> The open source method may be effective if enough people play along, but it does not make money in itself. Moreover, by definition it implies that open source projects have many more mistakes, bad code and failed efforts on their way to succeeding, compared with conventional projects.
Which definition of open source (method?) do they go by?
> A regular company couldn’t have experimented with creating 10 versions of an online photo album, then picked the best one.
Yes, but often 10 companies try to make an online photo album and the market picks one (or a few). The comparison between open source projects and the dynamics inside a company doesn't seem like the right one.
Its extremely difficult to make money with open source. With the notable exception of Red Hat there aren't a lot of open source companies that are known to be consistently profitable.
Most of the bigger companies raised money when VC's believed in the open source business model and are still operating on those funds. That's an avenue that is no longer open to today's open source entrepreneurs.
One possible exception might be the WordPress model of offering hosting.
The way you make money with open-source is by not trying to sell the open-source product. Apple releases lots of stuff as open-source, but it's not a problem because they don't want to sell you Clang. Similarly, Red Hat aims to sell big corporate contracts, not licenses to their open-source software.
> The way you make money with open-source is by not trying to sell the open-source product. Apple releases lots of stuff as open-source, but it's not a problem because they don't want to sell you Clang.
In other words, the way you make money with open source is by not making money with open source. The downside of this is that you don't make money with open source.
Unless you count using open source (as it contributes to the development and monetization of the non-open-source software you make money with). But then you could say everyone is using open source to make money, which I think misses the point the article is trying to make.
>But then you could say everyone is using open source to make money, which I think misses the point the article is trying to make.
Which is that you can't make money by selling open source software? That's pretty obvious. I think the point of the article is to discuss possible ways (or whether it is ultimately impossible) to make money for a company that produces open source software, which is not something everyone can claim to be doing.
Right, you make money off open source software by monetizing its complements. Intel pays for Linux kernel developers because having a quality Linux kernel means that companies are more likely to buy some commodity x86 box than they are to buy a commercial Unix running on SPARC or POWER. Red Hat spends money on all kinds of Linux development because they can monetize their support effort. This works well for some kinds of software that have complements that lend themselves well to monetization by the same people who are funding development. Other kinds of projects aren't as lucky.
It should be clear by now that this is a fallacy, many open source projects have attempted to sell support but there's just not that much demand. A good product does not need to offer much support, and consulting doesn't scale. iText is a good example of an open source library that has tried pretty much every open strategy and reverted to closed-source.
By definition you're not going to make money through open source selling licenses, but you're also not going to make any real money selling something else, unless you keep some aspect of it closed, like Google do with their search indexes.
The open source software which most companies publish has nothing to do with how they make money, it's usually some implementation detail of their infrastructure that represents an incremental improvement on some similar open software which came before it.
I'm not sure how profitable they are, but puppetlabs is a very fast growing company that offers an open source product. The 'pay for support' model works well with certain open source products.
I'd say it's an example of the opposite, the partly closed-source Puppet Enterprise is where the primary focus is now, support and training are still available but more of a sideshow. Also Puppet Labs got their initial funding almost 10 years ago now, back when investors believed in open source, as the parent comment observes. They've received around $85m to date, half of that since their pivot to Puppet Enterprise.
If you take away licensing restrictions, then those business models that depend on licensing revenue are naturally going to suffer. But that's almost a mathematical truth and fairly obvious.
But there are lots of other businesses out there that make heavy use of open source software to make a lot of money selling something other than licenses. And those businesses often recognize an interest in contributing to the software upon which they depend.
... 90% of its revenue, yes, we all know this. The uppost was basically saying that its revenue is derived at least partially from the use of open source (not in the selling of services around the use of open source by its customers, as say Red Hat does). IT's a digression from the point of the original article, but in short: many very successful commercial enterprises depend heavily on open source to get their work done. It wouldn't make sense any other way- the licensing costs to assemble the equivalent of GNU/Linux and the code that runs on top of it, using only proprietary software, would be ruinous, and in many cases, there just aren't superior commercial equivalents anyway.
I don't think open source is really about making money for most companies, it's more about improving quality and keeping costs down. At a former employer we did follow the give-it-away-and-charge-for-a-service model, but we also worked on several of the open source libraries and frameworks we built our products on.
We shared all of this work with the upstream projects, as we had no interest whatsoever in maintaining private forks and because better quality infrastructure attracts more users, which again leads to further improvements or at least ensures the project stays maintained.
> it's more about improving quality and keeping costs down.
Yes. This point seems to be lost on the author of the article. Even the founders of the Open Source movement thought that.
I can't find it now, but I remember once reading an essay by Bruce Perens where he pointed out that the vast majority of software that's written lives on the cost side, not the revenue side, of a company's balance sheet, because the vast majority of software that's written is custom in-house software. If you can share those costs with others who have similar needs, everyone benefits.
As a kid who had only ever thought of software as "something you have to pay for when you want a copy for your computer", this was eye-opening for me. The idea that the software I was familiar with was the exception, rather than the rule, made it plausible that there were other ways to get high-quality software through besides paying a license fee to Microsoft. As a result, I found that idea very attractive.
The economic system in the US in based on very, very strong property rights. This greatly skews the advantage from the ones doing the work to the ones who own the result of this work.
If US property rights were not so strong there would not be such enormous advantages given to those who produce proprietary software. This isn't as much an issue with open source business models as much as it is with US economic policy.
Dual licensing, like Trolltech did originally with QT, could also be a viable commercial strategy for certain open source projects, but that seems to be going out of fashion lately.
What about the software maintainers that are often salaried employees?
And what about the main users of the software that are saving money by using free software?
But of course your whole competitive edge is that you are giving away those rights to keep the software proprietary and to dictate who can use it, for what and for how much. You cannot simply expect to put a price tag on it and sell it, "as is".
It's a shame our options have been reduced to a false dichotomy of pay-for software where you have no access to the source, and pay-nothing-for software where all you have is the source. There was supposed to be another option for pay-for software where you have access to the source and may make changes to that source to suite your needs: "free" was supposed to refer to "freedom" and not "cost." But no one seems to have gone that route. Is it just not viable?
There's nothing stopping most commercial products from releasing their source. It wouldn't impact sales. It wouldn't be a competitive disadvantage. The Code Combat team open sourced everything, and as far as I know they didn't suffer for it.
Management is just skittish. Nobody wants to open source because nobody else wants to open source.
A number of companies are successfully charging for use of their open-source software as a hosted service. For example, getsentry.com and travis-ci.org.
The problem is that for-pay software that gives you some freedoms manages to piss off everybody. It's not "open source". It's not "free software". It can be characterized as trying to astroturf your way into free goodwill. And you have minimal real protection against someone cloning your software.
I am currently trying to build a product on this model, because I think it is pragmatic. Not sure how it will work out.
FTFA: "Moreover, by definition it implies that open source projects have many more mistakes, bad code and failed efforts on their way to succeeding, compared with conventional projects."
Wait, what? How, exactly, does OSS "by definition" have more mistakes?
That statement alone is enough to dismiss the entire article.
Perhaps if you include every half-hearted attempt to throw code in the open as a "failed effort", and ignore such half-hearted attempts for proprietary software (because they are hidden), then perhaps you could reach such a conclusion.
But I doubt the author even bothered to look for any data; instead he just printed it because it made sense to him at the time.
I guess that's also how he reached the conclusion that "it’s time to admit that this idea didn’t work out" despite the wild growth, unimaginable reach, ubiquitous use, and general acceptance of open source.
Linus Torvalds knew that "the GPL would turn as many people away as it would attract" and exactly of that (it attracted a lot) he considers his decision for the GPL "to be the best decision he ever made" (though he disagrees with the too strong ideological believes of RMS "I want to (be able to) reprogram my printer's software"... sorry what?!). Linus used the big wave (pro & cons discussion on the GPL) for getting his Linux off the ground (which he succeeded with). And he is so incredibly smart and doing things right (showed it again with git... slamming CVS&SVN basically to death in a talk he gave).
Google, Facebook and others are open sourcing a good amount admitting, that this attracts skilled people (probably to be hired) and benefits their code (in the open)... "companies that use your code are never competitors". Though they remain the major force behind these projects, e.g. being accused to be show offs (hiphop vm) "look how great we are, but you'll never get it to work".
The default license on git hub is MIT. Sorry GPL, but forcing everyone to open source on modifications they make. NO way for me: Take it (http://www.use-the-tree.com) and make it proprietary (and better and make a business), you are so welcome. When you succeed with that, then congrats to you.
Enterprises want to pay and they go for FUD "no accountant was ever fired for buying IBM/Microsoft/HP". It's hard to break into that and they are even right with it, because when buying from a large vendor they are also buying into the huge support/manpower.
How does Ubuntu want to generate income? Selling shirts and Coffee-cups with "Ubuntu" printed on (okay advertising for Amazon in Unity) and diverging from "Linux for people/the desktop" to Ubuntu-Server/Cloud? But the fact that they sell T-Shirts and coffee cups (and others sell likewise silly things, like boxes with CDs and printed manuals), wow... that's a confession of failure (for this type of doing open source).
So "open source (almost) everything" (2011) [1], what a late(!) and great contribution to this discussion (free from ideological believes and focused on business).
Does RMS still live for free on the Harvard/MIT campus? And sorry I was in the group that was turned away back in the times.
The sentence immediately before the one you quoted is: "Remember how the open source software movement was supposed to be like Woodstock, with everybody sharing and everything free?"
I think the author wants to express that open source has changed a lot from the early days. As far as I remember it, open source originally was driven mainly by students, academics and hobbyists, and by service companies like Red Hat or maybe IBM. Nowadays, many of the larger, successful and enduring open source projects are mainly developed by professional developers working for companies that use the open source software for their commercial products or infrastructure.
I understand the article as making the argument that in the past companies tried to make a business out of the open source project itself while nowadays the open source project is more often a byproduct or a basic infrastructure component for other commercial projects.
I don't think the examples and quotes used by the author are great, but I do think that there is some truth to this argument.
Remember how the open source software movement was supposed to be like Woodstock, with everybody sharing and everything free?
No?
I'll go back and finish the article, but it's starting to look like the something you'd see in a tweet from @nytonit "Building a business around a product you give away is hard and The Times is ON IT."
I don't find open source as a product has the benefits we assume it must.
Theoretically, yes, if I have the source every change and customization I want to make is just a recompile away.
Practically speaking that almost never happens because now I'm likely forking a huge code base and almost always abandoning the support contract I paid money for to do so.
Unfortunately, because theoretically speaking anyone can change anything, authors seem to neglect the whole documentation and API portion of open-source software, and that's what I typically want in practice.
Obviously there are exceptions, but most people pay money for support, warranties, integration with existing systems, documentation and available training. All of those things have very little to do with whether the source code is open or not.
I work on open source software with two other developers and our business makes enough to pay for three full-time salaries (they're low-end software developer salaries, but we all live comfortably in the USA and Mexico).
We sell two WordPress plugins, both of which have free counterparts (think Lite/Pro). The source code is GPL for both versions.
Why do people buy the Pro version of our software? That's we've asked many times over the past few years. I feel they buy the Pro version for both the features and, more importantly, for the fact that it will be _updated_ and _maintained_, and they will have _access to support_. Site owners that want to use one of our plugins recognize that by paying for the plugin they get access to support and features not available in the Lite version, whereas the free version has limited features and only community support.
Could someone copy the source code for the Pro version and sell it? Sure. (And people have.) But can they copy our Pro support and sell that? Not so much. Can they copy our reputation? Nope.
I've learned over the past few years just how valuable "access to support" is for many people. Even if they never contact us, just knowing that we're there _in case something goes wrong_ is comforting enough for them to pay for that comfort.
If more open source projects started offering support, I bet you'd have a lot more open source developers making money.
For example, if I decided to set up `mutt` for the first time to switch over from web-based email, would I buy a "mutt support service" from someone that allowed me to submit `mutt` questions through a ticketing system while I was setting up and getting familiar with `mutt`? Yes! I would, especially if that person had a reputation in the community as someone who knows `mutt` inside out.
It's also worth noting how offering support helps drive documentation. If the documentation is great then a developer can easily point people to the relevant information and save him/herself time (so that support doesn't take up all available time). I basically used this type of "ask me a question, I'll create documentation for it and then reply with a link" method with my Independent Publisher WordPress theme project [1].
Building a viable business on top of an open source project is totally possible. But, like building lots of other types of businesses, it can be difficult.
These challenges cut close to some things we are aiming to solve at Assembly [http://assembly.com]. One such question we hope to answer is this: "how can people build a real, profitable company in a collaborative, open manner like they build in an open source environment?"
An example is Coderwall, which makes more than $25,000/month [1] and is built and maintained by the community. Each month, revenue goes to the people who are building the product. All the code is publicly available [2], and licensed to be used non-commercially. A core team of contributors guide vision and quality control, but anyone can participate.
It's not exactly open source, but products on Assembly are built in the open and anyone can dive in and help out. Some products have lofty ambitions to make big revenue, and others are are more aimed at being a public offering like a traditional open source project.
41 comments
[ 3.1 ms ] story [ 102 ms ] threadWhich definition of open source (method?) do they go by?
Yes, but often 10 companies try to make an online photo album and the market picks one (or a few). The comparison between open source projects and the dynamics inside a company doesn't seem like the right one.
Most of the bigger companies raised money when VC's believed in the open source business model and are still operating on those funds. That's an avenue that is no longer open to today's open source entrepreneurs.
One possible exception might be the WordPress model of offering hosting.
In other words, the way you make money with open source is by not making money with open source. The downside of this is that you don't make money with open source.
Unless you count using open source (as it contributes to the development and monetization of the non-open-source software you make money with). But then you could say everyone is using open source to make money, which I think misses the point the article is trying to make.
Which is that you can't make money by selling open source software? That's pretty obvious. I think the point of the article is to discuss possible ways (or whether it is ultimately impossible) to make money for a company that produces open source software, which is not something everyone can claim to be doing.
By definition you're not going to make money through open source selling licenses, but you're also not going to make any real money selling something else, unless you keep some aspect of it closed, like Google do with their search indexes.
The open source software which most companies publish has nothing to do with how they make money, it's usually some implementation detail of their infrastructure that represents an incremental improvement on some similar open software which came before it.
If you take away licensing restrictions, then those business models that depend on licensing revenue are naturally going to suffer. But that's almost a mathematical truth and fairly obvious.
But there are lots of other businesses out there that make heavy use of open source software to make a lot of money selling something other than licenses. And those businesses often recognize an interest in contributing to the software upon which they depend.
Google makes a lot of money from advertising.
FTFY
We shared all of this work with the upstream projects, as we had no interest whatsoever in maintaining private forks and because better quality infrastructure attracts more users, which again leads to further improvements or at least ensures the project stays maintained.
Yes. This point seems to be lost on the author of the article. Even the founders of the Open Source movement thought that.
I can't find it now, but I remember once reading an essay by Bruce Perens where he pointed out that the vast majority of software that's written lives on the cost side, not the revenue side, of a company's balance sheet, because the vast majority of software that's written is custom in-house software. If you can share those costs with others who have similar needs, everyone benefits.
As a kid who had only ever thought of software as "something you have to pay for when you want a copy for your computer", this was eye-opening for me. The idea that the software I was familiar with was the exception, rather than the rule, made it plausible that there were other ways to get high-quality software through besides paying a license fee to Microsoft. As a result, I found that idea very attractive.
If US property rights were not so strong there would not be such enormous advantages given to those who produce proprietary software. This isn't as much an issue with open source business models as much as it is with US economic policy.
Why? Can you elaborate on that?
Digia still does dual licensing of Qt and doing OK.
In any case, I wonder whether the original success of Qt would still be possible in today's business and open source community environment.
And what about the main users of the software that are saving money by using free software?
But of course your whole competitive edge is that you are giving away those rights to keep the software proprietary and to dictate who can use it, for what and for how much. You cannot simply expect to put a price tag on it and sell it, "as is".
Management is just skittish. Nobody wants to open source because nobody else wants to open source.
And the vendors that later move to other platforms carry the idea, for example remobjects.com & devexpress.com.
This make sense. Having the source not mean I want to change it.
P.D: But the source is not licensed as open source. So this is the other model
I am currently trying to build a product on this model, because I think it is pragmatic. Not sure how it will work out.
Wait, what? How, exactly, does OSS "by definition" have more mistakes?
Perhaps if you include every half-hearted attempt to throw code in the open as a "failed effort", and ignore such half-hearted attempts for proprietary software (because they are hidden), then perhaps you could reach such a conclusion.
But I doubt the author even bothered to look for any data; instead he just printed it because it made sense to him at the time.
I guess that's also how he reached the conclusion that "it’s time to admit that this idea didn’t work out" despite the wild growth, unimaginable reach, ubiquitous use, and general acceptance of open source.
Google, Facebook and others are open sourcing a good amount admitting, that this attracts skilled people (probably to be hired) and benefits their code (in the open)... "companies that use your code are never competitors". Though they remain the major force behind these projects, e.g. being accused to be show offs (hiphop vm) "look how great we are, but you'll never get it to work".
The default license on git hub is MIT. Sorry GPL, but forcing everyone to open source on modifications they make. NO way for me: Take it (http://www.use-the-tree.com) and make it proprietary (and better and make a business), you are so welcome. When you succeed with that, then congrats to you.
Enterprises want to pay and they go for FUD "no accountant was ever fired for buying IBM/Microsoft/HP". It's hard to break into that and they are even right with it, because when buying from a large vendor they are also buying into the huge support/manpower.
How does Ubuntu want to generate income? Selling shirts and Coffee-cups with "Ubuntu" printed on (okay advertising for Amazon in Unity) and diverging from "Linux for people/the desktop" to Ubuntu-Server/Cloud? But the fact that they sell T-Shirts and coffee cups (and others sell likewise silly things, like boxes with CDs and printed manuals), wow... that's a confession of failure (for this type of doing open source).
So "open source (almost) everything" (2011) [1], what a late(!) and great contribution to this discussion (free from ideological believes and focused on business).
Does RMS still live for free on the Harvard/MIT campus? And sorry I was in the group that was turned away back in the times.
[1] http://tom.preston-werner.com/2011/11/22/open-source-everyth...
If this article was published in 1998, and we revisited it in 2008, we would have concluded that the author was proven wrong by a landslide.
The idea that someone still thinks this in 2014, and is confident enough in his opinion to publish it, is ridiculous.
I think the author wants to express that open source has changed a lot from the early days. As far as I remember it, open source originally was driven mainly by students, academics and hobbyists, and by service companies like Red Hat or maybe IBM. Nowadays, many of the larger, successful and enduring open source projects are mainly developed by professional developers working for companies that use the open source software for their commercial products or infrastructure.
I don't think the examples and quotes used by the author are great, but I do think that there is some truth to this argument.
No?
I'll go back and finish the article, but it's starting to look like the something you'd see in a tweet from @nytonit "Building a business around a product you give away is hard and The Times is ON IT."
Theoretically, yes, if I have the source every change and customization I want to make is just a recompile away.
Practically speaking that almost never happens because now I'm likely forking a huge code base and almost always abandoning the support contract I paid money for to do so.
Unfortunately, because theoretically speaking anyone can change anything, authors seem to neglect the whole documentation and API portion of open-source software, and that's what I typically want in practice.
Obviously there are exceptions, but most people pay money for support, warranties, integration with existing systems, documentation and available training. All of those things have very little to do with whether the source code is open or not.
We sell two WordPress plugins, both of which have free counterparts (think Lite/Pro). The source code is GPL for both versions.
Why do people buy the Pro version of our software? That's we've asked many times over the past few years. I feel they buy the Pro version for both the features and, more importantly, for the fact that it will be _updated_ and _maintained_, and they will have _access to support_. Site owners that want to use one of our plugins recognize that by paying for the plugin they get access to support and features not available in the Lite version, whereas the free version has limited features and only community support.
Could someone copy the source code for the Pro version and sell it? Sure. (And people have.) But can they copy our Pro support and sell that? Not so much. Can they copy our reputation? Nope.
I've learned over the past few years just how valuable "access to support" is for many people. Even if they never contact us, just knowing that we're there _in case something goes wrong_ is comforting enough for them to pay for that comfort.
If more open source projects started offering support, I bet you'd have a lot more open source developers making money.
For example, if I decided to set up `mutt` for the first time to switch over from web-based email, would I buy a "mutt support service" from someone that allowed me to submit `mutt` questions through a ticketing system while I was setting up and getting familiar with `mutt`? Yes! I would, especially if that person had a reputation in the community as someone who knows `mutt` inside out.
It's also worth noting how offering support helps drive documentation. If the documentation is great then a developer can easily point people to the relevant information and save him/herself time (so that support doesn't take up all available time). I basically used this type of "ask me a question, I'll create documentation for it and then reply with a link" method with my Independent Publisher WordPress theme project [1].
[1]: http://github.com/raamdev/independent-publisher
These challenges cut close to some things we are aiming to solve at Assembly [http://assembly.com]. One such question we hope to answer is this: "how can people build a real, profitable company in a collaborative, open manner like they build in an open source environment?"
An example is Coderwall, which makes more than $25,000/month [1] and is built and maintained by the community. Each month, revenue goes to the people who are building the product. All the code is publicly available [2], and licensed to be used non-commercially. A core team of contributors guide vision and quality control, but anyone can participate.
It's not exactly open source, but products on Assembly are built in the open and anyone can dive in and help out. Some products have lofty ambitions to make big revenue, and others are are more aimed at being a public offering like a traditional open source project.
[1] https://assembly.com/coderwall/posts/coderwall-coinholder-up... [2] https://github.com/assemblymade/coderwall