TRANSCRIPTION 🔗
WEBVTT
00:00:00.160 --> 00:00:03.279
Hey, it's Jason, host of the Code with Jason podcast.
00:00:03.520 --> 00:00:04.480
You're a developer.
00:00:04.639 --> 00:00:06.240
You like to listen to podcasts.
00:00:06.400 --> 00:00:07.759
You're listening to one right now.
00:00:07.919 --> 00:00:13.679
Maybe you like to read blogs and subscribe to email newsletters and stuff like that.
00:00:13.839 --> 00:00:14.960
Keep in touch.
00:00:15.199 --> 00:00:23.440
Um email newsletters are a really nice way to keep on top of what's going on in the programming world.
00:00:23.679 --> 00:00:26.800
Um, except they're actually not.
00:00:27.199 --> 00:00:36.479
I don't know about you, but the last thing that I want to do after a long day of staring at the screen is sit there and stare at the screen some more.
00:00:36.640 --> 00:00:40.240
That's why I started a different kind of newsletter.
00:00:40.320 --> 00:00:43.679
It's a snail mail programming newsletter.
00:00:43.920 --> 00:00:44.479
That's right.
00:00:44.560 --> 00:00:51.600
I send an actual envelope in the mail containing a paper newsletter that you can hold in your hands.
00:00:51.840 --> 00:00:59.280
You can read it on your living room couch, at your kitchen table, in your bed, or in someone else's bed.
00:00:59.359 --> 00:01:02.159
And when they say, What are you doing in my bed?
00:01:02.399 --> 00:01:04.879
You can say, I'm reading Jason's newsletter.
00:01:04.959 --> 00:01:06.159
What does it look like?
00:01:06.640 --> 00:01:12.640
You might wonder what you might find in this snail mail programming newsletter.
00:01:12.879 --> 00:01:20.480
You can read about all kinds of programming topics like object-oriented programming, testing, DevOps, AI.
00:01:20.719 --> 00:01:23.760
Most of it's pretty technology agnostic.
00:01:24.239 --> 00:01:42.000
You can also read about other non-programming topics like philosophy, evolutionary theory, business, marketing, economics, psychology, music, cooking, history, geology, language, culture, robotics, and farming.
00:01:42.239 --> 00:01:46.239
The name of the newsletter is Nonsense Monthly.
00:01:46.560 --> 00:01:50.560
Here's what some of my readers are saying about it.
00:01:50.879 --> 00:01:57.120
Helmut Kobler from Los Angeles says, thanks much for sending the newsletter.
00:01:57.280 --> 00:02:00.879
I got it about a week ago and read it on my sofa.
00:02:01.120 --> 00:02:06.400
It was a totally different experience than reading it on my computer or iPad.
00:02:06.719 --> 00:02:12.159
It felt more relaxed, more meaningful, something special and out of the ordinary.
00:02:12.960 --> 00:02:15.199
I'm sure that's what you were going for.
00:02:15.520 --> 00:02:18.879
So just wanted to let you know that you succeeded.
00:02:19.199 --> 00:02:20.800
Looking forward to more.
00:02:21.919 --> 00:02:31.520
Drew Bragg from Philadelphia says, Nonsense Monthly is the only newsletter I deliberately set aside time to read.
00:02:31.759 --> 00:02:44.159
I read a lot of
Load full transcript
great newsletters, but there's just something about receiving a piece of mail, physically opening it, and sitting down to read it on paper that is just so awesome.
00:02:44.479 --> 00:02:46.080
Feels like a lost luxury.
00:02:47.360 --> 00:02:57.919
Chris Sonnier from Dickinson, Texas says, just finished reading my first nonsense monthly snail mail newsletter and truly enjoyed it.
00:02:58.240 --> 00:03:03.120
Something about holding a physical piece of paper that just feels good.
00:03:03.520 --> 00:03:04.479
Thank you for this.
00:03:04.639 --> 00:03:06.240
Can't wait for the next one.
00:03:06.560 --> 00:03:16.319
Dear listener, if you would like to get letters in the mail from yours truly every month, you can go sign up at nonsensemonthly.com.
00:03:16.960 --> 00:03:20.000
That's nonsensemonthly.com.
00:03:20.800 --> 00:03:22.159
I'll say it one more time.
00:03:22.400 --> 00:03:25.599
Nonsensemonthly.com.
00:03:25.759 --> 00:03:29.360
And now without further ado, here is today's episode.
00:03:49.919 --> 00:03:54.800
Hey, today I'm here with Julian Donjo, CEO of Mergify.
00:03:54.960 --> 00:03:55.680
Julian, welcome.
00:03:56.479 --> 00:03:57.759
Thanks, David.
00:03:58.800 --> 00:04:04.800
So you and I met because we are evaluating your product to use where I work.
00:04:06.479 --> 00:04:11.919
And you and I talked and you helped us understand the product and stuff like that.
00:04:12.080 --> 00:04:24.240
And I found it interesting because I am working on a product on the side of my own that's in a similar area of the CI, CD, and stuff like that.
00:04:24.399 --> 00:04:26.639
And so I was like, oh, interesting.
00:04:26.800 --> 00:04:30.399
Let's let's have Julian on the podcast to get deeper into this.
00:04:30.639 --> 00:04:34.480
Anyway, do you want to tell us a little bit about yourself and the product and stuff like that?
00:04:34.959 --> 00:04:35.199
Sure.
00:04:35.360 --> 00:04:38.240
Yeah, so I'm Julian, I'm 42 years old.
00:04:38.399 --> 00:04:46.720
Um I started my career like 20, 25 years ago as an engineer, uh, starting in open source, actually.
00:04:46.800 --> 00:04:52.319
So that was my first gig was writing free software just for the fun of it.
00:04:52.399 --> 00:04:54.319
Uh I learned a lot doing that.
00:04:54.480 --> 00:05:01.759
Um, I grew from that to doing fun stuff in software engineering in like various companies, startups for a while.
00:05:01.920 --> 00:05:12.000
Uh, worked a bit at Red Hat and Datadog before that, and working on a large open source project a few years ago, uh, which is OpenStack for people knowing it.
00:05:12.079 --> 00:05:18.240
Uh, it's like huge uh cloud project, like a cloud provider project.
00:05:18.639 --> 00:05:32.800
Um, there was a lot of CI CD stuff that we were doing in that project, uh, with a lot of things I never met before, like Jenkins at the beginning and things like that, which was a lot of discovery for me, which was great.
00:05:33.040 --> 00:05:37.759
And I stumbled upon a
thing that we call merge queue, which I found very interesting.
00:05:37.920 --> 00:05:44.560
And there was nobody doing that at that time, or almost nobody, or you know, no very large usage of that tech.
00:05:44.879 --> 00:05:48.959
So with my co-founder Medi, we started to dig into that.
00:05:49.120 --> 00:06:01.759
He was also working on OpenStack, so that's where we met actually, and we started to build kind of own system, merge queue, and things going to another thing that started a company that was mergified.
00:06:02.079 --> 00:06:02.399
Okay.
00:06:03.040 --> 00:06:05.120
And around what what time was this?
00:06:05.199 --> 00:06:06.240
Like what year?
00:06:06.959 --> 00:06:11.279
I think we started like seven years ago now, so like 2019 something.
00:06:11.680 --> 00:06:12.879
Yeah, okay.
00:06:13.199 --> 00:06:19.360
Okay, and for anybody who's not familiar, what is a merge queue and what's the idea behind that?
00:06:20.480 --> 00:06:30.079
So the idea of merge queue, it's something a pain, which is when you have a lot of code to merge, so usually a lot of pull requests, for example, if you're using GitHub or merge requests or whatever.
00:06:30.480 --> 00:06:38.160
What people tend to, I mean, 15 years ago you had no CI or nothing, so you would just push in code into the main branch and that's it.
00:06:38.319 --> 00:06:50.480
But crossing your finger that the billboard or whatever it was named at the time was you know building the thing and testing it and saying, yes, it works, or no, it doesn't work, and you have to fix something, but it's too late because you merged it.
00:06:50.639 --> 00:06:56.319
So the idea of CIC was to do that before and that continuous integration thing, testing before you merge anything.
00:06:56.560 --> 00:07:12.720
But that did not solve all the problem that you have at a very large scale, which is when you have like two developers sending two changes that are different and that are being tested at the same time, they are maybe tested independently and work independently, but when they are merged together, they don't work.
00:07:12.879 --> 00:07:34.560
And if you test them, let's say on the same day on Monday, and then you merge them uh at the same time the day after, then maybe the main branch move, maybe they don't work together, maybe so you don't know, and you just merge something that was working on Monday but doesn't work maybe anymore on Tuesday to merge it and then you rerun the TI after the merge, which is what a lot of teams are still doing nowadays.
00:07:34.800 --> 00:07:38.399
Well, they don't have a merge queue and you read that and that doesn't work.
00:07:38.560 --> 00:07:47.839
You broke something and then you have to dig into what is broken, why, and stop maybe merging new code, deploying and dig into that, which is a pain.
00:07:48.000 --> 00:07:54.000
So the idea of the merge queue is to say if you have something that is seems ready to get merged, don't merge it.
00:07:54.160 --> 00:08:04.160
Put that into a queue and test every change one after the other and one with the other to make sure that they are both compatible, test with your CI again.
00:08:04.240 --> 00:08:08.240
And if that works, then you can merge it and then deploy if you want to go release, etc.
00:08:08.560 --> 00:08:16.959
So it's really making sure that you're not trying to click the merge button on 10p at the same time and then cross your finger that it's still going to work in the main branch.
00:08:17.680 --> 00:08:20.160
I'll give my perspective on it too.
00:08:20.480 --> 00:08:49.200
Um so we we the I I've encountered a problem at multiple places that I've worked, uh, the same problem you're describing, where somebody pushes their branch through CI, all the tests pass on their branch, and then somebody else does the same thing with their branch, all the tests pass there, both developers merge their work into the main branch, and then tests fail.
00:08:49.360 --> 00:08:51.120
And it's like, whoa, how is this possible?
00:08:51.279 --> 00:08:54.879
What everything passed on its own, but now things are failing on the main branch?
00:08:55.039 --> 00:08:56.639
Like, how how did this happen?
00:08:56.879 --> 00:09:08.480
Um and of course, the way that that happens is that two changes, even though they work independently, like you said, when they're brought together, they create some sort of bug.
00:09:09.519 --> 00:09:21.200
Um and the further downstream, of course, that bugs are allowed to travel, the more expensive they are to correct.
00:09:21.440 --> 00:09:27.440
You know, it's like the the cheapest time to find a bug is before you even write it.
00:09:27.600 --> 00:09:34.639
It if you if you figure out that what you're about to do is is problematic, just don't do that and do something different instead.
00:09:34.720 --> 00:09:38.960
So like the design stage is the cheapest time to catch it.
00:09:39.279 --> 00:10:00.080
Or then in your local development cycle, if you find a bug there, um and then it's more expensive if you if you uh have a test failure on your branch in CI or something like that, but then it's even more expensive if it makes its way to the main branch, because then you have then you're affecting a lot more than just yourself.
00:10:00.240 --> 00:10:06.559
And of course, if it makes it all the way to production, then that's a a significantly more expensive problem to fix.
00:10:07.679 --> 00:10:18.480
But the the more we can prevent these bugs from getting further downstream and causing greater expense and havoc and stuff like that, the better.
00:10:19.039 --> 00:10:30.000
So just to re-explain your what you explained in my own words, um we we want to prevent that.
00:10:30.320 --> 00:10:47.840
So instead of testing changes uh um uh in parallel and merging things uh uh one at a time, we test them serially.
00:10:48.720 --> 00:10:57.840
Um I've been using an analogy to explain it to the people I work with, because this is kind of
I think a tricky concept to understand.
00:10:57.919 --> 00:11:12.000
So my analogy is this there's a party happening, and some people at the party don't get along with each other, or some people who are invited to the party don't get along with each other.
00:11:12.240 --> 00:11:25.679
So let's say you and me, Julian, we're waiting in line to get in the party, and there's a bouncer at the party, and before I come in, the bouncer says to everybody at the party, they're like, Hey, Jason's about to come in.
00:11:25.840 --> 00:11:27.120
Is everybody cool with Jason?
00:11:27.279 --> 00:11:29.440
Okay, great, Jason, go in.
00:11:29.840 --> 00:11:32.240
And then actually you're in a separate line, Julian.
00:11:32.320 --> 00:11:37.360
You're you're in a different line, and the other bouncer there is like, hey guys, are you all cool with Julian?
00:11:37.440 --> 00:11:38.080
Can he come in?
00:11:38.240 --> 00:11:39.200
Okay, great.
00:11:39.440 --> 00:11:43.279
But you and me, we don't know that each other is gonna come into the party.
00:11:43.440 --> 00:11:45.120
You and me have problems.
00:11:45.360 --> 00:11:54.000
And so we we see each other and we're like, oh, bro, it's it's on, and then we fight and we ruin the whole party and it's terrible.
00:11:54.320 --> 00:12:18.879
So the idea is instead, instead of having these multiple separate lines to the party and letting people in without knowing about each other, we have just one single line, and people are coming in in groups, and maybe you and I are are placed in the same group, and we're like, hey, everybody at the party, are you cool with this group?
00:12:19.120 --> 00:12:23.039
And also everybody in this group, are you cool with each other?
00:12:23.279 --> 00:12:28.000
And so you and me see each other, and we're like, oh, this is this is no good.
00:12:28.799 --> 00:12:39.759
And so we never even go into the party, so the party can continue peacefully, and we can work it out in the parking lot or whatever, and then we can come back once we've resolved our differences.
00:12:40.159 --> 00:12:43.600
So that's that's a way that I that I explain it to people.
00:12:44.000 --> 00:12:46.000
Yeah, that's a that's a pretty good analogy.
00:12:46.399 --> 00:12:53.279
Uh, I think most people fail to understand that it's uh kind of a real problem because it's it can be pretty rare in small teams.
00:12:53.360 --> 00:13:01.440
Um the example I always give that is like pretty simple to understand in a lot of like programming language of you know modules or files that you import.
00:13:01.679 --> 00:13:14.240
Um you just have to imagine the the common scenario where you have one PR which is like I'm going to move that file somewhere else, and the other PR that is importing that file in the first place it was uh at.
00:13:14.480 --> 00:13:20.399
And once you merge both, there is no conflict in terms of code conflict because they don't overlap, they don't change the same files.
00:13:20.639 --> 00:13:24.639
One is just importing a file that is moving under its feet, and that's just going to break.
00:13:24.720 --> 00:13:44.399
And this kind of change that they don't happen if you have a team of five people pushing one PR a day, but when we have like 200 people pushing 200-300 PR every day, that happens a lot of time because you have people manufacturing, people moving things around, people changing the linters, people just adding new features that you know touch a large part of the scope, etc.
00:13:44.559 --> 00:13:44.799
etc.
00:13:45.120 --> 00:13:50.960
So the simple case that can grow into a lot of like larger cases as you grow.
00:13:52.000 --> 00:13:55.279
So I'm curious about the business side of Mergify.
00:13:55.759 --> 00:13:59.759
Um I think you said you started it around six years ago.
00:14:00.399 --> 00:14:06.320
I'm I'm always curious for people who start their own businesses, how you made it work.
00:14:06.480 --> 00:14:08.159
Because these things are not easy.
00:14:08.399 --> 00:14:10.399
Getting customers is really hard.
00:14:10.960 --> 00:14:13.279
Um can you tell me about that?
00:14:13.440 --> 00:14:18.720
How how did how did the beginning go and how did you get your first your first customers?
00:14:19.440 --> 00:14:29.679
The first users of Mergify were uh ourselves, basically, and the open source project we were working on at the time, uh, which was a pretty small base.
00:14:29.919 --> 00:14:35.519
And we were lucky enough to have a pretty good network of other open source developers working at Red Hat.
00:14:35.759 --> 00:14:38.960
It was like easiest to get those people from there.
00:14:39.039 --> 00:14:54.960
So we started to talk about the project, which was not really a company at the time, but more a side project for me and Medi because we were not thinking there was going to be uh kind of commercial attractiveness to that project.
00:14:55.120 --> 00:14:59.440
So we started to work on that, providing that service to open source projects.
00:14:59.759 --> 00:15:11.279
And after a few months, I kind of challenged uh Medi saying, okay, let's go to X Tribe, put a you know a credit card form on that and see if there's anything that could be interesting into solving that issue.
00:15:11.360 --> 00:15:12.960
Is that really an issue at a pain point?
00:15:13.600 --> 00:15:21.919
So I started that and started to go into online forum, you know, randomly and stating okay, we started to build that.
00:15:22.399 --> 00:15:24.000
Is that something that might be interesting?
00:15:24.080 --> 00:15:28.240
Is that the early days of the kind of indie maker, indie hacker movement?
00:15:28.320 --> 00:15:31.360
So I started to walk into that, trying to find people.
00:15:31.519 --> 00:15:40.080
So I found one guy interesting, I think it was a team of three or four people that used it for a month, and then said does not really solve anything for me, so that's just stop.
00:15:40.320 --> 00:15:45.840
But I was like, I
still was able to convince one guy to subscribe, so that might be something.
00:15:46.320 --> 00:15:48.000
And we continued to push that.
00:15:48.080 --> 00:16:01.440
And the thing is we were doing some other things at the beginning that GitHub was not doing, which is what solving small pain points like automatic merging of the PR, which is not beneficial for if you're using GitHub, but was not there at the time.
00:16:01.519 --> 00:16:15.679
So there was a lot of small pain points for GitHub we are solving, which brought some first startups, small companies, users like 10 people, 20 people team, which was all early kind of users.
00:16:15.759 --> 00:16:24.159
And from that, mostly the growth has been you know word of mouth, people talking to each other, people switching teams, people saying, Oh, we were using that before.
00:16:24.480 --> 00:16:36.480
Up to the point after I think a couple of years, where we started to have larger companies knocking at the door, asking for that because merge queue are a problem that are pretty old.
00:16:36.639 --> 00:16:39.360
Um, and you can find traces of the first merge queue.
00:16:39.519 --> 00:16:43.519
Like I read a blog post about that, like the origin story of merge queues.
00:16:43.840 --> 00:16:50.559
Um you find the earlier system like 20 years ago or more, which were not called merge queue, but the idea was kind of the same.
00:16:50.799 --> 00:16:57.919
And there was always you know a staff senior engineer in the team that was like, Well, this is a problem I want to solve.
00:16:58.320 --> 00:17:00.960
And the pain was maybe not that big at the time.
00:17:01.120 --> 00:17:12.000
It is bigger nowadays because you have a lot of larger teams, and the you know, the rise of the LLMs are pushing code to a new kind of border in terms of productivity.
00:17:12.079 --> 00:17:22.240
So you have this kind of problem we were solving six or seven years ago at a different scale now, and that's how we started to grow basically from team to team, and with pretty good stories.
00:17:22.400 --> 00:17:38.000
From we have one of our largest customers today, which is coming from one guy uh writing code in Scala on the weekend for fun and going into Fortune 500 companies on Monday to work and saying, Hey, by the way, I used that this weekend, it's pretty great.
00:17:38.079 --> 00:17:43.839
We should uh buy a license and deploy it uh here, and which we did like four or five years ago.
00:17:44.000 --> 00:17:48.160
So kind of crazy story coming from just open source networking and things like that.
00:17:50.880 --> 00:18:00.720
Um yeah, I'll I'll tell you a little bit about some of the challenges that I've encountered with my product, and I'm curious if you've had any of the same challenges.
00:18:01.359 --> 00:18:07.519
Um, first of all, what my product is is it's called Saturn CI.
00:18:07.599 --> 00:18:10.160
It's a CI platform, continuous integration.
00:18:10.640 --> 00:18:19.119
Um and I had been a user of Circle CI and GitHub Actions and some of these other products uh for a long time.
00:18:19.279 --> 00:18:22.319
I've been using Circle CI since like 2013, I think.
00:18:22.880 --> 00:18:27.279
Um and I've been frustrated with it basically the whole time.
00:18:27.519 --> 00:18:45.759
Um all these products, you know, uh CircleCI, GitHub Actions, GitLab Pipelines, Jenkins, um, and and probably others too, when uh when there's a failure in a test suite run, obviously you want to know what went wrong.
00:18:45.839 --> 00:18:48.640
You want to see the error so you can address it.
00:18:48.880 --> 00:18:52.400
But these products make you dig to find the error.
00:18:52.720 --> 00:19:02.319
And I've always thought that's so dumb because it's like you're a computer, you can figure out what the error is, and you can show it to me.
00:19:02.559 --> 00:19:13.200
The creators of the project of the product just decided that it's not a priority to build the product that way so that it parses out the error and gives it to you.
00:19:14.079 --> 00:19:21.039
And I had this idea maybe around 2017, 18, something like that.
00:19:21.200 --> 00:19:24.400
It was like not long before GitHub Actions came out.
00:19:24.880 --> 00:19:32.000
Um, and when I had the idea, I'm like, no, this is like too ambitious, this is this is too big of a bet.
00:19:32.079 --> 00:19:34.160
I don't want to bet more than I can afford to lose.
00:19:34.480 --> 00:19:36.480
I'm gonna try to put this out of my mind.
00:19:36.559 --> 00:19:58.559
And GitHub Actions came out, and it's like, okay, now I'm really not gonna pursue this because anything I could do, GitHub Actions with their vast resources, or GitHub with their vast resources, could just replicate anything I could do, which is funny because now I realize that that is a totally naive, frankly, way to look at it.
00:19:58.799 --> 00:20:04.400
Um, but GitHub Actions came out, I don't remember exactly when, like 2018 or something like that.
00:20:04.720 --> 00:20:05.359
Yeah.
00:20:05.759 --> 00:20:11.599
The years passed, and year after year, the product kept not improving.
00:20:12.319 --> 00:20:15.039
Um, and there's other weaknesses with it too.
00:20:15.119 --> 00:20:21.599
Like I always find the site architecture of GitHub Actions to be very disorienting.
00:20:21.839 --> 00:20:22.880
Circles CI too.
00:20:22.960 --> 00:20:25.200
It's like, hang on, where am I?
00:20:25.359 --> 00:20:29.039
Uh, how do I get from here to here and stuff like that?
00:20:29.279 --> 00:20:33.519
Just very disorienting and confusing.
00:20:33.839 --> 00:20:37.599
And so I'm like, all right, I think I can build something better.
00:20:37.839 --> 00:20:40.480
I really want to, but like, no, no, no, no, don't do that.
00:20:40.640 --> 00:20:41.920
Bad idea, no, no, no, no.
00:20:42.079 --> 00:20:45.759
And then like some time would pass, and I'd be like, man, I I hate these things.
00:20:45.920 --> 00:20:47.200
I really want to build my own thing.
00:20:47.279 --> 00:20:49.680
But no, no, no, no, no, bad idea, don't do it.
00:20:49.839 --> 00:20:52.240
And that happened about a million times.
00:20:53.039 --> 00:20:57.440
And then finally, one day in 2023, I'm just like, fuck it.
00:20:57.519 --> 00:20:58.319
I'm just gonna do it.
00:20:58.400 --> 00:21:02.720
Like, I don't even care if it works out as a business, I just need to get this out of my system.
00:21:02.880 --> 00:21:04.559
I'm gonna build something.
00:21:04.720 --> 00:21:09.839
And this is around the time Chat GPT, uh, I think it was four, came out.
00:21:10.000 --> 00:21:20.000
And you know, ChatGPT 3.5 to 4 was a big change, 4 was much more capable than 3, and so I used that to help me get started.
00:21:20.240 --> 00:21:27.519
Um, and honestly, I don't know if I could have done the whole thing without AI, that was a huge boost.
00:21:28.400 --> 00:21:42.720
Anyway, long story short, I was able to build it, and by now I've gotten paying customers, and I think it's going to work as a business, and so I'm putting my foot to the floor and and giving it everything I have.
00:21:42.960 --> 00:21:45.839
But here's here's the challenge that I wanted to bring up.
00:21:46.160 --> 00:21:53.519
Um I know what I don't like about the existing CI products, but not everybody is me.
00:21:53.680 --> 00:22:01.359
Different people have different pains with different products or the absence of pains in In products.
00:22:01.839 --> 00:22:09.039
And so they're not going to be bothered by the same things that bother me and be interested in the same things that interest me.
00:22:09.119 --> 00:22:16.319
And so my question to myself was who are my first few customers going to be and why are they going to be attracted to this product?
00:22:16.559 --> 00:22:27.680
I don't want to go out there and pitch a specific value proposition just because it resonates with me, because I don't know if that's going to resonate with other people.
00:22:27.839 --> 00:22:30.400
I have to like empirically find out.
00:22:30.559 --> 00:22:34.960
You know, let's get our first few customers and then find out why did these people come in the door?
00:22:35.119 --> 00:22:38.240
Okay, let's go get more of those people.
00:22:38.720 --> 00:22:43.119
I'm curious how much of that resonates with your experience.
00:22:44.480 --> 00:22:46.079
Uh it does quite a bit.
00:22:46.240 --> 00:22:55.839
I think or through is a little bit different because what you did is that you try try to and you're trying to attack something that exists, which is great.
00:22:56.000 --> 00:23:04.160
And for us, the early years were we were alone, so it's not like a blue ocean of there is nobody knows what a merge queue is for the first year.
00:23:04.240 --> 00:23:08.319
So every talk we have with engineers about what do you do, guys, at Mergi5?
00:23:08.480 --> 00:23:09.200
Like we do merge queue.
00:23:09.359 --> 00:23:10.160
What is a merge queue?
00:23:10.319 --> 00:23:24.799
Well, it does it,
and then we go back to the beginning of the podcast where we explain everything to everyone, which is not a way to sell any anything to any engineer when you have to explain the problem, but like, well, if you have to explain the problem, it's not a problem that I have, so I don't need a solution, which is very hard.
00:23:24.960 --> 00:23:42.000
So, lucky for us, um, MergeQ evolved to the point where GitHub uh started to provide one merge queue into their product in 2023, which uh brought us a bit more on the map, saying, Okay, that is a real kind of a category, you know, that's a real thing.
00:23:42.160 --> 00:23:44.559
There are multiple uh companies doing that now.
00:23:44.640 --> 00:23:47.519
We have competitors uh for a for a few years now.
00:23:47.680 --> 00:24:04.000
So that is to grow from something where it was a blue ocean to kind of a red ocean, not that many people to competitors doing it, but still where in I think in that why we started Multi5 because I know myself back from the years and Medine.
00:24:04.160 --> 00:24:16.720
I'm sure if you were trying to do that, and there was like 10 other companies doing it, we were like as you saying, okay, people are already doing that, there is no point that we do something better, so let's not do anything, and we probably don't know about that.
00:24:16.960 --> 00:24:26.319
Which I think is a pretty bad reason for not doing the thing, but as I think that's I call that my engineer mind, which is like very rational.
00:24:26.640 --> 00:24:36.000
Say, well, there are multiple alternatives, and there are a lot of products and projects, maybe open source projects doing that, so let's not start a new one because there's no point in it.
00:24:36.240 --> 00:24:57.599
But I think in your case, there is value in doing that because and you say that at some point, which is you have you are opinionated about what CI should be and should not be, which is uh kind of a luxury you can have, which I think GitHub cannot have, for example, because they are very large and they have to cover every user on the planet, basically.
00:24:57.680 --> 00:25:07.839
So if you look at GitHub Action from uh design engineering perspective, it sucks and it's great, it's great because you can do anything with it, and it sucks because you can do anything with it.
00:25:07.920 --> 00:25:18.240
So it's a very large framework where you can build a CI, but you can also build a tool that rearge PR with put labels on it, which is not CI, it's just workflow automation.
00:25:18.319 --> 00:25:22.079
You can do both and the same language and same YAML file, which makes no sense.
00:25:22.240 --> 00:25:36.559
But uh, and that's where everything is very fragile at the end of the day when you do GitHub action, for example, and they don't go to the point where they're going to let's say integrate and say this is the problem, this is the test that fails, and that's not the product.
00:25:36.720 --> 00:25:43.440
The product is a very large level brick, high-level brick where you can build stuff with it, and that's it.
00:25:43.599 --> 00:26:16.880
And that's I think always what GitHub tries to do because they have millions of users and you don't have them, so you can say, Okay, this is going to be my CI for this pain point, and start with a very like small niche or vertical or anything, and see this is what I'm going to build, and I'll have a few customers, and then you can see out expanding that if you have that works and that resonate with a few users and say, Okay, I'm going to address people using this technology, this working in this industry, whatever, which makes it easier when you start, which is very different than what large companies are going to do.
00:26:16.960 --> 00:26:24.079
So if you go that and you know you can start with building an audience, usually about like this is a pin point that I hate, I'm going to solve it.
00:26:24.240 --> 00:26:34.960
Like, I don't know, CI should know about test uh failure, and I don't know, should parse their log files by default and tell me what's wrong, etc., and be smart, then that's going to be your play.
00:26:35.119 --> 00:26:42.160
But it's not going to solve every use case of like GitHub action or Cycle CI or whatever, but you can start with that and iterate and go on.
00:26:42.400 --> 00:26:47.039
And that's that's a lot of that can fail if nobody cares about your problem.
00:26:47.119 --> 00:26:52.400
But if that works, that's pretty powerful to start and grow and then your product after that.
00:26:54.319 --> 00:27:04.720
Yeah, I I think of it a bit like a campfire where like if you just hold a lighter up to a log, it's never gonna work.
00:27:05.119 --> 00:27:17.440
Um you have to start with some tiny little twigs and then put some bigger twigs on top of that, and some small sticks, bigger sticks, and you can work your way up to the logs.
00:27:17.680 --> 00:27:26.079
Uh if I just put it out there and say, hey world, here's a head-on competitor to GitHub Actions, nobody's gonna notice.
00:27:26.160 --> 00:27:29.119
I'm not gonna get any traction anyhow.
00:27:29.440 --> 00:27:32.480
Um and I think there's multiple ways to do it.
00:27:32.640 --> 00:27:41.279
Like one way to do it is by starting with something very narrow and specific so that you're not really competing head-on with anybody else.
00:27:41.440 --> 00:27:54.640
You're doing something something so specific that you have no competitors, and so that makes it easier in that way, and then you can, you know, that's your like tiny twig, and then you can expand outwardly from there.
00:27:54.799 --> 00:27:56.400
That's what I did with this podcast, actually.
00:27:56.559 --> 00:28:11.440
It started as the Ruby testing podcast, and then I expanded to Rails, and I called it Rails with Jason, and then I expanded again, and now I just call it code with Jason, and it's just a general programming podcast.
00:28:11.680 --> 00:28:17.759
If I had started with just a
general programming podcast, maybe that would have gotten traction, maybe not.
00:28:17.920 --> 00:28:29.200
Um, but the way I did it, I felt like was it safer bet I was I was the world's best Ruby testing podcast, because it was the only Ruby testing podcast.
00:28:30.480 --> 00:28:42.319
Um But then a different way to go, which is the way that I'm going with Saturn CI, is my twigs take a different form, which is just people I know.
00:28:42.480 --> 00:29:00.240
So what I've been doing for the last while is just having as many calls as possible with people I know and just demoing the product and just saying, Hey, friend, look at this thing, and then that's all there is to it.
00:29:01.039 --> 00:29:05.599
And sometimes it's like, hey, friend, look at this thing, and they're like, oh, interesting.
00:29:05.839 --> 00:29:08.480
Like, hmm, I I never thought about this.
00:29:08.640 --> 00:29:10.000
Like, hmm, does it do that?
00:29:10.079 --> 00:29:10.960
Blah blah blah.
00:29:11.200 --> 00:29:15.279
And then we start talking, and it's like, hey, I I actually want to try this out.
00:29:15.359 --> 00:29:27.359
I got a project I want to use this on, and then you know, bit by bit, people like that accumulate, and that's how I've gotten my first few customers.
00:29:28.000 --> 00:29:37.279
It's it's not really the product, isn't that differentiated from Circle CI, GitHub Actions, etc., in terms of the sales pitch.
00:29:37.519 --> 00:29:43.359
It's just that the connection was made because there was a personal relationship there already.
00:29:43.759 --> 00:29:56.400
And then, you know, after I after I squeeze all the juice out of these personal relationships, because I only know so many people, then I can move on to whatever the next thing is.
00:29:56.480 --> 00:30:02.880
But luckily I know a lot of people, and so uh uh it'll probably take me a while to go through all those.
00:30:03.119 --> 00:30:13.920
I'm curious how much of it, how much of that was part of your story, just deliberately going to people you know and be like, hey, I'm working on this thing, take a look at it.
00:30:14.960 --> 00:30:16.240
Uh that's a good question.
00:30:16.480 --> 00:30:26.960
So I think it's not that that's I recommend usually because at least maybe that's uh I don't know, some kind of you know, difference in terms of like culture or whatever.
00:30:27.039 --> 00:30:31.920
But I think like if I go to my friends and I ask, like, do you think a merchant is a good idea, would be like that's a good idea.
00:30:32.079 --> 00:30:44.000
But if I was to, I mean, years ago, I was going to ask, uh, okay, so give me your credit card and I'll subscribe you to the things, I'll be like, Whoa, no, no, it's a good idea, but for you and maybe for people, but not for me.
00:30:44.240 --> 00:30:53.920
Because like if you if you do test your product generally to friends are going to be very nice to you and you know be asked a bit and say, Yeah, sure, that's a good idea.
00:30:54.000
--> 00:31:02.240
That seemed like very smart from you, and sure, I mean you'll find people to buy that for sure, which might not be the case because nobody might be having the the problem you think you're solving.
00:31:02.319 --> 00:31:29.440
So, or maybe it's not worth paying for solving it, like there's a lot of things like that where it's very hard, so it can be great, but uh, I'm always like you know, uh be careful when advising that to other like entrepreneurs and all, because it's it can be a trap if you don't have enough you know insight saying okay, they're saying it's nice, but I don't think they're really going to buy it, even if they don't give you money, just like you know, they're not going to use it or whatever.
00:31:29.599 --> 00:31:49.680
So um for that, but for the other part of networking for business, uh took me a few years to discover that, but I actually think that you're right in the sense that you'll get a lot of customers from your networks, not maybe your first level friends, but you know, friends and friends and acquaintance, etc.
00:31:50.400 --> 00:31:57.440
Because doing business, whatever you sell at the end of the day, is a lot of relationship and talking to people and and building trust.
00:31:57.519 --> 00:32:05.119
It's which again when you come as an engineer, you know, you do tech and you're like, I'm just going to put something online and sell it, and that's it.
00:32:05.359 --> 00:32:13.599
Which is the we are seeing a lot of engineers doing that, you know, the product let growth kind of move and just put a credit platform and that's it.
00:32:13.839 --> 00:32:23.839
Which I thought was going uh was going to sell mergify for the early years, but was my first move like let's connect Stripe because that is the problem to solve, right?
00:32:23.920 --> 00:32:44.400
To sell Mergify, I need to connect Stripe, which is definitely not the problem I had to solve, but that was my first reflex as an engineer doing that, which is more way more efficient to put a calendarly link actually in the on the website to talk to people, and I like 90% of the sell we do is like talking to people and teams saying, Okay, we're having this issue, can you help us?
00:32:44.559 --> 00:32:48.240
And then we just people come to your website and then they schedule a call.
00:32:48.559 --> 00:32:48.960
Exactly.
00:32:49.119 --> 00:32:57.839
That's like 90% of the cells are like that, which was uh no, I think it's normal, but like a few years ago, I was like, This is weird, like people want to talk to me.
00:32:58.000 --> 00:32:59.599
Why do you want to talk to other people?
00:32:59.680 --> 00:33:03.920
Like it's no need to, you can just click and subscribe and build the docs online, and that's it.
00:33:04.079 --> 00:33:08.160
But you have to build trust, you have to to to answer questions to people.
00:33:08.319 --> 00:33:14.319
They want they're not experts in their their area, so they want to know if you can solve their problem.
00:33:14.559 --> 00:33:23.519
Sometimes they don't really understand fully what you're building as a product, they just have a vague idea that maybe this is something you can help us solving.
00:33:23.599 --> 00:33:36.160
So sometimes yeah, I have like once I have a I had a conversation with a prospect which was about database and SQL, and there was no way we're going to help them because we don't do database at all.
00:33:36.319 --> 00:33:41.039
But they stumbled upon a blog post that we wrote about Postgres real migration, etc.
00:33:41.279 --> 00:33:48.799
That we did internally, just to share some knowledge, and they were like, Oh, these guys might know about databases, and they just called and and I was like, I'm going to help you.
00:33:48.880 --> 00:33:55.359
That's going to be 30 minutes of free, you know, engineering consultation for you guys, but we're not going to sell you anything because we don't have anything to sell.
00:33:55.440 --> 00:33:58.960
I mean, we're not going to spend time helping you, that's not what we do.
00:33:59.119 --> 00:34:12.400
But that's that's kind of the thing that you that I learned the kind of the hard way is that yeah, you have to build trust and you have to talk to people, which is where if you have a large network of people, that's actually a pretty good leverage to to fast track.
00:34:12.480 --> 00:34:18.800
Where I did that in the early years with my open source networks, which is probably not as large as yours.
00:34:18.880 --> 00:34:42.400
So after a few months, you're like, I talk to everyone I know about what I'm doing, and they talk to everyone they know more or less about what I'm doing, and then we are at that point that we need to get people I don't know, and then you start doing like marketing online and whatever that can be, uh, which is not way um harder because you talk to random strangers that don't know you from the other side of the planet sometimes.
00:34:42.480 --> 00:34:44.880
So we're like, okay, let's maybe do business together.
00:34:44.960 --> 00:34:48.079
So that's a lot of work, yeah.
00:34:48.800 --> 00:34:54.559
Yeah, especially with a new product, it takes a lot of trust.
00:34:54.800 --> 00:35:02.719
Um, and sometimes your customers don't know how new or old your product is, and so then it doesn't come into the picture so much.
00:35:02.960 --> 00:35:13.199
But like a lot of times I think it's fairly obvious if a product is pretty new and not very mature yet, and so people have to take a bit of a gamble.
00:35:13.519 --> 00:35:21.280
It's like if I use this product, is it still gonna be around in a couple years or even six months?
00:35:21.440 --> 00:35:24.079
Uh how well maintained is this gonna be?
00:35:24.239 --> 00:35:32.559
Like, that's a big risk, and so that's of course where a pre-existing personal relationship helps a lot.
00:35:33.039 --> 00:35:45.920
Um most of the people who have expressed interest in using Saturn CI for real um are people I've I've known for
years and maybe even worked with in the past.
00:35:46.559 --> 00:35:54.400
Um your like uh thing to schedule a call on your website, that seems like a good idea.
00:35:54.719 --> 00:35:58.719
Um it's you know, it's not the first time in my life I've like heard of that idea.
00:35:58.880 --> 00:36:05.840
I'm I obviously know that people have that on their websites, but for some reason it never occurred to me for myself to do that.
00:36:05.920 --> 00:36:06.960
So I made a note just now.
00:36:07.039 --> 00:36:10.400
I'm like, oh yeah, maybe I should put that on my website.
00:36:10.480 --> 00:36:14.559
That's a low-cost, high potential upside kind of a thing.
00:36:15.840 --> 00:36:24.559
Um did you did you have any of those challenges of people being like, hmm, I don't know, like this is this is a new product.
00:36:24.719 --> 00:36:25.920
I I really don't know.
00:36:26.079 --> 00:36:27.920
Did did you have any of that at all?
00:36:29.199 --> 00:36:31.199
Not that much, to be honest.
00:36:31.440 --> 00:36:41.920
Um and I think I mean I don't know why, maybe because we were lucky, because we were good at selling our vision.
00:36:42.000 --> 00:36:49.679
I think part of that is like we we s try to sell each time what we are doing, how long we've been doing for what kind of customers, all of that.
00:36:49.760 --> 00:36:58.159
So, and we can share you know case studies and we can put you know uh existing customers together, like connect them with like prospect if needed.
00:36:58.239 --> 00:37:01.119
We we did that sometimes if that helps.
00:37:01.360 --> 00:37:06.960
So, you know, you build trust basically, so anything we can help and use to build trust helps usually.
00:37:07.199 --> 00:37:08.159
So it's pretty rare.
00:37:08.239 --> 00:37:16.800
I think also because a lot of our customers are startups, which are you know, there's a lot of risk going on when you build a startup.
00:37:16.880 --> 00:37:21.840
So like that's we're not at that uh using a service that might disappear next week.
00:37:21.920 --> 00:37:23.280
Okay, that's not a big deal anyway.
00:37:23.360 --> 00:37:50.000
We have a lot, a lot of over risks, and for larger companies, it's it takes a lot of paperwork usually to prove that you're okay, that you're going to exist in a year, that you're going to you know uh be trustful in some way, but we never had a large blowback or you know, people closing the door saying okay, it's not going to work because you're too small or too young.
00:37:50.239 --> 00:38:04.400
It happened once, I think, that some a company uh picked a competitor because they seemed to uh I think to raise more money than we did, and we're like, Okay, this one is going to live longer than you, and so we're going to pick this one.
00:38:04.480 --> 00:38:07.039
And that happened once in six years or so.
00:38:07.199 --> 00:38:10.079
And that was kind of funny, it was like, what's a weird decision?
00:38:10.159 --> 00:38:14.159
But I think they need a reason to
pick one or other another, so they pick that one.
00:38:14.320 --> 00:38:22.320
But that's pretty rare and almost fun in some way, but other than that, luckily enough, it was not a huge uh problem for us.
00:38:23.360 --> 00:38:26.079
That reminds me, I was talking to a friend of mine the other day.
00:38:26.239 --> 00:38:34.239
I was kind of trying to sell them on Saturn CI, and he's like, No, I'm afraid it wouldn't make it through our procurement process.
00:38:34.320 --> 00:38:43.519
Uh on the one hand, you have Circle CI, a multi-billion dollar company, and then on the other hand, you have a one-man side hustle.
00:38:43.679 --> 00:38:45.039
Like, that's not gonna fly.
00:38:45.119 --> 00:38:46.800
And I'm like, Yeah, fair.
00:38:48.960 --> 00:38:53.679
Um, I I recently learned about this concept.
00:38:53.920 --> 00:39:02.639
Um it it wasn't really news to me, but I hadn't heard it framed this exact way: the concept of a buying window.
00:39:03.440 --> 00:39:21.760
So every customer when they're gonna buy something, they go through their journey in life, and there's a certain time when they're not ready to buy yet, and then a time when they are ready to buy, and then they've made their purchase, and they're they're not gonna buy after that because they already have.
00:39:21.920 --> 00:39:24.320
So there's that buying window in between.
00:39:24.480 --> 00:39:31.840
And the vast majority of people you come into contact with, they're not gonna be in the buying window at the time that you meet them.
00:39:32.079 --> 00:39:37.599
And I think it's un it's helpful to understand what this buying window looks like.
00:39:38.320 --> 00:39:51.360
Um and what is the trigger that causes people to have a need or a desire to either adopt a product or switch from one product to another one.
00:39:51.679 --> 00:40:03.199
That's part of my challenges is if somebody's gonna use my CI product, it's either because they don't have one or they do have one and they're gonna switch to something else.
00:40:03.440 --> 00:40:10.320
Um, and there's a cost to that, obviously, of of switching from one thing to another thing.
00:40:10.639 --> 00:40:19.760
Um, and so I've been asking myself the question of like, okay, what causes people to switch from one CI product to another one?
00:40:20.079 --> 00:40:21.039
Is it cost?
00:40:21.199 --> 00:40:22.239
Is it something else?
00:40:22.400 --> 00:40:24.079
What is the story there?
00:40:24.480 --> 00:40:34.480
And I'm curious, in your case, do you have a feel for what that buying story looks like and what those triggers are and stuff like that?
00:40:35.039 --> 00:40:35.599
Right.
00:40:35.840 --> 00:40:53.679
I mean, for the window, you're totally right, and that's why you see a lot of companies uh doing you know kind of a brand awareness kind of things in terms of marketing, which is extremely costly, but you can't really do at a small size, even for us, it's pretty rare to do things that we just say, oh, that's going to
be brand awareness.
00:40:53.920 --> 00:41:04.079
But it does work, like uh people engineers they keep seeing your name and your brand, and at some point they're getting familiar with the name, and they're like, What do these guys do?
00:41:04.159 --> 00:41:09.840
By the way, okay, they do that, so maybe at some point I'll have this problem, and then they're reminded about oh yeah, okay.
00:41:09.920 --> 00:41:21.599
I remember that I met this at some point, and that was this kind of issue, I should look into it, and then you start to enter the window, which means they're going to enter that when they feel the pain that you're going to solve.
00:41:21.840 --> 00:41:33.840
In CI case, even if they don't have any CI and they're looking for a new one, but that's might be pretty rare, and there's little chance they're going to pick you because you're new, nobody knows you, and you have probably not a lot of users, all of that.
00:41:33.920 --> 00:41:38.400
So nobody's going is going to be fired for buying IBM, that's a whole thing, right?
00:41:38.480 --> 00:41:48.159
So they're going to buy cycle CI and not buy you or get action because that's just that's there, and everyone using it, so I'm not going to get fired because that's a risk I'm not going to take.
00:41:48.639 --> 00:42:04.880
So that's where it's going to be very hard for this kind of you know, um, new customers, new users to to buy your product, but people using already um CI, they're going to fill up pain, which at some point is going to be so costly that they're going to look for alternatives.
00:42:05.039 --> 00:42:28.400
If I take for merch queue, for example, either they go from no merch queue to a merch queue and they'll look at the market and there are a few competitors and see okay, who has the best feature, what's the best pricing, who is going to provide me the best aero I because it always boils down, especially as you start to build your case to hire like bigger companies, it's always boils on to what is my return on investment.
00:42:28.639 --> 00:42:31.360
So, how much money I'm going to save if I buy your product.
00:42:31.599 --> 00:42:39.360
That might not be true for tech products in smaller companies and startups because that won't ask for hard case about aero.
00:42:39.679 --> 00:42:47.119
They will be like, okay, that might be just easier, they won't ask for numbers or you know, a multiple of returns on their investments.
00:42:47.280 --> 00:42:51.440
Just going to say, okay, this is going to be the new tech, there's going to be to be something we can build around.
00:42:51.519 --> 00:42:58.800
This is going to be a different kind of thoughts to or mindset to buy a product like that.
00:42:58.960 --> 00:43:05.840
But for larger companies, it's going to build on okay, if I use your CI, um, how much time is it going to take me to switch?
00:43:05.920 --> 00:43:18.960
Uh, how much money is going to cost me, but how much money I'm going to save, or hours of human time going to save at the end of the month of the day, and where is going to be that win basically from using whatever I'm using right now?
00:43:19.119 --> 00:43:20.400
So you can see I could have action.
00:43:20.559 --> 00:43:41.679
If you're able to build a case like that, and sometimes it's very hard because the pain might be obvious, and the worst case is like where you're going to save, I don't know, 10 hours per engineering per month, which is really usually ballpark number of hours, and everyone knows that it's like, yeah, sure, why 10 or 2 or 12 or whatever?
00:43:41.760 --> 00:43:49.119
It's always random numbers because not totally random, but it's like a bit random because every team is different, every case is different.
00:43:49.199 --> 00:43:59.599
So it's very hard to find how much hours people are spending on on let's say debugging a log file from a CI, for example, and then they won't have to do that.
00:44:00.159 --> 00:44:09.840
CI, so if you are able to count that and say okay, you won't have to do that anymore because it's going to be automatic and you're going to win on that, or win on I don't know, compute time, etc.
00:44:10.079 --> 00:44:13.679
If you can do that and build a case like that, it's pretty easy to sell.
00:44:13.840 --> 00:44:23.599
Um, I think like the larger the customer is, the more you can build a error ICAs case, the easier it's going to be uh to sell your product at the end of the day.
00:44:24.559 --> 00:44:41.199
Yeah, and in your case, I'm sure you've already thought about this plenty, but like there's also the cost in frustration and morale and stuff like that when the master branch is constantly going red.
00:44:42.000 --> 00:44:44.880
There's a feeling of like, oh, this again.
00:44:45.039 --> 00:44:48.559
Here we go again, and you have to wait for it to get fixed.
00:44:48.960 --> 00:44:51.119
It's like, is anybody looking at this right now?
00:44:51.360 --> 00:44:58.719
Okay, I'll try to keep an eye on the status and see when things are green again so I can continue to focus on my work.
00:44:58.960 --> 00:45:03.840
And there's just this feeling when stuff like that happens, it's just like uh we suck.
00:45:04.079 --> 00:45:06.719
You know, our processes suck.
00:45:07.039 --> 00:45:08.079
This sucks.
00:45:08.320 --> 00:45:09.280
And that's no fun.
00:45:09.360 --> 00:45:16.559
Nobody wants to be part of an organization that has broken processes and systems and stuff like that.
00:45:17.039 --> 00:45:35.760
And so if you can make those things no longer be broken, there's not just the like hard, quantifiable uh cost savings in terms of hours and money and stuff like that, but also that emotional penalty that that goes away.
00:45:37.280 --> 00:46:06.400
Yes, so I've been enough in that field to see how things change, and for example, the whole um discourse around developer experience was a lot trendy, like three, four years ago, and every team was like, Okay, we need to provide the best DX to every
team, so we want the engineers to be very happy to go to work, so we don't want to have a broken CI, for example, or cold coffee or whatever, which I think tend to switch a bit.
00:46:06.559 --> 00:46:14.239
For example, for the last couple of years, I don't tend to see a lot of content around that, and like developer experience is still a thing, it does still exist.
00:46:14.559 --> 00:46:27.920
I feel like companies are less prone to spend money on just improving the day-to-day work of their engineers, or like maybe it's not as important as it was, or maybe it just became mainstream.
00:46:28.000 --> 00:46:34.320
I don't know, but I see a lot less marketing, let's say, beaten around that ID, which I think is still true, is still very important.
00:46:34.400 --> 00:46:43.599
I mean, I don't want to work in a team where the CI is broken like twice a day, it just sucks, and there's no way to that's that's not a good um spending of my time trying to fix the CI all the time.
00:46:43.679 --> 00:47:10.159
So, but I I think it's not a good argument anymore, and maybe because that you know, the feeling of 2026 where you got a lot of team where a lot of layoff, a lot of you know, the dynamic seems to be very different in terms of engineering, so it's less about you know making the day-to-day of engineers really great, it's more about efficiency and where is the money going um to be spent.
00:47:10.320 --> 00:47:32.320
If I compare right now to five years ago, like before the first you know, kind of crash of like 2022, it was very easy to sell SaaS to any companies, like we were having call every week, uh, just to any team that they wanted a new SaaS, they were looking at like Mergi 5, we were just selling very easily.
00:47:32.559 --> 00:47:42.320
The year after that, like 2022, 2023, it was like every team we were talking to, they were like, There is no budget anymore, there was layoff, no budget anymore.
00:47:42.400 --> 00:47:46.079
So, like improving the day-to-day of no, that's not a good reason.
00:47:46.239 --> 00:47:56.960
So, we were we that's where we started to say, okay, there's actually an ROI that we can say on let's say CI time, for example, because we're going to do batches on pull requests and stop breaking the mailing.
00:47:57.119 --> 00:48:20.719
So we had to switch from just selling something that is you know kind of fun technically and great and improve your daily life to okay, we're going to save you a lot of CI time and engineering interruption about issues, all of that, and you have to put numbers and dollars in front of that, and that gets easier to actually sell your solution if you do that because that works whatever the weather is in business, basically.
00:48:22.719 --> 00:48:33.760
Yeah, do you think that that was you know, things got harder, it sounds like um around 22, 23, whatever.
00:48:34.159 --> 00:48:37.440
Do you think that was a positive thing?
00:48:37.679 --> 00:48:52.639
You know, like maybe in the moment when
things got harder, that obviously made things harder, but like it forced you to change some things and maybe put you ultimately in a stronger position.
00:48:52.800 --> 00:48:54.480
Do you think that's the case?
00:48:55.119 --> 00:48:56.800
I think that's the case, yeah.
00:48:56.880 --> 00:49:22.320
It was very easy for us the first years, basically because the like business atmosphere was great, there was no competitors, so it was the only solution on the market, so you have no efforts to do, so you're like in a comfort zone where you don't have to do much when when it switched a few years after when you started to see competitors, um, you know, turning business.
00:49:22.400 --> 00:49:24.239
It's like, well, it's a bit different now.
00:49:24.320 --> 00:49:30.800
We're not going to buy uh new SaaS, so we're going to use GitHub and stay on GitHub, maybe, and and not use anything else.
00:49:30.960 --> 00:49:42.239
So let's remove MergiFi and other ticketing system we might be using, and you know, consolidate everything on a few SaaS that we use, which was a trend like three years or four years ago.
00:49:42.639 --> 00:49:53.920
Um, you have to think twice about what you're doing, either you're going to be dead very soon, or you have to prove your value and to make sure that you have a pretty good value for your cost and that you have a pretty good error.
00:49:54.159 --> 00:50:21.039
So that was a lot of getting out of our comfort zone and saying, okay, we need to actually uh bring the product on a next level technically and in business too, saying okay, this actually has a real value and you're not paying for nothing, and and you learn that way, which I think we did, otherwise we would be dead by now because everybody would probably have left at some point and we would have been no new business, etc.
00:50:21.360 --> 00:50:23.199
Which was not the case, luckily.
00:50:23.360 --> 00:50:31.119
But yeah, that's I think if you want everything changes every few years because of like economic crisis, you know, you name it.
00:50:31.199 --> 00:50:42.800
Uh, and so you have to always be learning what's going on and always getting out of your comfort zone saying, okay, this is something new, business that changed, so either you adapt, either you die, basically.
00:50:43.840 --> 00:50:45.760
Yeah, exactly.
00:50:46.159 --> 00:50:56.079
Um, and I used to kind of envy people who had easy successes, they just start something and it takes off almost immediately.
00:50:56.239 --> 00:51:09.920
Um, but I don't envy those people anymore because they they don't necessarily have to learn much when that happens, and maybe they don't know why they don't know what explains their own success.
00:51:10.480 --> 00:51:18.159
And so just the smallest change could wipe them out because they were never forced to build certain skills.
00:51:18.559 --> 00:51:25.440
With me, you know, I've been trying to build a business like the one that I'm building now since 2008.
00:51:25.599 --> 00:51:29.199
So as we record this, that it's been 18 years.
00:51:29.679 --> 00:51:48.559
Um and so I've had to really fight and learn a lot, and by the time I finally am successful with it, like it'll take a lot to wipe me off the map because I've had to gain such a deep understanding in order for things to work.
00:51:48.880 --> 00:51:50.800
Um, and I think that's better.
00:51:51.039 --> 00:51:56.800
Um okay, it's it's probably time for us to start heading toward a conclusion soon.
00:51:57.039 --> 00:51:59.920
Um one or two last questions.
00:52:00.079 --> 00:52:07.599
Is there anything that we haven't covered with uh related to Mergify or anything else that you wanted to touch on?
00:52:08.559 --> 00:52:10.320
Uh what can I say more?
00:52:10.480 --> 00:52:23.119
Um well uh kind of a bit about how things changed this year, I think, with the ride of you know LLMAI and all of that, it changed a lot the way people work in engineering for sure.
00:52:23.199 --> 00:52:31.840
And so we're seeing that trend where people send ten times the code they were doing last year.
00:52:32.000 --> 00:52:51.119
So, for example, in our case a lot of team are now looking into merge queues because of that, because the pace of changes is going crazy, um and they need more stability into what they ship and overship it, which merge queue is a great tool for.
00:52:51.360 --> 00:52:59.599
Um, but probably not the end of the story because now the new button leg is going to be CI con reviews, so that's a lot of things we are looking into right now to improve.
00:52:59.679 --> 00:53:06.400
I mean, like you were saying, there's a lot of things to do in CI, I think, where that tech has been is very old.
00:53:06.559 --> 00:53:12.079
If you were talking about Circle CI from like 12 years ago, 13 years ago, it's pretty old now.
00:53:12.239 --> 00:53:16.639
Um, GitHub Action is not that old, but still and I anyway.
00:53:16.719 --> 00:53:19.840
We have a lot of customers using Jenkins, which is very old, also.
00:53:19.920 --> 00:53:25.760
So that tech did not change for the for a lot of time over the last years, and I think there's a lot of things to do in that area.
00:53:25.840 --> 00:53:36.159
So, I mean, I'm looking into that because I'm pretty excited to see what's going to happen in the next years about that whole development workflow, which I think is going to change a lot.
00:53:36.320 --> 00:53:40.320
Uh, we're not going to stay using pull requests on GitHub forever.
00:53:40.400 --> 00:53:41.920
I think that's going to change at some point.
00:53:42.079 --> 00:53:43.199
That can't that can't be it.
00:53:43.280 --> 00:53:44.880
That can be the end game, right?
00:53:45.039 --> 00:53:46.960
So that has to change at some point.
00:53:47.840 --> 00:53:50.000
Yeah, I hope you're right about that.
00:53:50.320 --> 00:53:53.440
Um, final question for you, Julian.
00:53:53.679 --> 00:54:01.039
Um is there any any links that you want to share where people can learn more about you and mergify?
00:54:01.840 --> 00:54:02.320
Sure.
00:54:02.400 --> 00:54:12.000
Um, I have a kind of a blog/slash newsletter that I write every week or so on uh julian.donjou.info, and I will put the link somewhere, I guess.
00:54:12.159 --> 00:54:20.159
Uh so if you're curious about what I think, what I write about, um, I've been writing for like 20 years or so on on my blog.
00:54:20.400 --> 00:54:28.079
So a lot of archive if you're curious, but I try to write about I mean CI of check or whatever crossed my mind every week.
00:54:28.159 --> 00:54:32.880
So feel free to subscribe if you want to read more about my ramblings.
00:54:33.760 --> 00:54:34.320
All right.
00:54:34.480 --> 00:54:36.239
We'll put that stuff in the show notes.
00:54:36.320 --> 00:54:38.480
And Julian, thanks so much for coming on the show.
00:54:39.039 --> 00:54:39.840
Thanks, Jason.