no additional consumption of resources vs (the programmer) simply existing.
jandrewrogers 4y ago
The rate of innovation in software is not constrained by capital or physics (and to a lesser extent regulation) to the extent it is in other engineering disciplines, which overweights the importance of individual skill and execution in outcomes. Software isn't an intrinsically high-margin business, but the cost of creating and operating in new high-margin business frontiers is uniquely low, allowing individuals to leave the rapidly commoditizing part of the software business in the rear-view mirror. The margins would be single-digits if the pace of change and innovation possible was like other engineering disciplines. The business of software engineering evolves at an incredible rate that would be unimaginable in the engineering discipline I went to school for.
olalonde 4y ago
In a free market, there is such race in every industry. It's just that demand for software is growing too damn fast and supply can't keep up.
pjc50 4y ago
Software's network effects make it much easier to construct a "monopoly" which gets to have very large margins. Where you see margins, look for moats.

I think you're right. It's my suspicion that eventually software engineers will be paid the same as copy writers and scientists. But not yet.

insightcheck 4y ago
I was initially going to disagree that copywriters get paid in the same range as scientists, but the numbers in the US are roughly similar from the Bureau of Labor Statistics and Indeed. -Indeed average copywriter annual pay [0] is $54,998 USD, versus a staff scientist's [1] at $69,091 USD -BLS median copywriter pay for 2021 [2] was $69,510 USD, versus a medical scientist's [3] at $95,310 USD I'm surprised that the median annual salary difference is only roughly $15,000 to $25,000 USD in favor of working as a scientist, considering the additional years doing a Master's/PhD, whereas copywriter jobs often don't need even a Bachelor's. In addition, I would've guessed that the median copywriter salary would have been far lower, as my perception was that there is a far greater supply of candidates for copywriter jobs than scientist jobs in the US (maybe this is offset by fewer scientist positions?). Alternatively, maybe the numbers are missing important contextual information in some way. [0] https://www.indeed.com/career/copywriter/salaries https://www.indeed.com/career/copywriter/salaries [1] https://www.indeed.com/career/staff-scientist/salaries https://www.indeed.com/career/staff-scientist/salaries [2] https://www.bls.gov/ooh/media-and-communication/writers-and-authors.htm https://www.bls.gov/ooh/media-and-communication/writers-and-... [3] https://www.bls.gov/ooh/life-physical-and-social-science/medical-scientists.htm https://www.bls.gov/ooh/life-physical-and-social-science/med...
- simmerup 4y ago
  The scientists I know see it as almost a vocation. They want the prestige of having their names in journals and the feeling that they're helping advance human knowledge. I can't really blame them... I class it as a bit like development in the way they undervalue their skills so that they can get it. I guess the other side of it is that the people paying scientists really don't know if their grant money will pay off so they want to hedge their bets on costs (eg salaries)
- t-3 4y ago
  > In addition, I would've guessed that the median copywriter salary would have been far lower, as my perception was that there is a far greater supply of candidates for copywriter jobs than scientist jobs in the US (maybe this is offset by fewer scientist positions?) Probably, but given that ~20% of Americans are functionally illiterate, while ~40% have a college degree, the supplies might not be too different. Reading and writing are also skills which are neither prestigious nor especially lucrative, while scientists of any kind are esteemed pretty highly.

insightcheck 4y ago

The distribution cost of selling software is far lower than selling a physical product. For example, to manufacture and sell cars, you currently have to worry about supply chain considerations and chip shortages. You don't need to worry about this to sell software to an interested purchaser.

atwood22 4y ago

The marginal cost of software is basically zero. The marginal cost of hardware is high.

ccorcos 4y ago

Marginal cost of an additional user is zero. And instant global distribution with the internet. And then there’s the freemium B2B SaaS business model (Slack, Notion) that is just insanely lucrative…

majormajor 4y ago

A different answer than everyone else so far: Leverage. Super-simple example: a single developer focused just on internal tools for a 10 person company. Let's say that this company has a full plate, everyone is busy 100% of the time. They can't really bring on new clients. They could hire a new person to do some of the work, but let's say that that new person will cost almost as much as a new client would bring in. Their margins are very tight, they wouldn't be able to expand fast at all, and losing a client would be very perilous. (This is how a lot of small companies work. People getting laid off because a client went away is something I've seen firsthand.) Let's say over the course of a year an engineer could build out enough basic, CRUD-app-style automation to free up 30% of everyone else's time. Now you can get multiple new clients for the cost of one new person! And if you can further automate and reduce the drudgery, even more! (This is also a partial explanation for why the outsourcing, salaries-will-go-to-zero expectations of the early-2000s haven't come to pass: it requires tight communication and a good understand of people's workflows, which is much harder to do across time zone and language barriers. Hardware, on the other hand, needs a firmer spec and less ability to iterate rapidly with rapid communication and feedback.) Of course the real world is more complicated and some of that automation comes from SaaS tools, etc (a less-direct way of paying those engineers, and something that could be more outsourceable in the long run) but a surprising amount of companies still need hands-on internal expertise to glue everything together. Less code tends to result in more code, because now that you and your competitors have gotten more efficient too, you have to get even more efficient to move the needle again. The thing about business processes is that they will always evolve as old companies die, new ones start, and consumers (both individuals and other companies) want different things.

est31 4y ago

There will be, but right now software is eating the world (and establishing entire new markets) faster than developer talent enters the market. So prices for developers go up. Eventually this will end, either by the demand curve slowing down, or developer supply meeting demand. Whether it happens within the next 5 or 10, or even 20 years, who knows.

dougabug 4y ago
This saturation of demand for software talent has been predicted for literally decades. In 1992, Ed Yourdon published “Decline and Fall of the American Programmer,” only to reverse himself four years later with the publication of “Rise and Resurrection of the American Programmer.” If anything, the rate at which capable software professionals can create value has dramatically increased over time. Software is not a zero sum game, cumulative advances make it possible to solve an expanding frontier of problems with increasingly effective and efficient solutions.
- est31 4y ago
  Frontier is a good term because just as the US frontier age eventually ended, the software frontier expansion is also eventually going to stop.
  - dougabug 4y ago
    I don’t mean frontier as in “the American Western Frontier.” I meant frontier as in, “the frontiers of human knowledge.” It’s not clear to me that the latter is limited in the way that the former is. Creation of knowledge products both facilitates the further production of new knowledge products and expands the scope of new knowledge products which can be created. This process is like a chain reaction where the potential energy unleashed may ultimately exceed any practical scale modern economics has meaningfully considered. Software costs almost nothing to deploy, even on a global scale, versus its potential value delivered; and it can be created or improved with almost no additional consumption of resources vs (the programmer) simply existing.

jandrewrogers 4y ago

The rate of innovation in software is not constrained by capital or physics (and to a lesser extent regulation) to the extent it is in other engineering disciplines, which overweights the importance of individual skill and execution in outcomes. Software isn't an intrinsically high-margin business, but the cost of creating and operating in new high-margin business frontiers is uniquely low, allowing individuals to leave the rapidly commoditizing part of the software business in the rear-view mirror. The margins would be single-digits if the pace of change and innovation possible was like other engineering disciplines. The business of software engineering evolves at an incredible rate that would be unimaginable in the engineering discipline I went to school for.

olalonde 4y ago

In a free market, there is such race in every industry. It's just that demand for software is growing too damn fast and supply can't keep up.

pjc50 4y ago

Software's network effects make it much easier to construct a "monopoly" which gets to have very large margins. Where you see margins, look for moats.