Episode 83
· 50:28
Having, like, recognition within your peers was kind of like, if you focus on that, then generally some sort of commercial opportunity will arise.
Jack Bridger:Hi, everyone. You're listening to Scaling DevTools. I am joined today by Ben Romich, who is the CEO of Flaggedsmith, which is a feature flagging open source feature flagging tool. And Ben is also the host of the craft of open source. Ben, thank you so much for joining.
Ben Rometsch:Yeah. Sure. No problem. It's great to be here.
Jack Bridger:So this episode is gonna be a really interesting one, I think, because I put this kind of, like, call out on Twitter of, like, are there any, dev tools which are kind of open about their numbers? Because I feel like it's we often don't have, like, a a real, like, transparent view of the growth of, like, early stage dev tools or, especially. And I know that you're not really early stage anymore, but it'll be super interesting to hear about, your progress. And, also, extra interestingly, you're bootstrapped, open source, so you're taking, like, a lot of very interesting boxes. So, yeah, I wondered if you could tell us about FlaggedSmith and also how you got started.
Jack Bridger:This episode is brought to you by Work OS. At some point, you're gonna land a big customer and they're gonna ask you for enterprise features. That's where Work OS comes in because they give you these features out of the box. Features like skin provisioning, SAML authentication, and audit logs. They have an easy to use API and they're trusted by big dev tools like Vercel as well as smaller fast growing dev tools like Nock.
Jack Bridger:So if you're looking to cross the enterprise chasm and make yourself enterprise ready, check out WorkOS. We've also done an episode with Michael, the founder of WorkOS, where he shares a lot of tips around crossing the enterprise chasm, landing your 1st enterprise deals, and making sure that you're ready for them. Thanks, Workhorse, for sponsoring the podcast and back to the show.
Ben Rometsch:Yeah. So, yeah. I mean, it's kind of crazy now. Flowsmith's, like, 14, 15 people. It's been growing pretty rapidly in the last couple of years, but, it actually, it started, quite a long time ago and the sort of growth rate was very, very slow at the at the initial, kind of inception of the project.
Ben Rometsch:And, actually, the the project lived, I think, for, like, 3 or 4 years before we even incorporated a company, like, a a legal entity around it. So, yeah, so, like, the kind of the the the positive origin story is that, I've been running a, or, yeah, I was been running a software agency in London. We built software for people, built apps, cloud transformation for large companies, APIs, all sorts of software applications, front end, back end infrastructure, all manner of stuff. And, we used to build, when we had downtime because agencies are are kind of feast family businesses. When we had downtime, we used to build, stupid things, like, that has zero financial, potential, basically, but just things that we thought would be fun to do.
Ben Rometsch:It just kind of to let off steam a little bit, I guess. And then, around in 2017, we decided to build a to kind of scoring model for for ideas, because we were always having ideas about things. And as engineers, you know, like, everyone's built a I feel it feels like everyone or maybe, like, 15 years ago, everyone was building, like, support ticketing systems or, you know, like, everyone had, like, their own side project agencies. Like, you know, most of those came out of agencies. And and so we we we put together a set of criteria to to try and rate these ideas.
Ben Rometsch:And, Kyle, who I was working with at the time, he still works on Flowsmith, and still works for the agency. He, was looking at feature flags for one of our customers. And at the time, like, there wasn't really an open source tool. There was George Darkly who kind of invented the the segment and, you know, started building a really impressive tool, but it was quite expensive. And so yeah.
Ben Rometsch:So he he wrote the original kind of 1 page or 2 page document, which was the idea for Facebook. And it was like way it was way better than it scored way higher than any of the other ideas. Like, the main two reasons were, a, because we were the target customer, ourselves, which is, like, really, really valuable and powerful. You know, if you're, like, building a, you know, I don't know, like an app for new moms or whatever, and you're a, you know, 25 year old man, then, you know, so that that was that was that was super powerful part of the idea. And then the other one was that you can build a really kind of decent MVP without a huge amount of code.
Ben Rometsch:Right? You don't have to do a huge amount of heavy lifting. You know, you can get started with something pretty quickly. So, and and actually and the I remember there's a so there's a there's a GitHub link for an issue in Firebase remote config. So we almost didn't start the project because Firebase at the time had a remote config component that was really good, but their SDKs only supported it on mobile native SDKs.
Ben Rometsch:They didn't have a web SDK. And, Kyle actually, this is, like, 7 years ago posted on that issue and said, like, why is that why, you know, why isn't there web support for this? Like, we really wanna have web support. We're thinking of building a feature flagging tool because this this remote config feature doesn't have, like, web support. And it got, like, 40 thumbs up or something.
Ben Rometsch:And so we were kinda like, that's kind of weird. Like, you know, you don't normally you comment on a GitHub issue, it's pretty unusual to get 40 thumbs up. So, yeah, so we got some validation there. And, yeah, like, we just started building in our in our spare time, basically. So myself, Kyle, and and Matt, who's now the CTO, we just spent you know, if we had some spare time during the day, we'd work on it.
Ben Rometsch:And, then the 3 of us a little bit in the evenings as well because we were just kind of interested in it. And so, yeah, so we, you know, it it didn't really because we were literally doing it in our spare time. You know, it's kind of analogous to your solo founders who build something literally at, you know, in their evenings and weekends or whatever. We kind of did that as a company sort of thing. So, yeah.
Ben Rometsch:So I don't we never kind of, you know, if we'd have if someone would have paid us to build it as an agency, I think we probably would have spent maybe sort of few £100,000 working on it. Agencies in London are quite expensive, so, that might sound like a lot, but it's not a huge amount of actual developer hours working on that. But, yeah, like, we just we started building it and and, and nothing happened at all. Like, no one started. No one was interested in it.
Ben Rometsch:Absolutely nothing at all.
Jack Bridger:Yeah. So so and that was for, like, 3 years.
Ben Rometsch:Yeah. That was that was, yeah, that was for, like,
Jack Bridger:a couple
Ben Rometsch:of years, and, we built this at the time, it was called bullet train. We did so many things back at the at the start of the project that we gave no thought to whatsoever. So, like, the the idea for the name, I was just like, you know, I don't know, thinking about going to Japan and I was like, why don't we call it Bullet Train, which is a terrible name, because you're you're if you type bullet train into Google, you don't get a new, you know, newly creative feature flagging project that no one's heard of. You get you get trainers. And so, we built a really sort of sketchy website, and we put a hosted API up.
Ben Rometsch:We thought, okay. We'll we'll build a hosted API and people will pay for that because they don't wanna have to host it themselves. It's kind of like what everyone else is doing at the time. You know, that's how that very influenced by by GitLab. You know, that's kinda how they were making money.
Ben Rometsch:But it's so kinda interesting, actually. That's how it looked like they were making money, but, you know, I think a lot probably a large proportion of their revenue now, I'm sure it's public information, is is is in is in, you know, enterprise on premise installations like ours is now, which we could talk about later. But, yeah, like, we we just very, very, very slowly started to get, paying customers through to our API, basically. But it was literally like 1 of 1 a month or 2 a month, and we were charging, like, I think at the time, the cheapest plan was $29 and the most expensive was a 100 a month, I think. So it was really, really slow to get going.
Ben Rometsch:And, obviously, like, if you're making, like, you know, if you're if you're an agency and you're making $500 a month out of service, you know, that's not even gonna cover I mean, you know, it doesn't even cover, like, the lift maintenance service con do you know what I mean? It's like
Jack Bridger:Yeah.
Ben Rometsch:It's like it's basically 0. It's it's as good as 0.
Jack Bridger:Do you know where those early customers were coming from?
Ben Rometsch:I I I mean, I guess people who, wanted to use a a service like that, but were maybe more price sensitive, to have like the time we were, I think we were like 2 or 3 times cheaper than LaunchDarkly, but they had like a way, way, way more powerful products, way more sort of like fully free fledged products. At the time, we only had like I think we supported, like, 3 or 4 SDK languages. So, yeah, I think it was just people who who were who were more price sensitive, basically. Yeah. But we we admit we we also knew immediately learn how, sort of much of an issue or how problematic hosting of a feature flagging API can be because the, you know, if if you're, let you know, let's say you're an issue tracking SaaS company, your your traffic is gonna be almost perfectly linear growth with your number of customers.
Ben Rometsch:But we had very early on, we had a big UK customer. I think it was the postcode lottery, and they just literally signed up. And then immediately, like, they were doing, like, 5 times more traffic than all of the other customers combined because they had a fairly popular website, and it was on their it was on the the main page of their website, basically. So we suddenly realized, like, you can't and also, like, they had very, very unpredictable traffic patterns because they would get a load of traffic when, they've done a draw, I think.
Jack Bridger:So Yeah.
Ben Rometsch:Like, we were like, we hadn't really sort of considered that too much. And funny enough, that that problem still we we built a a a very, very, very much more scalable, service API that's that's globally replicated and stuff, which means we don't have to deal with these problems anymore. But we still got some customers on the old API, and they still generate you know, occasionally, you know, one of them will send out a push notification, and, they'll get, like, a ton of traffic. And so, yeah, so it we we sort of it was kind of a it was kind of a pain in the ass at the time. It was kind of like because, you know, we're like, oh, you know, we have to we've got this client work, and we're being paid for, but we have to go and sort of, like, nurse the the bullet train API again.
Ben Rometsch:You know? And then, like, Matt, my business partner, he he was kind of like less interested in in it. And he was kind of grumbling, like, you know, why we do so many times. Yeah. Like why we're spending so much time for like to, you know, we make I mean, I don't mean to sound brash, but like an agency charges like £120, £130 for an hour for someone's time and for like a, you know, a senior software engineer, you know, we were making, like, 1 hour's work Yeah.
Ben Rometsch:A month from the platform and, you know, and there was a hosting cost and that sort of stuff. So, yeah, like, for a while, it was kinda like touch and go whether we'd even, we'd even carry on, to be honest. Yeah.
Jack Bridger:Yeah. Yeah. And why did you carry on? Or or, like, what so how did it go from this kind of, like, time sync sort of thing?
Ben Rometsch:It's I mean, it's sort of it slowly got it. You get these kind of like sort of positive feedback loops. So, you know, we got to I remember when we got to like $1,000 in MRI, like that was a, you know, that was like, okay, this kind of, you know, it was slowly the the rate of the rate of change was increasing as well a little bit. And then we were kind of like, you know, like, let let let's just, I don't know, like, build more SDKs, improve the quality of the SDKs, add more features to the platform, you know, improve the hosting. And, you know, if you sort of looked if you squinted hard into the future, you kind of see, you know, the sort of very, very slow appearance of a of a profitable company.
Ben Rometsch:But, you know, you had to squint pretty hard, you know, and, like, I yeah. So that's why it took such a long time because, you know, we weren't we weren't venture backed. We didn't put any, you know, like, the we I I mean, we didn't spend any cash apart from hosting, I think. Like, we didn't buy ads, pay for content markets. Like, all all the all the, value that we were providing to the business was in engineering, effort and production of of the platform.
Ben Rometsch:It wasn't we never spent money on it at all. Mhmm. And so you're naturally, like, constrained in your growth because, you know, you're you're not doing all of those things that that grow you know, yeah, that typically grow businesses, basically. Yeah.
Jack Bridger:Yeah. And it was just organically coming in somehow, I guess, like Google search or something.
Ben Rometsch:Yeah. Like, yeah, we were slowly, I think, actually, like, you know, like, open source speech flags. I do remember we we we started ranking on page 1 for that fairly easily. It'd be much hard to see that now. But, yeah, like, so, you know, and it's like, I was also kind of fairly certain that the direction of travel for companies like feature flagging companies or projects was commercial open source.
Ben Rometsch:Like, that just felt like a natural thing that would that would happen. And so engineers would more and more so wake up one morning, get get to work and go, I want an open source. Do you know what I mean? I'm gonna go and find an open source feature frame platform. 1 must exist.
Ben Rometsch:Right? And and at the time, that was kind of one one of the main reasons why we open source it was because it didn't basically. So, yeah. So so that that was the kind of yeah. Like, people weren't actually going I don't know.
Ben Rometsch:Yeah. Like, like, a common thing was, like, I've used I've changed job. I used all the stock clean my old job. My g new job, they can't afford it. And so Mhmm.
Ben Rometsch:That's like, what are you gonna do as an engineer? You're just gonna Google, like, what you know, can I self host it, or is there something cheap, basically? Yeah.
Jack Bridger:Yeah. That makes that makes total sense. I could totally imagine people doing that. So I guess maybe I don't know if this is a good boundary to draw it to, but, like, how did you go from, like, 1 k MRR to, say, 10 k MRR?
Ben Rometsch:Yeah. So, there's this the simple answer to that is we did a deal with 3 American guys who, had started a kinda interesting unusual company called Polychromes. So we met them very fortuitously. I was getting called by loads of VCs because, they're all kind of, like, massively, deal flow constrained at this time because the market's ex you know, the VC market's exploding. And we were, like, way too small, and it I was gonna give up, stop answering the emails, and stop taking the calls because they were a waste of time because I kind of wasn't really that interested in raising money, and we were tiny and growing really slowly.
Ben Rometsch:So they they weren't interested in us either. But one of the guys on the last call that I took was a friend, with Matt Orthouser, who founded Polychrome with his 2 friends. And they he put me in touch with Matt, and that was just very, very, very lucky. It was just as COVID had started. And so I remember it very clearly like being in a little room in the in the, you know, loft of my house.
Ben Rometsch:I met Matt, and and immediately, like, within 10 minutes, it was like, you know, this is just gonna be the best deal of my life. Like, not, you know, like and it was. And so because they had a very, very different model. They they gave us a small amount of money and agreed to work themselves on the business until it got to 40 k MRR. And basically so they they'd invest sweat equity in the business and they all had like Matt was CEO of Amplitude.
Ben Rometsch:Alex was a finance director at Twilio and
Jack Bridger:and Wow.
Ben Rometsch:Yeah. So yeah. So they had these site and Greg Greg had been working as head of sales in a number of big value companies. So and they they kind of didn't know anything about engineering, open source, developer tooling products, engineering, anything like that, but we didn't know anything about go to market, or building and selling software to large organizations. And that's how they'd all, like, made their careers, basically.
Ben Rometsch:So we had this site perfect. Yeah. It's it's just this perfect deal, basically. The the timing of us being introduced to them was absolutely perfect. So, yeah, that was that was us off to the races, basically.
Ben Rometsch:We were at the time, like or before that was happening, I was that we were starting to get leads from large companies saying, how do I host this ourselves? And then we were just kind of like
Jack Bridger:Just have a go.
Ben Rometsch:It's open source. You can host it yourself. And then they were like, no, no, no. We want, like, an enterprise license. We want our own contracting.
Ben Rometsch:We need our own mass service agreement. You need to be insured. You need we need it deployed on Kubernetes. You need to help get us up and running and sizing the infrastructure and all that sort of stuff. And we were like, I was like, I know how much does all that cost?
Ben Rometsch:And I was like, I've got no idea. Like, I've absolutely no idea because I've been working running a small company for like 20 years. And so I literally had no idea how much people pay for that. But Matt and Greg and Alex knew exactly how much people would pay for that. That's and that's that's how we started growing.
Jack Bridger:Just you know, like ballpark just for people to know, like, roughly what?
Ben Rometsch:So, yeah, like, you start basically, if you're if you're selling software like this to large companies, if you say a number that's less than 5 digits, they basically don't take you seriously.
Jack Bridger:And 5 digits per month for 5 digits. Oh, yeah. Sorry. Yeah. So at least, like, 10,000 Yeah.
Jack Bridger:Dollars a a year at least. Yeah. Yeah.
Ben Rometsch:That they're they're literally like
Jack Bridger:What's wrong with
Ben Rometsch:you? These, yeah, like, you don't you you like, you've lost you've lost all trust Yeah. All my trust because you you've you don't know what you're doing basically yet. Yeah. So yeah.
Ben Rometsch:So that's and and then, yeah, figuring out how to package that, how to sell it, how to market it, all of that sort of stuff. Like, we had no idea about really we knew how to do the the you know, knew a lot and and cared very deeply about kind of selling it to the engineers.
Jack Bridger:Mhmm.
Ben Rometsch:And but the engineers aren't the people that are buying the actual software. You know, it's the procurement depart departments that are buying that. So it's kind of we kinda had this interesting quandary because and this is still the case today. Flowsmith gets it gets the decision to buy it gets made by the engineers in a quite a bottom up manner. Like, you know, a a team lead in a department of a large organization would be like, let's use that.
Ben Rometsch:Like, we've done a POC with it. We really like it, you know, blah, blah, blah, blah, blah. And so we we were really well tuned to talk to those people because we were those people Mhmm. Pretty much. But then the actual, like, point at which they say contact us to say, like, you know, we wanna buy an enterprise license for FlashSphere for 3 years.
Ben Rometsch:How much is it? How much is it you know, what is it what do I have to pay more for this and this for that? Blah blah blah. That was where Polychroma, like, you know, designed and built all of the stuff around, and that is an operational business. Yeah.
Jack Bridger:Yeah. And, actually, going back to, Polychrome, so they put in some amount of money. Is that, like, amount of money for you to go work on it full time or is it
Ben Rometsch:So that that was like because we were like, well, like, if we if me and Matt and Kyle spend, you know, the next 3 months working on this full time, then there's gonna be a hole in the the revenue of solid state. And the revenue of solid state is paying everyone's mortgage or or rates, so that can't happen. So, basically, yeah, exactly that. Yeah.
Jack Bridger:And so for, like, a year or for, like, all of you for working on
Ben Rometsch:Well, and then the so that, you know, and then we had a we had a cash flow projection of, like, what we thought the growth rate rate would be and at what point because because 40 k a month was also, like that was the point at which it could pretty much be self sustaining in terms of, like, it could pay Amazon, it could pay me, you know, it could pay a couple of engineers, and, yeah. So, like, that that was yeah. So there was a it was it was it was very much like you're kind of a runway really in terms of, like, you know, like, if we hadn't have if we hadn't have got to that point, the speed that we predicted, it would have been easy because we could have just spent less money for the next 6 months or whatever and just waited for the revenue to catch up. But it turned out that we didn't need to do that because, they were really, really good at selling enterprise licenses of the platform. And back then, it was, like, pretty easy because there was, like, one competitor Yeah.
Ben Rometsch:In the open source And you
Jack Bridger:were cheaper.
Ben Rometsch:Than the space, basically. So well, there was one SaaS competitor, but that we weren't even we weren't even competing against them because the companies that we were talking to literally couldn't buy them. Like, they'd have been fired, if they'd have bought them, basically. So Yeah.
Jack Bridger:Like, we yeah. That's amazing. And so they those guys were full time on, the
Ben Rometsch:3 of them were
Jack Bridger:yeah.
Ben Rometsch:I think they did another deal as well, with Browserless who are like a Oh, yeah. Headless, like, Chrome, platform. So, yeah, Joel and I were, like, the first to that investments isn't quite the right word, but I don't know. But, like, engagements they had.
Jack Bridger:And and then,
Ben Rometsch:yeah, back then, like, they were, like, just, like, full time, like just full bore. Like, let's design this. Like, and we we actually had a we did we changed the name of the company partly because there was a Ruby on Rails start framework thing called bullet train, and they
Jack Bridger:Oh, yeah.
Ben Rometsch:They released, like, 4 months before us. But we were kind of we we really realized we used to change. So, you know, we did, like, we changed the name, designed a new brand, built a new website, put together a bun you know, got a master service agreement written by a lawyer, all of that sort of stuff, and then just started, like, pounding sales as hard as we could, basically. Yeah.
Jack Bridger:Yeah. And that I guess that was probably, like, largely connections with stuff they had in the space.
Ben Rometsch:Or No. No. It was main no. It was mainly inbound leads that we were getting. Oh, really?
Ben Rometsch:We yeah. Yeah. People contact us and, like, we wanna buy or we're thinking of buying your software. Yeah. I mean, that's how enterprise software gets sold.
Ben Rometsch:Right? Yeah. Like, that's that's what happens.
Jack Bridger:Inbound stuff and then following up. Yeah. That's really cool. So it just, like, took off essentially in terms of, like, revenue. So Yeah.
Ben Rometsch:I mean, it it didn't it wasn't like, I don't know. You know, it wasn't like I was coming down from the loft and saying, like, right. You know, like, we can go and move to wherever. I don't know. Like, not in a hill or whatever.
Ben Rometsch:But but, Expensive place. Yeah. Or not not familiar
Jack Bridger:with London.
Ben Rometsch:But, you know, like, it wasn't it wasn't like that. And Yeah. But it but it but it was like it was deaf you definitely felt like we'd unlocked a a massive, like, sort of massive aspect of the business that had been basically sort of, like, sat there the entire time, but we just didn't know what to do with it, basically.
Jack Bridger:That's that's amazing. Are you able to say where you guys are at in terms of revenue these days?
Ben Rometsch:So we're like we're closing in on 3,000,000 ARR now. Last year, we grew, like, 250%.
Jack Bridger:200 and 50%. Yeah.
Ben Rometsch:So, yeah, so we're growing really quickly. The team's growing, so, like, we're doing you know, reinvesting it in in the team. But, yeah, like, the the market's maturing a lot. It's also, you know, it's maturing in in different ways as well. Like, there's there's more open companies, there's more closed companies, but there's also much more of a spread and variety in terms of, like, who they're selling to and what they're selling because there's kind of a lot of some platforms that experimentation see themselves as experimentation platforms that use feature flags as the way of bucketing and modifying user experience.
Ben Rometsch:And then others like flags with a much more like design for engineers and engineering teams to try and improve their velocity and aren't so focused on experimentation, multivariate testing, and and things like that. So there's kind of this kind of quadrant that's sort of appearing out of the midst of, like, oh, what which which is at the way that we see the market, which is, like, open and closed in one dimension and kind of designed for experimentation and product managers in 1 in one end and and designer for engineers and engineering teams at the other.
Jack Bridger:Yeah. Yeah. That makes that makes sense. So it's more like rolling back stuff, like
Ben Rometsch:Yeah. And concept up. Yeah. Yeah. And and and things that, you know, you know, like, I don't know, like, change requests for feature flags and approvals, feature flags, the scheduling them and things like that, like compliance stuff that the bank would need, for example, whereas, like, the experimentation side is more or, you know, would would spend more time working on things like, I don't know, like user session tracking and, you know, like, the the the hardcore statistical stuff and and things like that.
Jack Bridger:Yeah. Yeah. I think I I when I used feature flagging before, it was, like, it was Firebase, and it was really painful because it was well, it was just, like, very rudimentary, I guess. And, like, we were doing growth experiments, and that was what we used it for.
Ben Rometsch:So I guess yeah. But, I mean, you know, a lot of engineering teams and a lot of products that, you know, people are building, they just have a they just have a tough time just deploying software.
Jack Bridger:Yeah.
Ben Rometsch:And, I do I do one of the things that I feel quite strongly about is that, the the sort of engineers are my I I kinda overly influenced by Valley West and West Coast stuff. So, you know, if you're a Valley company and you're paying $400,000 a year to your average engineer or whatever it is. I mean, it's probably a bit less now, but, you know, you've got, like, this crack team. You've got, you know, you got massive bundle of money that you've you've raised relatively easily. I know things have changed a little bit, but, you know, like, you know, you're you're you've got, like you're building a platform from the ground up using all the latest frameworks and languages and technologies and infrastructure and all this sort of stuff.
Ben Rometsch:And so it's kind of trivial for you to go, oh, yeah. No. Yeah. We'll start doing, like, a b and multivariate testing and, you know, optimize the onboarding process and all this sort of stuff. Most most most software engineers don't most software engineering teams, that's not their life at all.
Ben Rometsch:They've got a 15 year old Java application that's running on IBM WebSphere that no one knows. Do you know what I mean? It's like it's horrible. It's hard. It's like secure you know, you're getting, like, hacked because you've got 15 year old Log 4 j dependency that's been do you know what I mean?
Ben Rometsch:It's like Yeah. So I I do feel like a lot of engineering is is is, sort of overly influenced by this kind of, like yeah. Like, not not there's anything wrong with what those companies are doing. It's just not the reality for most engineers. You know?
Ben Rometsch:Like, most engineers, it's like it's just hard for them to get their code released. It's hard for them to fix bugs. It's like they're working in, like, large, complex, old, decaying platforms, not you know, like, what you know, that have been supported for 15 years, that have had, like, 7 different people working on the code base that you're looking at at the time. You know, like, they've you know, there's all this stuff that they can't use because they're on, I don't know, old infrastructure or or whatever. You know?
Ben Rometsch:That that's the reality of most engineers, and it's not, you know, like, you know, all all roses and, you know, like, 15 builds to production a day. That's not how most people live.
Jack Bridger:Yeah. And what do you think are, like, the things that they care about that my teams who come from that perspective might not realize?
Ben Rometsch:Well, yeah. I mean, they they just they just want something simple that'll help them, you know, like, deploy their software. And so that's the kind of I mean, that that's that's the kind of the beauty of of using feature flags is that you get, you know, a large amount of the value of the from the of the concept by just using them in a very, very simple way by you know, of decoupling deployment and releasing features. Like, that's that was that was why we designed that was one of the reasons why we built the platform in the first place. We had firsthand experience of working with, a big customer or a couple of big customers.
Ben Rometsch:This happened twice in succession where they they couldn't release their software because, you know, they do a 2 weekly release. They said it was agile, but it wasn't like that. You know, they were like, oh, yeah. We'll we'll release on the second Friday of every, you know, every other Friday. And they they were like large teams that were all interdependent on each other, and then someone would find a bug on Wednesday, and then that would scrap the release on Friday.
Ben Rometsch:And then they'd wait for the following 2 weeks on Friday, and they had 4 weeks of code to release, and then it would happen again. And it just got worse and worse and worse. And that was the that was for that was when Kyle was like, what why don't we use feature flags to to solve this problem because we can not have to, deal with this problem of, like, we like, there was a so we were getting these these teams that we were working, and we're getting in this situation that everyone knew that when we deployed it, it was gonna break something or or multiple things. And that wasn't it wasn't anyone's fault. It was just the circumstances that those teams were working in.
Jack Bridger:Yeah. That makes sense. So they're kind of pushing code that's buggy out in the release, but then they're just making sure that it's not actually activated until it's fixed.
Ben Rometsch:Yeah. So it's like, you know, or like there's some feature that we're waiting on. The team a is waiting on teams b and c for some additional thing, but the teams b and c are delayed because blah. And so, you know, like, the team a, like, hold the release of team b and c and then, you know, like, that will get, like, worse and worse and worse. So, yes, you know, feature flags, like, you don't need to worry about, like, you know, progressive rollouts or multivariate flags or AB testing or anything like that.
Ben Rometsch:You just you just want you just want a button that turns the feature on or off. Yeah. That's it.
Jack Bridger:Yeah. That makes sense. And then sorry. Just going back to, like, when you said simple, maybe this is, like, something that's, like, obvious to you, but it's, like, I think it's quite interesting because I don't like, what what what makes something, like, simple or not simple to, like, an enterprise custom customer, like, big enterprise?
Ben Rometsch:Well yeah. I mean, they they want they want the platform to be easy to to host and to understand. They they want, you know, they they want really, really, really good support getting their teams onboarded and getting the platform up and running, you know, because it's quite a big deal. Like, if you've got, you know, if you're a large organization, you know, like, we some of our our large customers, they're running, like, you know, mission critical stuff for, you know, a massive multibillion dollar company and, you know, putting feature flags in that in that infrastructure is like, a nontrivial thing to do for them, sort of. Not necessarily technically, but just as a kind of, like, state of mind type of thing, basically.
Ben Rometsch:And so, yeah, they they just, they wanna sort of, you know, like, dip into the pool as gently as possible rather than dive in head first. You know what I mean? Whereas some of our other customers would just be like, right. Yep. You know, we have this, like, high functioning team will do that, and we'll suddenly put, like, a ton of traffic through the platform and all this sort of stuff and we're good to go.
Ben Rometsch:But others like, you know, it can take them I don't know, like it can take them like a year to get, to go from like, you know, buying the platform to actually, you know, starting to use it and get, and get value out of it just because it's they're just used to moving just very, very slowly, basically. Yeah.
Jack Bridger:Yeah. That makes sense.
Ben Rometsch:And that's not like that's not a criticism of them at all or the engineers on those teams or whatever. It's just, like, just culturally what for, you know, for all there's there'll be, like, a number of reasons why those companies move slowly or those teams move slowly or or, you know, they're just not comfortable, yeah, like, with with making change to the to the process and and to the platforms that they're working with. Yeah.
Jack Bridger:Yeah. That makes sense. I guess if, like, you're HSBC or someone, like, as a customer, I'd I don't care that I don't want HSBC to just be, like, AB testing and just absolutely, like, redesigning the whole platform to slightly improve retention. And and, like, I just want them not to lose my money.
Ben Rometsch:Yeah. Like, you know, if you're a regulated UK bank, you you you can't you basically can't do that. Like, you can't you need to have, like, you know, sort of a chain of of of evidence of, like, making changes to your application and your platform and and things like that. You can't just, like, have some person randomly toggle a flag in your platform and some big behavior, you know, that that's like kind of you know, it you can literally go to prison if you did that in an egregious enough way. Do you know what I mean?
Ben Rometsch:Like, as the CTO, for example. It doesn't often it doesn't happen, but, you know, if you're in if you're in a regulated industry like that, you know, you you you you can you can it can be a criminal offense to to do something that that's, like, you know, professionally negligent. So, yeah, that you know, that's that's a good example of why a lot of people don't move quickly yet.
Jack Bridger:Yeah. That makes sense. And then they like flagsmith because they can see all the code, as you said, simple deployment, and they haven't got to, like, yeah, rely on you to, you know, if you go down and that's like a big risk for them.
Ben Rometsch:Yeah. Yeah. So, I mean, there's there's we do a lot of work around, like, defensive coding and, like, educating people who are using the platform about using default flag values and things like that. But, yeah, like, you know, it's it's, the the other thing as well is, like, you know, the platform's open so they can contribute to it. One of the things that I've been most surprised by is that it's really common well, not really common, but it's not unusual for people to buy the platform and then us to get, like, 5 pull requests from one of the engineers that's been on the call as the in the sales process because they're like, oh, you know, the Java SDK doesn't have, like, capability to do, I don't know, like, JSON logging or or whatever.
Ben Rometsch:So, like, here's the pull request to do that.
Jack Bridger:That's really cool.
Ben Rometsch:Yeah. So, like, that you know, and they can't I with SDKs, they could probably do that with with closed platforms, but definitely not, you know, part of the core part of the core products itself. So yeah. So that that's been that's been something that's been really surprising, actually. Yeah.
Ben Rometsch:It's very it's I think probably half yeah. I think I think, you know, I don't know. I I we'd we'd have to do some analysis on this, but, like, some significant percentage of our external third party contributions will be from people who are paying us money to run the platform. Yeah.
Jack Bridger:That's really cool. Ben, so, 2 final questions. One is, have you got any kind of key takeaways from this that you would advise to another founder that was maybe a bit of an earlier stage than FlaggedSmith?
Ben Rometsch:Yeah. Well, I mean, I think for us, like, we just kept at it. Right? Like, we just, you know, like I mean, this is true for a lot of things in life. Like, if you just get up and do a bit every day, slowly, those things accumulate.
Ben Rometsch:And, you know, like, we were the the the the the sort of the the nice thing about working on on Facebook is, like, it was an interesting thing to start working on, you know, just just from a technical point of view. It's a bit like it's a bit like a, you know, like a sort of a to do list or a a bug tracker. Like, you know, everyone would have an opinion on on how that should work and what the processes and workflow should be and what it looks like and all that sort of stuff. And so we we've we've you know, like, if we we we never we never started writing that code with the goal of, like, hitting 10,000,000 ARR or anything like that. Like, we never started it about that.
Ben Rometsch:And we we just started it because we wanted to really, we wanted to have a a successful open source project. And, that's been like one of the re really cool things like, you know, like the the the first sale we made was kind of exciting, but, like, I remember the we got, like, around about the same time, we got a contribution from this random guy who we'd no one had ever spoke to before who'd, like, upgraded Django or no. He'd upgraded from yeah. I I can't remember whether it was on the j I think he upgraded Django and pie the Python version that he ran on and, like, just gave us this big pull request. And we were like we we were just like in bits.
Ben Rometsch:Like, we couldn't believe someone is sort of, like, contributing their time to do this. Yeah. And it was like all of this work that he'd done was like, we could just, you know, review it, merge it, and then, you know, we didn't have to worry about doing that. So that was kind of a bit of a light bulb moment for us. So for us, it was very much, I think, being, sort of driven by, like, non like, not you know, making money wasn't like the the the primary focus of the project.
Ben Rometsch:It it was it was just building something that, you know, like, we had a sales call the other day and, like, 2 of the people on the on the customer side had used Fracic Week in 2 separate previous companies. And that was like that was like re you know, I was kinda like jumping up and down mentally in my chair going like, wow, that's kind of nuts. So, yeah, I think, like, having, like, recognition within your peers was kind of like the you know, because getting a pull request is kind of a valid you know, that that's kind of like a a sort of like a byproduct of that, I guess, in a way. So, yeah, from an, you know, from an open source point of view, and and then, like, you know, I think if you do that and you get, you know, some traction in terms of, like, contributions or stars or just people raising issues or or wanting to come up with ideas or whatever, like, I think if you if if if you focus on that, then generally some sort of commercial opportunity will arise as a function of that rather than trying to do it the other way around.
Jack Bridger:And that
Ben Rometsch:that's something that I think, like, you know, like, we we never went out. We, you know, we just saw a flight smooth and go, right. Like, not not many people sit down and go, I wanna start an enterprise b to b sales business Yeah. Because they you know, it's actually not as bad as people make out. It just takes a long time, and it's you need to have knowledge about things that, you know, you you probably don't have knowledge about.
Ben Rometsch:But, yeah, like, think that so that kind of right appeared out of the mists as kind of like, right, this is definitely a strategic thing to do and and kind of, like, realign kind of the priority of features and stuff based on that. So I think, yeah, like like, just being consistently but working on the project in a consistent manner and not not I mean, I guess it was easy for us to say not focus on making money because, you know, we we kinda had this kind of free engineer resource. But, you know, I mean, if you're an if you're an engineer yourself, then, you know,
Jack Bridger:you You're also You're free also.
Ben Rometsch:You are also free as well. Yeah. Yeah. So I think that would be but I guess that's 2 things, but that's miles.
Jack Bridger:Yeah. So stick with it, and just do it for almost this intrinsic enjoyment or because you want to do it essentially at first, make a great open source project.
Ben Rometsch:Yeah. I think so. Yeah.
Jack Bridger:Yeah. That's really good. And then finally, are there any dev tools, outside of Flightsmith that you're kind of excited about at the moment?
Ben Rometsch:I mean, I we we I I spend, like, a huge amount of my time in GitHub at the moment, and I'm kind of really I love the the the sort of the tooling ecosystem that's building up around GitHub, especially around, like, actions and stuff automating things and just so it's I guess it's not well, I guess GitHub actions, I guess, would be the answer, but, like, the way that the way that the when they designed it, we use GitLab CI a lot, which is basically just a it's basically like a YAML file which does bash commands in Docker. And when when GitHub actions came out, I thought they were, like, horrendously overdesigned. But, I I I just I really love them and, like, there's stuff that we've found, you know, like, I don't know, like, if we release a version of the core platform, there's an action that we found that then goes to our Kubernetes helm chart and bumps the version of that and creates a pull request for that. And just the kind of like or like just just things that add quality of life improvements to engineers in terms of, like, being respectful of their time.
Ben Rometsch:You know, so like those like, you know, if we like to do it to do a production deployment of our core platform now, we've got a, there's a really good one called release, please. And that came out of Google, I think, where, it looks at your pull requests and it creates a new release for you as a pull request. And then when you merge that PR, it tags a new version of the application and and then all our sort of tests run and then our stuff gets deployed into production. So I mean, it does all the semantic versioning for you as well. And if you create conventional commit titles for your pull request, it creates a release notes for you and all this sort of stuff.
Ben Rometsch:And so I just I really love all of that, like, convention. If you if you do things in a certain convention, you know, like, you you can just basically spend was it as an engineer, you can spend, like, such a large large proportion of your time just writing code rather than Mhmm. You know? Because doing all the crap that you you know, 15 years ago, you had to do. What you know, whereas now it's like, you know, all of this stuff.
Ben Rometsch:Right? There's, like, release notes that are automatic. All the versioning is automatic. Like, all of this, you know, all the testing, everything is like. So I really and yeah, like for those people old enough to know what software engineering was like 20 years ago, where you see it, you felt sometimes you felt like you were spending half your time just like, you know, doing 0 value stuff.
Ben Rometsch:Like, that's something that I've just been, really, really yeah. Finding a new GitHub action that will add some small quality of life improvement to the the team in the processes, that that's like I I I just I love finding those. Yeah.
Jack Bridger:Yeah. That's that's awesome. Thank you so much, Ben. Yeah. Cool.
Jack Bridger:Where can people learn more about Flacksmith?
Ben Rometsch:Yeah. So github.com/flagsmith, that's like our main, like, home, I guess. We've got a discord, community that, you know, if you've got questions or ideas or problems deploying the platform or anything like that, like, Discord's probably the best place to go. Yeah. They're the they're the 2 places that we're we're generally present on.
Ben Rometsch:Yep.
Jack Bridger:Amazing. Well, thank you again, Ben, and thanks everyone for listening, and we'll be back again next week.
Ben Rometsch:Thank you.
Listen to Scaling DevTools using one of many popular podcasting apps or directories.
