▲ 186 points
back
145 comments
Kind of Meta: The meaning of "business model," in a software/startup context is somewhat quaint:
“All it really meant was how you planned to make money.”
"Model" seems grandiose for a word that could be simplified to "income source." Comparisons to restaurants tend to be strained. Restaurant "business models" are not generally that complicated: People pay restaurants for food. Some deliver. The reason software people contemplate business models and restaurateurs don't is that "software," regardless of licensing, doesn't come with a standard business model. Every one is a snowflake.
In software, "business model" is a wide range of possibilities. Most of them failing. Many quitely succeeding. A few explosively winning. Is AirBnB a software business model? Is Adwords? Even MSFT, which does literally exchange software for money... MSFT's "business model" is a complicated, impossible to replicate thing.
The majority of software that's literally exchanged for money is done with some sort of services/support bundled in... and the customer is a business. If I give it away then why do people buy it?" is not even a sensible question. What does that question mean for Google, Apple or Amazon? There are plenty of exceptions: Photoshop, MS office... but even these are becoming online services and the "business model" is less directly "software for cash."
TLDR, software doesn't have a standard business model regardless. Weighted by revenue share, the vast majority of software revenue comes from advertising, luxury electronics, web services or some other "business model" that isn't contradicted by OS in any way. "Selling software" in a sandwich-4-cash comparable way is surprisingly rare, regardless of license.
> Every [software business] is a snowflake.
As much as I appreciate this comment, it's less limited to software than you paint it to be. When you start digging into how businesses actually make money, you quickly realize it's rarely what it seems.
Many gas stations that are wildly successful, such as Sheetz and Wawa, don't make their money selling gas, they make it selling food using gas as a commodity to get the customer in the door. Likewise, McDonald's doesn't really make its money selling food either, although this is a lot more complicated and heavily involves real estate and franchising.
In other words, nothing is really as it seems when you peel back the veneer. Software just happens to be relatively new and is poorly understood by most people.
This does not consider b2b at all, where selling software in a sandwich-4-cash comparable way happens on the orders of many, many billions of dollars per year.
Open source is an implementation strategy for closing a business model. Strategic use of open source can move a business model from a money loser to a money maker.
It can be extremely strategic to use source components in the supply chain of a software product. Using open source components allows the engineering team (if one is needed at all) to focus on the unique parts of the project. This can reduce development/maintenance costs and timelines. It can also make up for missing skills on a business team. Facebook did this with Open Compute Project.
It can also be strategic to release parts of a product as open source. For example, releasing a database, OS, or framework as open source can establish an industry standard, improve brand awareness, quality, and make it easier to hire new talent. Basecamp did this with Rails. Google did this with Android.
Open source can also commoditize a market segment to a point where large companies that might otherwise compete choose not to. WhatsApp did this with Erlang and their initial messaging stack.
However, if you’re releasing 100% of your intellectual property & technology as open source then you might attract enough attention to get funding, but you probably don’t have a business model.
> if one is needed at all
Depends. If what we are releasing is a product of engineering (even someone else's), we'll need someone that at least thinks like an engineer in the woodpile.
In my experience, engineering isn't really about balancing binary trees, designing FFT implementations, or writing functional code.
I believe that it's a discipline, and an approach to an organized, end-product-oriented, problem-solving way of working. I've met people that I'd definitely classify as "engineers," that have jobs with no obvious connection to tech, like soldiers, masons, carpenters, and even janitors. I've also met people with advanced technical degrees that can't take an engineering approach to save their lives. Their code and products have been abysmal failures.
As long as I've been in tech, there's been this myth of the "engineer-free" product. The one that is dreamed and executed by Corner-Office executives. A product that jumps straight from their genius, into customers' hands; with none of those messy tecchies fouling the mix.
Maybe, one day, it will become a reality, but I haven't seen anything that even comes close. I believe that the current "dependency-first" development model is not actually a bad one, but it is one that can go horribly, horribly pear-shaped, if done without a disciplined, deliverable-focused, engineering approach.
For myself, I've always enjoyed doing open-source. It gives me a nice track record to which I can refer, and it has forced me to do top-quality code. We always clean the place before we do an open house. If we do it often enough, it becomes habit. I'm currently working on a closed-source app, but that has little to do with IP or security, and more to do with a marketing embargo. I look forward to open-sourcing it in the future. Significant parts of it are already open-source (almost all written by Your Truly). It isn't really meant to be a profit-making app, but it certainly knocks a lot of commercial apps into a cocked hat; and it's only half-done.
Not a guarantee of quality, but it does work for me. I think that Atlassian open-sources all of their stuff. They seem to be doing OK (I am not a huge fan of their stuff, but can't argue with their success).
Your point about commoditizing an existing market segment is extremely underrated.
Companies should use open source where the existing solutions are blackbox but also commoditized. If everyone's doing it but keeping their doors closed, you have a chance to create an open standard. If it's successful, all the black box companies have to adapt it and basically have now wasted a bunch of time
I don’t agree with reduced maintenance cost because all internal patches must be ported to newer versions of the component all the time.
Open source is to software makers what Groupon is to restaurants: a way to get people in the door, but many of those people are not, and will never become, the customers you seek. Every open source business sells something else, often a solution to its own complexity, such as managed instances.
You hit the nail on the head, I think ghost blog is relatively easy to set up, but for $9 a month and you'll handle my hosting and my ssl certs, and most people with a bit of disposable income are fine with paying that. I just spent $80 on a nice meal and a couple of drinks, if my blog is anything like an income source I'll have no problem with spending $100 a month for someone to manage it for me.
I would still advise students, and other new developers to go through the motions of hosting things themselves via AWS or whatever. If you can get a Ubuntu instance on ec2 to run a public facing website you probably can walk into a tech company and get an entry level job. It's not necessarily hard to learn how to do these things, it's just tedious. Avoiding the tedious drama is how open source companies make money
Who are these imaginary foes who are presenting their business model to investors with simply the words “Open Source” on a single slide?
Early stage investor here. You’d be shocked. It’s extremely en vogue to start a company and have your value prop be “open source alternative to X,” without any regard for whether there is actually any logic to something being open source.
Open Source isn’t a business model but it CAN be a support model - for the right market the number of companies who will be willing to fund features/fixes that the company can’t be bothered to get noticeable.
It can also be shown as a “backup” - if we fail you have the source and can continue to maintain it.
I have worked at a few companies that had contracts written such that if we went out of business (or stopped supporting the product) they would get access to the source code. We would regularly upload our source to Iron Mountain for escrow as part of our contract.
"backup" is one of the main reasons I prefer open source. Even if it is funded by one company now (and I pay them) there is some sort of continuum plan if they go bust. I am not immediately screwed and I can consider options such as self-maintenance, forks or a migration. This also applies if they change the terms or price.
What this article misses is the presence of bad faith actors who make billions from open source but studiously work to prevent even a small amount of that flowing back to the companies that wrote the software. Kind of like the difference between a restaurant failing because they can't get enough customers vs. failing because people keep robbing them.
This is a bad faith analogy, others profiting from OSS they didn’t write themselves is a feature not a bug. Stealing from a restaurant is not.
Open source tends to be used as a user acquisition strategy for companies in the growth stage not worrying about monetization.
Once they get to the stage of trying to monetize their newfound userbase, they rethink the value of releasing their product as open source.
It's really common for Launc HNs, though. This search shows many just the last two months: https://hn.algolia.com/?dateRange=all&page=0&prefix=false&qu... (with some false positives)
Amusingly, two variants of "open source Datadog" the same week.
Are these kind of OS copies successful in YC context? They are apparently getting accepted.
The mentioned Redhat article Part 2 is a 404. Here's the archived version:
"How to Make Money from Open Source Platforms, Part 2: Open Core vs. Hybrid Business Models"
Open source sits alongside Free Software. Free Software as defined by the FSF is software that respects the users freedom. Freedom to use, copy, modify... They do say you can charge for it, but never claim it's a good business strategy.
Consulting. Services. Education. Those seem like words that might let you make money. You can attach "open source" to those words but it obviously doesnt stand on it's own as a business model.
> Consulting. Services. Education. Those seem like words that might let you make money.
And that's cool but to a very great degree these often boil down to an exchange of time for money (unless by services you mean SaaS which is really just a product business with a different revenue model). You can make a living but you're on a treadmill that you can't ever really afford to get off of, and it doesn't really scale other than by adding more people.
I mean it's OK, don't get me wrong - it's certainly better than a kick in the balls - and you can make a decent living from it. But I don't find it particularly appealing: I'd much rather be involved in creating products that you can sell over and over again. And, if what you've created is of value to others, why shouldn't you charge them a fee to use it?
(Btw, I'd also add paid support to your list but, again, I wouldn't want to build a business heavily based around providing support because support can really wear you down - or at least it can really wear me down. It's absolutely necessary, but not something I find at all enjoyable.)
The big problem is that there is no distinction between an individual user, SME or a big corporation.
The create grounds for exploitation of software developers, social divide and other unintended consequences.
The promotion of open source by big corporations is cheaper in the long run that spending money on R&D, salaries, taxes as there is always going to be a good enough project for such corporation to appropriate and make money off of, without having to pay anything to contributors.
I've been saying this forever. Open source is just a tool. It's a tool you use to attract customers who want it, or employees who want it, or whatever. It's not a product though, and it's not a sustainable business model. The choice of license you use too is a business tool.
Consortia are about the only successful open source business model. The most successful such consortium is the SQLite Consortium, which clearly makes D.R. Hipp and his team very happy. One key to SQLite's business success was keeping the best testsuite for it closed source, which makes it extremely difficult to fork the project, or even to contribute to it, which leaves the team as essential to those who depend on SQLite (which includes a lot of FAANG).
24 hour/365 production support is a requirement for large companies, sell that. Putting features behind a paywall doesn't work when the entire business can be forked and those features implemented for free. See AWS with Elasticsearch.
There's a reason enterprises use MSSQL, Oracle, etc.
Baking bread is not a business model, a bakery on the other hand.
The author linked to a wikipedia page[0] for examples of viable business models, and it interestingly includes "Professional Open-source Model" without explanation.
[0]: https://en.m.wikipedia.org/wiki/Business_model#Examples
Open source is as sustainable as a closed business model in some sectors as long as you are willing to be creative.
I've invested/funded/donated to several OS projects. I don't really expect to make money in most. I do it for the same reason I give large tips time to time.
Somebody has to do it because the value they are providing is IMMENSE.
I ask nothing in return. I give without really thinking about maximizing profit. The way I see it is that when you give to good cause the universe will reward you in ways you don't expect.
I encourage others who are successful to give to OS projects especially folks who work on projects full time with little to no way of supporting themselves because they can't memorize graph algorithms or do whiteboard interviews or neurally diverse.
As much as I have gripes with Github as a corporate entity, it's platform does serve a useful purpose in that it facilitates the movement of capital to OS projects.
It's just where you build software for.
Of software isn't your main product, just needed so you can better make your main product, than open source software can be very helpful for you.
It's a nice way to help each other, and it doesn't change your profit.
But if software is your main product, than is open source very hard to pay your bills.
Also if you build software for a company, very few will pay for something that will be shared to everybody. But I know a hospital that pay to write software for some hardware on Linux, and share it as open source so all hospitals that use that hardware with Linux can use it. That is helping other people if you have solve a problem.
Buy indeed, open source isn't a business model, but a way of life.
I would encourage people to read the article. VM Brasseur has a click-baity title, which is fine, but the point they are making is actually a bit more subtle. I'll quote from the article:
> There is a great number of potential business models, but “open source” is not one of them. It is, instead, one of the many tools that can be employed in order to make a selected business model work as expected.
> Therefore open source ... does not ... concern itself with business any more than food concerns itself with business. If there is a business that has a business model that is not living up to expectations, and if that business model uses open source as one if its tools, it’s illogical to blame open source for the failure.
The gist of the article isn't "open source and business are incompatible", rather, it's saying "don't blame open source for a failure of your business".
VM Brasseur focuses on open core specifically, a subset of an "open source business strategy", talking about what kind of expectations one should have engaging in that strategy, pointing to John Mark Walker's post on "How to Make Money from Open Source Platforms" [0].
As an aside, I would like to point out that there are successes in some businesses pursuing an "open source" strategy, for a broader form of open source, either using open source as a key feature or because the technology was open sourced, abandoned and picked up by someone else. Arduino [1] comes to mind, where they used the open sourced Arduino platform, the IDE, the libraries and the open source hardware schematic and boards, that was essentially abandoned by the original creators and picked up by Massimo Banzi and David Mellis. There were other factors to Arduino's success, such as focusing on artists and educators, there being a need for a cheap, accessible electronics/microcontroller platform, selling physical devices and various other economic factors that went into making the business viable, etc., so maybe I'm being overly pedantic and not really contradicting VM Brasseur's claim.
In case people don't know, VM Brasseur is a big proponent of open source and has written a book called "Forge Your Future With Open Source" [2] and is in the process of writing a new book about "FOSS strategy in business" [3].
[0] https://www.linux.com/news/how-make-money-open-source-platfo...
[1] https://arduinohistory.github.io/
[2] https://pragprog.com/titles/vbopens/forge-your-future-with-o...
[3] https://anonymoushash.vmbrasseur.com/2021/03/19/coming-soon-...
SaaS is out sourced IT. Plenty of SaaS could open source their entire code base and it would have little effect on sales; people would gladly still pay for it to be hosted.
In some categories, having the code open source helps gain adoption early in the process.
So yes, open source is not a business model, but plenty of open source software is monetized these days.
It's often used as a marketing model.
Baseball is not a business model either.
Open Source = the “Atoms” in Our “Digital Periodic Table”
Open source is certainly one way to do marketing, particularly developer marketing. I agree it's not a business model per se.
yes, open source is not a business model, it's a marketing model. All these startups (MongoDB, Elastic, etc.) they used open source like that only. A very good marketing model. These change of terms is also a marketing model. Now the slogan is: It's still open source for everyone, except for AWS!
Unfortunately it is a marketing term now, or new age jargon crap like "dont be evil"
It is not a business model. But it is a good communication and branding strategy
Do we expect copyright licenses to be business models too? Does anyone?
Straw Man
I've seen "not a business model" quipped a hundred times, as if that ends whatever debate is going on. I've yet to see anyone actually argue that open source actually is a business model, as opposed to having their argument bent that way. The author doesn't link to any examples. Who are they browbeating?
I don't see MongoDB arguing that the meaning of open source should address financial incentives for developers, either. I see them arguing that open source copyleft can be strong enough to demand openness from their competitors, as GPL would have in the packaged-software era. Mongo have a financial concern, but exactly to the author's point about means versus ends, they're not trying to write their business into any definition. They're reading the definition to cover a license that works for their business.
If open source didn't concern itself with business, it wouldn't exist. The whole pitch was to business. Cathedral and the Bazaar was a marketing manual for selling manager-types. O'Reilly bankrolled events and meetings. Early on in OSI history, they readily approved several stronger copyleft licenses for specific businesses. MPL. Sleepycat. QPL. IPL. RPL. Arguably AGPL. That's the history.
JMW does say there hasn't been a successful open core product, but he defines success as "ubiquity in the modern data center". In business, there's plenty of success between irrelevance and market domination. Ferraris and fast chargers aren't ubiquitous in modern garages.
What I see here is that open source is a business model to some interested parties. That model is add-on services---integration, customization, support, hosting---and no other. Never mind that dual licensing, open core, delayed release, and others existed well before RedHat and WordPress and IBM successes made headlines. Never mind that these models had their business standard bearers, too. Never mind that they sometimes offered support documentation or configurations for hosting that add-on service companies didn't.
The difference now is that the anti-commercial, permissive-focused, "open source means I never hear 'no'" faction view happens to correspond to the interests of the titans of industry, who are all cloud services companies. These companies want to commoditize their complements---"open source" means we can reap what you sow and sell it to our customers---but not the proprietary code they use to differentiate their clouds. They patronize the big open source foundations.
We shouldn't be asking what open source is or means, as if it's an end in itself. We should be asking what it's for. And accepting there was never any conflict-free consensus there. We brushed all that under a rug in the late 1990s and early 2000s, because there was fame and money and a sense of purpose to be had selling "open source" as a brand and a "revolution". Now that work is done, the victorious allies are squabbling again. So it goes.
I pretty much stopped taking this seriously after they compared restaurants to software development. Restaurants have very different operating cost structures.
A 100%. We try to make as much as possible of our online video editor (www.typestudio.co) free of charge and try to be as transparent as possible in our communication. But with open source we could never raise a seed round.
There are a million ways in which your business can go belly up, there is so much that can go wrong at any point. I see these open source startups and I just cannot fathom how clueless MOST of these people need to be to build their businesses around an ideology.
I guess if you start from a very specific strategy and you make sure that the FREE/OSS route is the optimal for the sake of your business, it can make sense. Honestly if this happens to me, I would probably get discouraged by the track record of such companies, and just backtrack into a model with a better one, in order to give my business the best possible chance.
Whats up with all the opensource hate on here lately? Has Bill Gates secretly bought this site or something?
(User was shadow banned for this post)
Oh yes it is. But like most products it has its market segment. And that segment is a group that requires a certain function that is not a competitive advantage for any in the group and wants to get that function for the lowest possible cost.
Ok, so much for abstractions. A stark example is that on Microsoft's Azure the majority of VM's are running Linux. A lot of companies want cloud based VM's and they don't want to pay for Microsoft's OS to run them when they can use Linux for free, so they don't.
Another example is OpenStreetMap. Many companies (including Apple) spend millions a year upgrading openstreetmaps. A lot of companies need a mapping function but it's silly for each one to carry the burden of creating their own map when they can all pitch in and use the final result.
This is the market segment where Open Source makes sense.