Adventures with Open Source Business Models
This is a modal window.
The media could not be loaded, either because the server or network failed or because the format is not supported.
Formal Metadata
Title |
| |
Title of Series | ||
Number of Parts | 43 | |
Author | ||
License | CC Attribution 3.0 Unported: You are free to use, adapt and copy, distribute and transmit the work or content in adapted or unchanged form for any legal purpose as long as the work is attributed to the author in the manner specified by the author or licensor. | |
Identifiers | 10.5446/67137 (DOI) | |
Publisher | ||
Release Date | ||
Language |
Content Metadata
Subject Area | ||
Genre | ||
Abstract |
|
FOSS Backstage 202410 / 43
8
11
00:00
Open sourcePrincipal idealWeb serviceOpen sourceBlogMusical ensembleProjective planeCASE <Informatik>Adventure gameEnterprise architectureSoftwareSoftware testingGradientPresentation of a groupOpen setComputing platformComputer animationLecture/Conference
01:15
Hill differential equationBusiness modelCodeNumberData miningOpen setMultiplication signMathematicsPurchasingFreewareSoftware developerSpreadsheetHand fanState of matterAxiom of choiceAdventure gameCore dumpTerm (mathematics)Process (computing)Statement (computer science)Information technology consultingVideo gameRow (database)Service (economics)CurvatureSoftwareProduct (business)Open sourceData modelMaizeClosed setSelf-organizationBitComputer configurationPlastikkarteParameter (computer programming)Projective planeDecision theoryCASE <Informatik>Different (Kate Ryan album)ScalabilityPlateau's problemBoilerplate (text)Branch (computer science)Computer animation
11:01
Open sourceMereologyPresentation of a groupLattice (order)ExistenceProduct (business)Personal digital assistantMultiplication signWritingMetropolitan area networkCore dumpInternetworkingComputer configurationCASE <Informatik>Projective planeSoftware developerHorizonRevision controlPay televisionStability theoryComplete metric spaceData model4 (number)Open setEnterprise architectureConnectivity (graph theory)Software maintenanceService (economics)DatabaseGodEmailNumberZoom lensCartesian coordinate systemSpreadsheetPoint cloudSoftwareFreewareBlogWeb syndicationFacebookServer (computing)Form (programming)InformationWeb 2.0RootWeb pageBusiness modelData conversionInformation securityProgrammer (hardware)Online chatSoftware repositoryOffice suiteLine (geometry)SpacetimeQuicksortGoodness of fitWebsiteSingle-precision floating-point formatInternet forumAutomatic differentiationSoftware bugWave packetVideoconferencingDecision theoryMoving averageValidity (statistics)Installation artElectronic mailing listComputer animation
20:48
Open sourceSoftwareMultiplication signBusiness modelProjective planeSoftware developerPoint (geometry)Musical ensembleMathematicsElectronic mailing listCodeMatching (graph theory)Design by contractSpectrum (functional analysis)Level (video gaming)NumberRight angleCopyright infringementCoefficient of determinationInformation securityOpen setCASE <Informatik>Product (business)Data conversionFreewareDistribution (mathematics)Stability theoryCloud computingFigurate numberRule of inferenceSystem administratorData miningOSI modelService (economics)Revision controlDecision theorySoftware testingGoogolGoodness of fitLetterpress printingPoint cloudLimit (category theory)Message passingLecture/Conference
Transcript: English(auto-generated)
00:07
My name is Taris Balog and I work at Amazon Web Services along with Rich Bowen and Spot Callaway who spoke yesterday. Prior to that however for 20 years I ran an open source business based upon a project called Open NMS
00:25
and I assume no one here has ever heard of Open NMS. Oh we've got we've got a few people. So it was a enterprise-grade network monitoring platform and in that 20 years we tried every possible method to make money. And
00:41
so I tried to distill what I've learned in 20 years into this presentation. Now I have a blog called adventuresinoss.com. It just turned 21 years old so you can go and look and I have detailed a lot of that that journey. Some of it hasn't aged well. Feel free to call out on me if that is the case. So how
01:05
many of you are old enough to remember text adventures? Like great then you'll enjoy this. So I made this like a text adventure. So the first thing when I'm talking about open source business models is I'm talking about
01:20
true open source companies. Which means if your software is open source you do not sell software. And when we talk about open source we go back to the cathedral and the bazaar. And so I wanted to start off talking about you know Cathedral was the proprietary software stack where everything was very regimented and we talked about open source being more like a bazaar.
01:43
More like an open-air market. And so as our adventure starts we have the choice to choose between proprietary regimented software or the bazaar. So of course we're gonna go south and go visit the bazaar. Now as you know when you when you start dealing with open source come on in please. You missed
02:05
the team picture. So there's a lot of open source companies. I'm a big fan of Red Hat. I live in the United States in a state called North Carolina which is where Red Hat was founded. But we work closely with things like the
02:20
Apache Software Foundation. So the reason I bring these two up is when you're starting to think about your open source business you have the choice of going corporate like a Red Hat where you're going to be a for-profit company or you can start a foundation. There's nothing wrong with that. I'm one of
02:40
those weird people who looks up tax records of foundations and directors of some of these foundations make a really good salary. So you can make a good living running a foundation and not having to run a corporation. This is a choice that's up to you. But it's a choice you have to make. So let's decide okay we decided we're going to make an open source company. So your
03:07
quest is to create a viable business model and the keyword there is viable. We want something that's actually a business in using open source software. And as I mentioned we're not going to be selling software but as someone who
03:24
did this for 20 years the 20 years I spent doing open source software were the most rewarding years of my life. And so I strongly recommend if you love open source and you really really want to make a business out of it this is something you should pursue. So my story was was kind of interesting. Open
03:42
NMS the project I was involved with I didn't start it. Some friends of mine started it and they had a company called Oculand. Your eye on the network. And they paid a lot of money for that. Anyway so I joined them to build a service and support business around open NMS because back
04:03
in 2001 this is kind of what we started off with. Let's build a services and support business. And so again we're not selling software. The way I used to describe my job was I'm a plumber. So I live on a farm out in the middle of
04:20
nowhere and I do a lot of work with wood. I do a lot of work with electricity. When it comes to plumbing I am horrible. I end up wet. I end up frustrated and I end up having to call someone anyway to fix all the mistakes I made. And so the idea when we were selling our services was you're smart. You have smart people working for you but we focus on our
04:45
product. So just like you would pay for a plumber for their expertise you can pay us for our expertise and we will save you time and save you money and you can get the most value out of this open source product. And it
05:01
worked really really well. We ended up having a lot of people paying us. One of the things that I recommend, you hear the term you know make things easier to sell. I'm a firm believer in make things easier to buy. And traditionally when you talk about consulting there's usually this very long process. You
05:22
do a statement of work, you go back and forth and you create these documents. And I worked as a consultant for many years and I can honestly tell you that 90% of that fancy document is boilerplate. And you just go in and fill in the 10%. So we decided to productize our services. So a lot of people were
05:43
like I need to get started with Open NMS and we'd say okay fine we have a green light project. And that was a week of services and a year of support and you just paid a flat fee plus travel and expenses. Nowadays with everyone being remote there's probably not travel and expenses anymore. But what
06:00
ended up happening was we started to plateau. Like our growth was very very strong in the beginning and then it just started to level off and so we tried to expand. And one of the other things we came up with was custom development. We had a number of customers who would come to us and say Open NMS does 95% of what I need it to do and I will pay you to do the
06:25
other 5%. And that turned out to be a very good business. Now it presented its own challenges because they'll say well I will pay you to do this but I don't want to share. I want you to make it just for me and I'd say okay
06:41
I'm gonna charge you five times more money. And they would go why? And I'm like well you know now I have to keep two branches. I have to keep yourself separate. I have to make sure that it's always when we make changes to the mainline code base that your changes changes work. And I was able to convince the first few custom development customers to let us open source and
07:05
include in the general product their features. And what that allowed me to do is in the future anytime asked me anyone said hey I don't want to open source this feature I would pull out my Excel spreadsheet and say well here are all the features someone else paid for that you're getting for free. And it's
07:23
like so if they came to me and said okay I need this feature and I'd say that's going to cost $100,000. And if they said well it's not worth it to me and I've said well that's a business decision. It's not an open source decision. If this feature is not going to save you $100,000
07:41
don't do it. But in most cases we could make the argument that it would save you $100,000. So you might as well do it anyway because it makes business sense. It's going to save you money and you can join this kind of corpus of special features that people paid for. And so a lot of our research and development was paid for by our customers which was a really cool
08:05
place to be. Now the problem with that is eventually that starts to plateau and you start running into this kind of scalability issues. And so I decided okay I'm going to seek inspiration. So I started looking all around at possible
08:25
ways to generate more revenue. And it was a bit of an adventure so I looked at all these different business models and in my adventure I basically went down this dark alley. So I went down into this dark alley and I was introduced to something I'm going to call you're in a maze with twisty
08:43
passages all alike. And so it was kind of confusing. Everyone was coming at me with different ways of trying to make money and they almost always came down to what we used to call open core. The woman who gave the keynote yesterday had called it freemium which is apparently the the new one. And I
09:07
know a number of businesses that do what I would call the open core model where some the core of the product is open source but there are certain features that they hide and they're kind of proprietary. Ultimately I'm for
09:25
any business model that produces more open source software. So I'm not here to criticize open core. If it is super useful the community has the option to fork it and make it their own. We've seen recently a lot of companies are changing their licenses and the community can respond by saying okay well
09:43
we're going to fork this and we're going to go our own way. But it's something I did explore. Now in here I basically said okay let's look at open core. And we agonized over whether or not we should do that and I'm a bit of a free software enthusiast so I was like very reluctant. And I came across an
10:06
article, now this is somewhat dated, it was an article by a guy named Brian Prentice for Gartner. And he basically said open core is the Emperor's new close. And this is Gartner. This is a very well respected organization and he
10:23
basically said if you were dealing with open core you have to treat it as if it was a proprietary software purchase. I mean I worked for very large companies before I started on my open source journey and if they
10:41
bought a product from even Microsoft or Oracle, if they spent enough money they could see the code. What we're calling source available. It's like hey we need to do our own audit and we're going to spend ten million dollars on this product. We want to see the code. And they would under NDA and everything they'd let you do it. So this idea of oh I'm going to be
11:00
magnanimous and show you the code, that existed long before we had source available kind of licenses. But anyway what was interesting about Brian Prentice article is it kind of put the end to open core which I was very very grateful for. Now what happened was companies would come out and they would have an open source component but they weren't exactly advertising
11:23
themselves as an open source company like we were. Because it used to be frustrating. I'd be at a conference and someone would come up to me and go yes but what about your enterprise version? And we're like well we only have one version and it's the enterprise version and it's open. It's free. So we definitely don't want to, we're just going to drop the whole
11:43
idea of being open core. And anyway I had to throw this in here because if you go down the open core path you're likely to be eaten by a group. Anyway there are other options. So I'm trying to find inspiration and part of
12:00
it was this company called Red Hat. You know Red Hat was probably the first and most successful open source company in existence. So my friend Spot who's the guy in the back, I said I want to meet Jim Whitehurst who was the CEO of Red Hat and he's like fine. He says write Jim at redhat.com and tell him I sent you. So I wrote a letter and probably within two hours he wrote me
12:25
back and it was obvious he wrote me back not Andrea his assistant. He's like sure. He says talk to Andrea my assistant and get on my calendar. And for about three years I got an hour of his time to talk about you know what does it mean to be an open source company. And we would come out with these
12:44
different ideas about how we could change our product. And I think if you ever look up in the US we have this legend of a man named DB Cooper who hijacked an airplane and jumped out of the airplane with hundreds of thousands of dollars and he disappeared forever. That also happened in North
13:01
Carolina where I live. If you ever look at the drawings of DB Cooper and a picture of Jim Whitehurst they are very similar. But anyway nicest guy in in the business that I'd ever met and so we started adopting a Red Hat model. What I found was customers were more willing to pay for stability than
13:23
they were new features. So at Open MMS we created two versions. We had horizon because the horizon is always moving and that was our public development version. And then we had meridian because meridians stay in place. Meridian was our little it's a version that lagged and you had to
13:42
have us a subscription to get access to our repos. They were password protected repos. Both products were published under the affair of GPL version 3. So they were both open source and if you if you read there's nothing wrong with with charging for open source. Now people who got our meridian product there was nothing preventing them from sharing it. But I
14:04
found that most people if they pay for something they're not likely to share it. And if it had happened I would have considered it a win. To me if like someone thought long and hard enough that they wanted to come and create like a centos out of our product. I would have been perfectly happy with that.
14:23
But anyway so. So dealing with Jim it is. So we came up with this idea that companies when it comes to open source companies are willing to pay for three things. They're willing to pay for simplicity. They're willing to pay for security and they're willing to pay for stability.
14:41
Otherwise they'll just do it on their own. And so anyway. We took that to heart. And I really want you if you take anything away from this this conversation remember those three S's simplicity security and stability. Now in preparation in preparation for
15:00
this I went to chat GPT and said what's the best open source model. And chat GPT wimped out and said well there are about six different models. So I wanted to cover some of them. And so I have these little three stalls here. And one of them that it suggested is advertising. And I don't know a single product or project that makes a living
15:23
like this. I can imagine perhaps you could come up with an open source project that people would want to come to your website or what's frequently enough or your forums frequently enough that you could serve ads. Maybe you could do training videos that served ads and you could generate revenue that way. So I'm throwing it out here
15:43
in completeness but it is not necessarily something I would pursue. I don't I can't imagine a situation where that would be a viable business model. Source for you. Yeah. So does anyone here remember source forage? That was the question from yesterday. Now there is
16:02
another option which everyone says is well you could do basically tips. This would be you know GitHub has GitHub sponsors and you could basically put out your tip jar and say here this is free. You know give us give us money. I in the early days of open M.S. I published my Amazon wish list and every so
16:22
often someone would spend 20 dollars or something and buy me something like in the movie Office Space I really want to do one of those red swing line staplers and someone bought me a red swing line staplers. But you can't eat a red swing line stapler. So again I don't think this is a viable business model. One that I wish I was talented enough to think about.
16:45
This quote here is comes from a post by this guy named Felipe Valsorda. And back in 2023 he published on his blog. I'm now a full time professional open source maintainer. Holy shit it works. And what he did is he
17:02
was he's a programmer. There's a good example of someone like this is Eric actually right here. If you build enough of a reputation in a particular open source project you can go independent and find people who use this project to sponsor you. Now you have to learn how to
17:20
manage these invoices and price things and do billing and things like this but you can be an independent contributor and get paid for doing open source. Like I said I am not a programmer. I have programmed the very few pull requests I submitted. You could see the CTO roll his eyes. It's like dear God not another one from Taurus. It's
17:42
just you know I'm not talented enough to do that. But I honestly think that is a valid business model. If you can build a reputation in a community where other companies find your project valuable you can go and do that custom development thing. I will make these new features. I will fix these bugs and you pay me for it.
18:03
But ultimately I came to the decision that the best model you can come up with is what I call managed open source. And the example company I use is WordPress. I'm old and I have this thing called a blog. There was this thing
18:22
called Google Reader back in the day and people had these things called blogs and instead of having Facebook or Instagram or the Snapchat we had Google Reader and it would go out and using a technology called really simple syndication would put everyone's blog in a way you could consume it very very easily. And one of the
18:41
companies that made this blogging software is called WordPress and I use WordPress. I mean it's great. They have this thing called a five minute install. And for me it takes maybe three and a half minutes. Basically you pull up your server you set up a database user you download a tarball you drop it in your web root you go to a web page fill out some form fill out some information hit
19:03
go you have a web page. But that assumes you have a server that assumes you understand how to do a database user. And to me I have to go always look it up. But that's pretty much two minutes of the three and a half minutes for me to install it. But if you don't want to do that you can go to WordPress.com and you can
19:21
get a free Web site as long as WordPress gets to serve ads or for like 10 bucks a month they will do it for you. That is that whole simplicity thing. And while WordPress may not be the ultimate example for for business models I honestly think that being true to open source yet creating a business that can scale like I
19:43
used to get when we had a services business I used to get emails that said I can double your sales and I'm here. God no I don't have enough people and that's not a problem you want. You want that problem. When Zoom came out I got a Zoom account I went and signed up for Zoom. No one at Zoom knew I existed right. I was a
20:01
number on a spreadsheet somewhere. And as a business person I love knowing my customers and I love working directly with my customers. But the idea to generate revenue without having to bring on headcount is kind of the Holy Grail. And one way to do that is manage open source. Anyway so I'm going to host my application.
20:20
I'm going to host it in the cloud or I'm going to host it on prem and I'm going to make it. And so congratulations you have won the open source business model. And anyway as a reward there's a crystal flask on the marble. And so let's get you flask. This is a very obscure Internet reference to strong bad. But in case you haven't seen it anyway that is my
20:41
presentation. I hope it was useful. I assume you have some time for some questions and be happy to answer any question I can. Questions.
21:01
Thank you for the great talk. My question is about the custom development. You said that you convinced the customers then to open source it. Had you had situations where the custom development wasn't really interesting for you to integrate into the open source mainline. That is a beautiful question. One of the so my
21:22
biggest regret in open NMS is we could have been Datadog. I don't know if you know Datadog but they're a monitoring their cloud based monitoring company worth billions and billions of dollars. A friend of mine was like employee number 10. He's 10 years younger than me and he's retired now
21:40
because they had a huge IPO. We had ideas but because we were bootstrapped we had no VC money. Our business model was called spend less money than you earn. It's crazy. I know the whole idea but what we ended up doing was sometimes a customer would come to us with a feature that was only good for that customer and it didn't. Now
22:03
luckily the majority of time like S&P version 3 support a bank paid for that and that everybody loved it because you know that when S&P security became an issue. But that's a decision you have to make. And we decided that we were going to do this work. And of course it was open source. We
22:20
were test driven development so we could keep you know every time we made a change to the main code we made sure that this feature this obscure little feature worked. But it does take away from something that may make you a lot more money down the road. So it is something that you have to balance and I think we did it wrong. You know if I had to change it I would have probably turned down said sorry that's really
22:42
obscure. It's going to take a lot of time and our resources need to be focused on our roadmap versus this little side road. But yeah it's a great question. I don't know the answer. I know we answered it wrong. Oh one question before that. Oh one question from
23:03
the online audience. Would you say that the passing of liability is also a reason why companies pay for open source. Yes. We again looked to Red Hat and Red Hat had a clause in there. It's called indemnification. And it's like hey I use your software and there's a
23:22
problem. Will you protect me. We were a tiny tiny company and we just copied Red Hat which offered indemnification at 4x your price. So if you paid us for support let's say you paid us fifteen thousand dollars a year for support. We would indemnify you up to sixty
23:41
thousand dollars. So that was kind of hidden in the contract. You know that there was a limit on it but we would talk about indemnification. I know in the world of A.I. right now there's a whole big talk about we you know if you get sued for copyright infringement we will protect you. And so they do look at that and that is a
24:03
selling point. In a lot of cases we ended up selling support just because the company that's what they did. I can remember it was this great conversation when we very first started our company. So what distributions are you going to support. And I'm a free software
24:20
enthusiast. I'm like Debian. That's it. We're going to support Debian. And my business partner was like well what about Red Hat. I'm like well Red Hat's kind of you have to pay for that. I don't want to do that. He was like well you realize people who pay for Red Hat are more likely to pay for our support. I love this idea. This is a great idea. We should support Red Hat. And so I did
24:44
have my mind changed. But yes indemnification is definitely something that that customers look for. You mentioned the need to grow. How about staying stable at a nice level. Would that work?
25:01
Yes. So I was I've been thinking a lot about the the was it really open security talk that we had yesterday. Now if you ever deal with venture capitalists they will pat you on your head and say you have a very nice lifestyle company. I have a friend who is worth 10
25:20
figures and he has a very nice lifestyle company. We didn't have a very good lifestyle. Like we worked really really hard and we made money and everyone got paid. But it was if you can work it out. There's nothing wrong with with doing that. And it seemed like the really open or radically open
25:41
security company. They had a beautiful lifestyle company. They got paid. They were able to donate a lot of money to charity and they got a lot of value out of it. And trust me when I was young and old people like me would say hey you need to do something you enjoy. And I'd say not just for the money. And when I was young I was
26:01
like no I'll make a lot of money and I can buy the things I enjoy. And it's not when you're doing when you're when you're running your own business you're probably at it 10 12 hours a day if you're doing it right. You got to enjoy it. And if that's it if you can enjoy it and make enough money to meet your lifestyle. That's that's amazing. With us we
26:21
always struggled with trying to make enough to grow. Now you've had your hand up since the start. So I want to make sure you get the microphone because this person right here. Thank you so much. I have a question with all the examples that you showed. You were exploring the business to business world. And I
26:41
was wondering if you also had experience when it comes to selling or providing services to public administration because in my experience public administration also wants stability and security and simplicity but they're not as willing to pay for quality as other businesses. So they want all these these these things but
27:03
they will choose when there's a tender they will choose the cheapest vendor. And then the project fails and they say open source sucks. We wanted it but it didn't work out. And I was wondering if you had any experiences or advice in that regard for open source businesses who deal with public
27:20
administration. Sure. Rule number one never sell on price. I mean it's one of the things you're trying to do when you're when you're running a business or open source it's free software. Never sell on price because you're good. You have a great product. You solve things you can eventually it could be like the third or fourth thing but never go into it
27:41
selling on price because everyone just goes on with the cheapest thing. The second thing is to say no to customers like our starting package was about twenty thousand dollars and I would be on the phone with a potential customer and they'd be talking and they would say you know oh I want this. They had this huge laundry list and well what's your budget. And they said twelve thousand dollars. I'm like get what's up Google. I'm like because you just can't even get
28:02
started with us. And when you run an open source company you need to be more particular about your customers because if you have a mismatch if you sell into a company that really isn't a good match you're going to spend way more effort supporting that customer than you would if you just said go away. With public companies we
28:21
actually did have a number of cities like the city of Portland was a customer of ours. And it was it seemed to be fine. They seemed to get it. In the U.S. again Portland is kind of on the more liberal side of the spectrum in America. Well it's probably the liberal side of Jesus. But the but they they had kind of this motion
28:44
for open source this this print you know open first and if they couldn't find anything that's open source. So we had an advantage and that's something you can play to your advantage when you're dealing with with with companies with with with public institutions. But like you said some public institutions are lowest bid wins. And the
29:03
thing is do not go that way. If it's not a good match and if you're not getting paid for your work you've got to say no. And there are times where you just really want that income and you want to say yes say no. You'll appreciate it. Thank you. Thank you very much. Well thank you
29:20
again for your attention. I'll be around and be happy to see you. Cheers.