"Open source, when professional, provides long-term stability and removes vendor lock-in. Even during internal development. He is perhaps even more critical. A couple of in-house programmers decide to walk away from poorly documented developments… You know for yourself that this is the case. And you can figure out the rest yourself. "
Milan Kolar is the founder of Pype.club, which helps study animation and visual effects to develop and maintain open source software. Their resources are used by animators and film post-productions from L. A. to Jakarta.
This business is extremely dependent on technology and managing huge amounts of data. In Pype.club, they write data management software for animators and a whole bunch of people around film and post-production. They find out the intersections of SW of all studies and look for common solutions so that everyone can take advantage of them.
They combine the software with other open source technologies and provide it completely free of charge. Commercially, Pipe.club provides companies with installation, support, consultation, training and, of course, further software development. Due to the fact that it is a professional open source, the SW is stable, constantly evolving and the studios have new updates.
The community of Pipe.club programmers is growing rapidly, thanks to which it actively communicates with clients and open source technology is developing so fast.
🔸What is the advantage of Open Source and what is the internal vendor lock-in?
🔸 Why doesn't Open Source mean free?
🔸 Where is the money in Open Source? Which model to choose?
🔸 How to scale the Open Source consulting business?
🔸 What to look out for when choosing Open Source for a company?
THE INTERVIEW
Martin Hurych (Moderator)
Hello, I'm Martin Hurych and this is Zážeh. Today's Zážeh will be with Milan Kolar, founder of PYPE.club, which, as you will soon learn, has nothing to do with pipes. This Ignition is going to be really special because Milan is from a completely different business than we normally discuss here. However, there is definitely a reason for it. Hello, Milan.
Milan Kolar (Guest)
Hey, good morning.
Introduction of Milan and Pype.club
Martin Hurych (Moderator)
We've already touched on pipes, we'll leave those aside. When I said you were special, tell us, what business do you come from and what do you do?
Milan Kolar (Guest)
Basically, I historically come from the business of visual effects animation, film post-production and commercial production to film production. I spent about 10 years in that at the creative level and in various technical positions. I ran a small animation studio here in Prague for 6 years called Kredenc and PYPE.club kind of gradually grew out of that because these animation studios and post-production studios are extremely dependent on technology at the moment. In the past it was drawing on paper, which is beautiful and it's still happening. At the moment it's huge amounts of data that these studios are generating every day, translating, somehow having to manage. And what does PYPE.club do? We write software to manage that data, to help animators and a whole bunch of people around film and film post-production so that they can focus on the creative and not have to worry about where to save the file, what it should be called, what format it should be in. We originally wrote a little internal software, then combined that with other open source technologies that were already available. But nobody has combined them into one big unit that you can install in a studio and say, this is your backbone, this is where you run your studio, and it's going to manage all this data for you. At PYPE.club, we provide that service in principle and we do it in a way that the software is free, it's open source, in fact 100% of the code that we're doing at the moment is completely open source, and the business is built on providing a professional service to those studios. That means installations, long-term support, training, consulting on how to do it and how not to do it.
What is the perception of Open Source in the Czech Republic and abroad?
Martin Hurych (Moderator)
That's a nice way to start out, you said the entire content of the podcast. However, I would start digging into it a little bit more. I was just intrigued by this business model, I was intrigued by a few things about it. My age bracket, a lot of people in the Czech Republic see open source as a bad thing. I saw in the preparation that you, on the other hand, consider open source as a way to gain client trust. So how do you have it laid out, how does it work for you, and how did you come to open source in the first place? Why didn't you program it yourself and, like a lot of software studios here, make boxed software and take it out into the world by way of licensing?
Milan Kolar (Guest)
There are several reasons for this, one of which is probably partly historical. For a long time in post-production and in the film industry in general, it worked exactly the same way that the studios that had the extra and had the business edge were the ones that had the big software departments that wrote it themselves. Those studios had the extra, not just with the creative, but with the technology, they could say, we can make a bigger set, we can make a bigger forest, a nicer forest, bigger water, better water. That got democratized at one point, software started to be bought off the shelf. You bought Maya, 3D studio, installed it and it actually did almost everything. And then the studios, in hindsight, 5, 10 years back, realized that the projects were so huge that they were going to have to work together. There's no studio in the world right now that could do an Avengers, Planet of the Apes, etc. sized movie by itself. It's basically always a collaboration of these big studios, and these are studios that have hundreds to thousands of people. It's impossible to do it in house. And by doing that, they find that they have their technology, somebody else has their technology, and they can't transfer the data in some reasonable form so they don't spend more money on that than they do on production. And what happened was that our whole business started tipping over into open source. By having two studios say together, let's do it this way, and it's already advantageous for us when we're doing it to put it out into the world and allow other studios, smaller studios, to work on our technology. So that it's easier for them to supply us with the data so that we can order it from them. This has been slowly happening, it's all flipped in post-production and suddenly everyone trusts open source a lot more for a simple reason. First of all, it can't be that there's something secretly hidden in the code. It's much more secure, and in the film business, security is an extremely sensitive issue because before the films are released, the visuals simply can't get out. And by their technical teams being able to see the complete code, they can see if it's safe or not safe. They can see how it works, and at the same time, in the long run, if I'm able to manage open source in my studio, I'm able to install it, work with it somehow, and have some professional support for it, then that gives long-term stability because it removes a lot of what's called vendor lock-in, which means that I'll buy something, or order something, or have it created, and then I find that in a couple of years I can't change vendors because my whole company is running on it and I'm dependent on it. And that happens in internal software development, among other things. When I develop my own software, the vendor lock-in is no less because it's developed by two or three people, it's not a team of twenty. These people get sick, get divorced, move away, have kids, and don't want to work as much. I'd say half of our clients have been in that exact situation. We have a data pipeline for visual effects, but we can't work with it anymore because it's aging for the last year because the person who wrote it has left, hasn't documented anything, and no one has any idea what's going on with it. For this reason, open source is gaining a lot of trust in our business.
Where is the money in Open Source? Which model to choose?
Martin Hurych (Moderator)
So far you've described it a bit like charity, because I program something, I give something to someone, they can still do a great job of looking it up, then they install it, they use it somehow. Where's the money?
Milan Kolar (Guest)
It's not a charity. There are more models. We've been validating one model for 3.5 years, we keep changing it a little bit, and now we've come to something that we hope is sustainable. However, it's
not just our invention, most companies that do open source software end up making a profit on a SaaS version of the software and provide it for free, actually counting on the fact that most companies will install it themselves and most users will use SaaS. We can't quite do that here because the software, by definition, simply can't work as SaaS. It's too big, so we're going with the other model, which is based fundamentally on long-term support for studies. That is, they can download it for free, they can install it, but they find out very quickly that it's not that simple, it's a huge complex system, and they pay us for long-term support, for training, for maintaining it for them for a long time, installing it for them, consulting with them on how to work with it, and at the same time, part of that is writing additional features and fixes. We try to combine the fact that the software itself is free in principle, but open source almost never means cheaper than a commercial solution.
Why is Open Source not a cheaper solution?
Martin Hurych (Moderator)
Before we move on, I'd like to stop here. What do you mean? Because open source is seen as something that's free. I often look for open source so I don't have to pay for a license. So what do you mean, in large companies, it's not a cheaper solution? What do I have to give for the cost against that?
Milan Kolar (Guest)
Most often, it's the people managing the solution. Most open source solutions cannot be run in one click and everything is live. Someone has to install it on their own infrastructure and on their own servers, they have to do some work around it. On the other hand, all the companies that are actually making money on it, thanks to some open source solution, they can't afford to do it without some professional support. At least in the mode that we have, they work with the developers and sponsor the project because it's important for them to keep the product going. Just because I download it for free on Tuesday doesn't mean it's not outdated by Friday, and it simply costs something to develop. Some projects do it as donations, completely voluntarily, some do it by support, and if they don't find it, the projects die. And that's where the problem is, it's free, but only until I actually start using it for business.
Martin Hurych (Moderator)
So am I understanding correctly that either I have my own team that takes care of this for me and your business model is that I can actually outsource this team to you and you take care of it for me?
Milan Kolar (Guest)
It's exactly right, and the mode is somewhere in between, of course. There are some studios that have someone who is able to get to a certain level but is no longer able to fix mistakes in it for example. So they come to us with those mistakes.
Martin Hurych (Moderator)
This is suddenly moving from production to consultation, if I understand correctly.
Milan Kolar (Guest)
Yes, that's right.
How to scale an Open Source consulting business?
Martin Hurych (Moderator)
You've got your company spread across major studios from LA to Jakarta, as you say. So how do you keep up? Everyone who comes here and has a start-up dreams of scaling and everyone tells me that it's impossible or very hard to scale people. How have you figured out your path?
Milan Kolar (Guest)
Our business is quite small globally, it's in the thousands, low thousands of potential customers when you take individual studios. There's no comparison. If I'm making software for anybody, it's all over the place. And by the fact that even the local market is relatively small, the studios are here but there aren't many, so we've scaled geographically very quickly and basically on our own. I really think we've got clients on every continent, across every time zone. We're actually at a stage at the moment where we're scaling roughly double every year, but we're already hitting that it's getting quite complex because scaling people is just challenging. What we're doing is we're looking for studios that have some overlap in what they do and what they need and we're convinced that you can write a product without having to fix an awful lot of things for every single studio. That is, a client comes to us, says they need a consultation, they need something fixed, and at that point we compare with other clients or even users. This happens of course and we look for a common solution. Most of the time it actually works out and that way we're able to scale the product and how it expands out into the world and we don't have to legitimately pull in a new person to spend full time on every new thing somehow. Of course, it's still complicated. We've gotten to the point where maybe 30%, 40% of the writing of that code is not us writing it. We're putting ourselves in the position of doing the governance and managing what it should look like. We have that insight into the different studios and into the usage, but a lot of studios want to write the code themselves and they'll ship it back, they'll actually put it into the program, and that's how it suddenly scales because we don't have to scale the specific people who are writing the code and consulting on every single thing, but the studios start telling each other within that community of ours. It's very visible at the moment, and financially it's working too, because then there are studios that are willing to put code into it, but they're also willing to put money into it because it's important to them that the product itself is sustainable in the long term.
Martin Hurych (Moderator)
If I put in my own code as your partner, what do I get?
Milan Kolar (Guest)
In the open source world, this is a big question, of course. Most often it's that part of the software works exactly the way you want it to. That's probably the biggest hook, come with us to do a new feature that no one has done before because now you have the opportunity to define how that feature will actually work. That's probably the most common approach. Of course, once you have x studios and x people writing into it, it becomes challenging and you have to balance it very delicately so that it's not written for one studio and works in another. But so far I think we're doing pretty well and it's all based on being extremely open. If we don't know something, we'll be upfront with the studios. Everybody can come on the call and it's clear to everybody that we don't have the recipe for the system, we're writing it with them. And that builds tremendous trust in the whole community towards that product. And it's really in that collective engagement, and it's terribly important for them, business-wise, to be able to work with other studies, because without that, you can't do it. Visual effects companies have historically been extremely divided, and they haven't had any standards between them for a long time, and we're just now getting to the point where those standards exist.
Martin Hurych (Moderator)
That's what I wanted to ask you, because even though we discussed in a recent episode with Honza Mašek that things are changing a bit and that companies in the Czech Republic are starting to learn from each other, until recently, and in some companies still, there is such co-isolation. We do it ourselves, we do it best here, we don't look to the right, to the left. This, what you're saying, is it just that everybody has to work with everybody else and we're in such deep trouble that we actually have to go forward together?
Milan Kolar (Guest)
It's partly due to that, I'm sure. Of course, there are companies for which it's completely unthinkable, but that's fine, they do it themselves and some do it maybe better, they have the resources, but it's also because it's not easy to find people. Generally in technology, in programming, to find a programmer who also understands the artistic part of our business and can put it together in a reasonable way is almost an impossible task even for us and the studios come to us simply saying we would write it ourselves but we find one or two people to do it and the software is so big that 10, 15, 20 people have to write it and at that point they have no choice or they can rely on it, that it's going to be written by a huge studio like Disney, which has made an awful lot of their technology available for free as open source, and other studios, Sony and so on, but nobody wants to rely on that ever happening, because in that time those studios are another four years ahead of them again, so in order for me, as a mid-sized studio, to keep up, I have to collaborate on everything on a technological level with the others. I can't do it any other way, and to do it I have to put my skin in the game because otherwise the whole product dies.
How to keep up with the market leaders?
Martin Hurych (Moderator)
Yeah, so how do you keep up with these big studies?
Milan Kolar (Guest)
By talking to them. And we talk to them on a regular basis, there are several business groups that actually discuss all these issues on a regular basis and we just try to keep up with them. We coordinate, we use their technology. We're big believers that everything we want to write,
somebody has already written it, maybe they've written it worse than we would have written it, or better, but we can't rely on the fact that something they have, we can write better. We know that X, Sony, Disney or somebody wrote it. For them it's marketing, for them it's just what I mentioned that others are using their technology and we're putting it together and making it a system that gives it a head and a heel.
How to finance Open Source development?
Martin Hurych (Moderator)
If I understand correctly, then for what you're writing yourself then, where you're pushing it forward, you're funding it from your consulting then?
Milan Kolar (Guest)
There are companies that do it explicitly just from consulting. We're of the opinion that you can't do it just from consulting, because either the consulting is extremely expensive and it's not that interesting for the studios, or we don't do anything else, we just do support and consulting and we don't write the product, we don't improve it. I mean, we're very open in that community and we're telling clients how we're distributing the money. If you pay us something, some of it goes to the time we spent on it, but most of it goes to developing the product and there's nothing you can do about it. If you want to work with us, you have to see the development of the product at the same time. In the beginning the studios were looking at us, why should they pay not only for our time but something more. Luckily, not many people in the world do what we do, so we're not completely short of clients and we were able to pick and choose the studios that wanted to go into it with us in this open way. What's happened over time is that now everybody has to accept that if they want to use it, they have to accept that a lot of the money they're putting into it is going purely into development. We've been telling them that up front from the beginning and we hope that it's sustainable in the long term.
Martin Hurych (Moderator)
I mean, the amount that I pay, say, extra to come to me to consult me on a deployment is how big relative to that?
Milan Kolar (Guest)
We're trying to make it fifty-fifty. An hour of your time means you're paying for an extra hour of development. I think right now it's split 40:40:20. 40% of your time, 40% is extra development and 20% of that goes to general research, but it's actually things that by the way we have to keep track of these other technologies, we're openly saying, you're going to pay for the time that we're playing with the technology and nothing's going to come of it. And we will use that 20% to do that and you have to agree with that because if we don't do that, we won't keep up.
Martin Hurych (Moderator)
That is, the 40% is not for development for me, but development for the community. That is, if I need a feature I pay for it separately and in the same adequate amount I pay for development, i.e. for the community.
Milan Kolar (Guest)
That's exactly right. Clients know this from the first meeting and if it's not acceptable to them, they can't use the product. But we haven't had a single case yet, which is interesting, where they haven't accepted that. Companies take it for granted because they want the latest technology. And that's not going to happen if nobody wants to pay for it. We've set it up this way, that we tell them and they have to accept it.
How to build a community developing Open Source?
Martin Hurych (Moderator)
About the community. It seems to be working great because it helps you grow. I think everybody wants a community these days. Every other business mentor recommends it, every first one has one, including me. How do you get that community to build up so that it actively supports what you're doing, it talks to you, it's often not only your fan, but it's your driver, that community is pulling you forward, how did you achieve that?
Milan Kolar (Guest)
I think this is much easier software than in other areas because in that software, when they get involved, they have a 100 percent feel of that ownership of the product. They're written in there and we communicate with them on a really level. Our community is about 370 people right now and we run on Discord, somebody has Slack, somebody has Facebook, it doesn't matter. It's around 370 people now and it's growing at an average of 1-2 people a day. It was slower before, now it's running and those people must feel like it's partly theirs. Just the fact that if someone helps us fix the documentation, we immediately put that person on the list of people who created the product. It's online, it's in the product, it's everywhere, and it's a microscopic psychological effect that really works. So in software I think this is easier, but of course we spend a huge amount of time doing it. All my programmers, all the people who do client support for us, they're also talking on that community every day. And we answer every single thing that comes up there, everything. Most of the time, we're responding within a few minutes. It's because, to make it simple, we live on Discord, the whole company communicates on Discord, we communicate with clients on Discord. So it's logistically manageable, but if the community doesn't live every day, it dies. For example, now we have 50 features open to see if they're going to be in the product or not, and we didn't write 30 of them. Someone else wrote it, so it takes us a very long time to even check if it's right, if it makes sense, but we believe that this is how we can scale in the long term. Not by us, but by the people around it. We'll build trust, they'll write the code, and that'll scale the whole thing.
Martin Hurych (Moderator)
I'm very impressed that a lot of small software studios that I'm around can't imagine connecting a programmer with a client. You have them linked on Discord?
Milan Kolar (Guest)
100%. We've got it to the point where we have a client that lets 50 employees on their private Discord channel, I let 15 of my people on there, and it works. We tell clients they can manage it themselves at home, but if they do it with us, we have enough people to help them. It doesn't mean they know them 100%, but they can help them. And we just do it openly, everybody knows about everything, logistically it's very challenging because you can't keep track of all the clients and you can't keep track of everything, but if somebody wants to, they can.
Martin Hurych (Moderator)
And now I'm wondering if you even keep track of how much time a programmer spends programming and how much time communicating with clients.
Milan Kolar (Guest)
We have tracked everything that we do for clients and it's more or less in the percentages that I said, about 40% we spend actually on the clients, 40% on the writing of the code and 20% on the ballast around that which can't be identified exactly but is very important. It's answering Discord and things like that.
What to look out for when building a non-traditional business model?
Martin Hurych (Moderator)
As a software studio, if I wanted to step outside the uncharted waters of SaaS or on-premise box solutions and wanted to go your way or another uncharted style, what should I look out for? Or how to do that, how to muster the courage to start trying something different that might work out there? Because your model works relatively successfully out there, but in our waters it's something unknown. So how to do it?
Milan Kolar (Guest)
It's very much based on who the target audience is. Doing open source for everybody, which is database systems, these kinds of things, is extremely challenging because when I release a product, I need to have huge hundreds of thousands of users, each of whom will give me a dollar, for example, as a donation to write it, or I need to make it SaaS as well. That's actually what all these big companies that have huge systems do. But I don't really have a recipe for that skill, because we, if you imagine if we had tens of thousands of users, we would have a huge onslaught of users who want fixes and don't want to pay for them, they keep asking how to use it, it's really challenging. So we went down the niche market route on purpose, which we know very well, and I think if a company has this model of knowing their target audience, communicating with that target audience, not abstract people somewhere on LinkedIn, but knowing enough of a sample of those people, it's really a lot about that openness. Everything has to be public. We also have conversations about what we're going to post and what it's probably going to look like. Everything can be tracked on GitHub or here, and that's the only thing that builds that community. Those people then believe that this is not a product that they're going to use in version 1 and version 2 will already be paid for. I think the openness is the most important part of it, but it's not for everyone, it costs time and it costs money of course. There are start-ups that scale up tens of thousands of users in two years. We have in 3.5 years low dozens of studies, 30, 40 studies using it. On the other hand, it's such a critical thing for them that they're willing to pay pretty good money for it to make it sustainable.
What to look out for when I choose Open Source for my company?
Martin Hurych (Moderator)
When I look at the other side of the barricade, we've already bitten the bullet, and as a business owner I might consider an open source solution. What to look out for and how to avoid the usual, which we've also talked about here, it's free, it's cool, it won't cost me anything?
Milan Kolar (Guest)
For me, first of all, look to see if there's a company behind the product or if it's completely a community product. I'm not saying there aren't purely community projects that aren't great and half the internet doesn't run on them. But there are a minority of them.
Martin Hurych (Moderator)
Can you name some so we have an idea? To give us an idea of the seriousness or otherwise of open source.
Milan Kolar (Guest)
In fact, of the ones that are purely community-based, I don't think I can think of any off the top of my head. But they're mostly smaller libraries that most websites use. They're written by one person or a few people, but they're pretty small projects. For the big projects, you have to be careful if there's a company behind it that's also making money from it, because if it's not, it's suspicious. So if there's a company behind it that has other income, for example, Facebook has released an extreme amount of open source technology that the whole world is using because it's clear that Facebook is using it itself. They're doing it as a pronounced promo and it's amazing the stuff they've released. Right now, a really big part of the internet is running on React and various technologies that have come out of these big companies. That's where it's safe. And then the companies that are making money on it either through SaaS or like us, that there's professional support for it, then there's not much to worry about. I've been pleasantly surprised that our clients went for it with us in the beginning because we had no track record and they actually went against exactly what I just said. Here's some company that pretends to do it, to support it, but actually has nothing behind it. I'd probably personally go more for a project where it's clear for at least a year or two that they're developing and then it's awfully easy to see if the project is alive. You look at the code on the internet, there's no need to understand it. I go to GitHub, find the project and see how active the repository is. There it's visible, there I can see that if 1000 changes have happened there in the last year, that means it's running and it's fine. If there's been two changes in a year, it's probably not quite right. While that might be totally great for me, I'd be worried that it's already dead, or the person in question doesn't care much about it anymore. That's the most important thing, if the activity is in the product, then it's mostly safe.
Martin Hurych (Moderator)
Okay. Thanks for a peek under a whole different hood. If any listeners or viewers want to contact you and ask how to do that, where can we find you?
Milan Kolar (Guest)
Either I'm on LinkedIn, under my name people can hopefully find me there, or email me at milan@pype.club. I'm happy to reply or connect. Of course we have websites and they are very professionally oriented, so pype.club is our website and there is more about us there.
Martin Hurych (Moderator)
Okay. Thank you for visiting.
Milan Kolar (Guest)
Thank you very much.
Martin Hurych (Moderator)
So it was Zahžeh with Milan Kolar. As I was saying, today we peeked into somewhere completely different, where at least I personally have no chance to see. If you're interested, if you have maybe a small, larger, medium-sized software studio and you're willing to try some new model, we'd love to. For the rest of you, definitely keep following us on YouTube or another podcast app. That's it for today. I thank you for your attention, I wish you success, and I'm crossing my fingers, thank you.
(automaticky přepsáno Beey.io, upraveno a kráceno)