What's it really like to work at Transistor?

Hey everyone. Welcome to build your SaaS this is the behind the scenes story of building a web app in 2022. I'm Jon Buda, a software engineer.

I'm Justin Jackson, and I do marketing.

I'm Helen, I do customer success.

I'm, Jason, also a software engineer.

We got the whole band together. Jon, this is the, this is the most people we've had in our recording, uh, closet.

It is, we'll see how this works. It's pretty, it's pretty tight in here.

Well, it's good to have. I mean, the whole reason we started this show is because we knew if we were going to be working on transistor, we needed to. Have a podcast and feel what it was like the podcast. Uh, did Jason, is this your first podcast ever absolute first podcast ever absolute first, Helen. I know you've been on, you've been on quite a few podcasts actually as a guest a few times. Yeah. I can think of at least three or four times. I've heard you as a guest.

true. Yeah. And probably about five, five or six.

Yeah, perfect.

Seasoned veteran,

a seasoned veteran. Well, I think so just as a background, uh, Helen started working with us full time, April 27th, 2020. And Jason started working with us. Full-time August 3rd, 2021. So Helen's one year anniversary is coming up and then in the summer it'll be one year for Jason as well. That feels like that long Helen.

It does. Yeah, I guess, um, with me kind of starting in kind of like a temporary kind of, um, capacity in 2019, it feels a lot longer because there's kind of. No differentiation between, um, part-time and full-time, it's just, it's just all transistor.

Yeah, that was, that was the, the nice thing is that you were helping us out part-time and then able to move into a full-time role. How about you, Jason? How does it feel like you've been working for us forever or does it feel brand new? It kind of feels a little bit

like. The time sort of has a weird amount of distortion over the last couple of years. So

yeah, it feels like both for me with Jason, just because I know James for so long. Yeah. I've worked with him before, so it's like, kind of feels like a long time, but also doesn't.

Yeah. I mean, sometimes it feels like we've always because we do these weekly meetings. We just had one. Sometimes it feels like we've been doing those forever. And then, uh, other times it feels like, wow, we're still just a new team trying to figure out how to work with each other and how to, you know, make, make this work as a, as teammates. You know, I thought it would be just to ease in to us all for being on the show together. We were just in this meeting and this idea of the.

A lot of things as we get really great feedback about the integration, um, it's kind of something that perhaps. Some translates to customers. Don't even know that's there. It may be a little bit tucked away, but I guess the people that do find it, um, take the time to let us know how much they enjoy using it and how helpful it's been. Uh, just to kind of like syndicate, um, the podcast across to YouTube and just to correct.

Yeah. Now this is a deep cut for build your SAS listeners because this YouTube integration played, it played a key role in early episodes. John did well in the sense that we talked about it a lot. Uh, in this, uh, because it was a pain initially. So I, and maybe you don't even remember. Do you remember, like there was like issues about it. It caused a lot of customer support initially, and then. What exactly happened like for awhile, Google implemented a new,

no, I am. Yeah, it was, it was painful. I mean, it worked for awhile. There was, we had to get, we had to go through this verification process to, to increase our number of accounts that could connect through our API integration with. I think at one point we were a, at like only a hundred people could actually author accounts to transistor through YouTube.

Yeah. So now. This is that's the background. And this is what's interesting about bringing on new people is that I think it's pretty easy, especially when you've been working on something from the beginning to just get a cranky about things like this. Just be like, I mean, I've had conversations with people where I'm like, we're just going to tear this thing out. Like it's just a pain.

don't know if Jason's really touched that at all

though. Have you touched that yet? Jason?

I haven't. I've looked at it a bit only because it's sort of loosely related to some of the audio processing stuff. Um, but I've w I've

withheld jumping in, does it process like the processing stuff? Is it, uh, is that still a problem? Does it process like having to process all of that video and stuff? Is that still a problem? Is it. Okay. No, it's,

I mean, it's slower than audio, but it's not, it's not terrible. It's, you're basically stitching together an audio file and an image, an image that's on a loop of like one or two frames per second. And you stitch it together with the same command line program we use for our audio processing.

Okay. And Helen, have we, cause I I'm actually a little bit disconnected from this now is do we get a lot of bugs of complaints about the YouTube. Do you have that sense of like, this is maybe the sense we used to have was it was too costly to support this thing. How do you feel about that? Yeah, I guess

that's one consideration when you kind of build an integration, you always end up doing a little bit of support for whatever company you're integrating with, whether that's YouTube, um, you know, apple, Spotify. Um,

just so just as an aside, what if we were, if we're going to talk badly about, uh, other companies, who do we do the most support for. Out of all the companies. Do you think what's, what's our top three or four, do you think I'm

amazing. Neutral,

like Switzerland. I mean, apple and Spotify have to be up there. It feels like we do a lot of support for apple, Spotify still, but I don't see YouTube as much.

Where we were answering. There are lots of questions about why the tube intubation isn't, wasn't working, um, and kind of trying to set people's expectations based upon, or perhaps not having had an update from, uh, or kind of clear communication. I think we are fortunate that a lot, lots of our partners do kind of communicate well with us though. So, um, we can kind of pass on that information. It was more a case of like educating people on how.

Yeah. But, but it has YouTube been coming up as much. Is it, is it like in terms of it not working in terms of there being customer support or is it more just people are interested in it now?

No, I think it's, um, since the, um, kind of off issue was fixed, there's very little support need really. Um, but the trouble is if something does go wrong it's then how do we go about. Communicating with Google to fix those issues. So if we did expand upon the integration, um, you know, how, how many more kind of features, what kind of speech set can we support, um, without kind of having better communication with

the YouTube side of things? That's actually an interesting way. I haven't even, I've never thought about it like that, where. You, one of the questions you're asking right now, Helen is like, well, we could make this better and promote it more. But if something goes wrong, who would we talk to? Like, do we have any channel to YouTube right now? Because now we have it with apple and Spotify, but

I don't think we do. I don't think we ever have. It's like almost no support for the API stuff beyond just a bunch of times. Yeah, unless probably unless you're a huge company and have a direct contact, but

yeah, that's th that would definitely be an issue on the flip side. Like I just had, I think I posted this. Slack, but Jonathan Stark, who does the ditching hourly podcast? He just like impromptu raved about that YouTube integration. Like you said, I just found this and he's like, it just feels like magic. Like suddenly my show is on YouTube and in parallel right now the podcast industry is talking a lot about YouTube.

address. Yeah,

yeah, yeah. That's a good point. I should get that. Uh, I'm actually going to be, I'm not going to that version of the conference, but they're the same time I'm in LA for family vacation is the conference. So maybe I should try to sneak out and go to one of the parties and. Get that guy's email. What do you think about all this, Jason? What are some of your thoughts about whether we should improve this feature? Promote it more, or should we just hide it? Keep it in the closet? Yeah, it's

interesting. The, I mean, we've gotten, we've gotten a few people that have pinged us about doing. Like video podcasts. So it's like this, this is sort of, you know, copying audio with some static image to YouTube, which to me, like you said, I don't know who was sitting there going to YouTube to listen to something like, I don't, I don't know who that person is. Apparently they, they exist.

I don't know how you and John think about this cause you actually have to build. But when you're hearing customers and remembering being a customer yourself, like when I was a podcaster, it's just like anything that happened. Automagically. I published something. And then a series of auto, automatic magical events happen. Like I publish an episode, it automatically tweets it for me. It automatically publishes a little video on YouTube for me.

so people are already going to the trouble of adding individual episode artwork. Um, and there was kind of an expectation that, that aren't work with that. Maybe a unique custom thumbnail for the video. Um, whereas currently we just use one static image, which is both the thumbnail and the background image for the entire video. So at the moment, any podcast episodes that get uploaded all have that kind of same image.

up on YouTube too. And Jason, you see how this. Inner laces with the stuff we're doing with podcasts websites, which we can talk about in awhile. But if we have a, an object called host and an object called guest, and both of them have a image attached to them, it's theoretical. We could generate those cool images automatically that have like the host image and the guest image and the name of the podcast. Right? Yeah.

That'd be cool for sure.

Yeah, there's definitely a lot of stuff we could do to improve it. Um, I dunno where it's, where it's at on our roadmap. I mean, there's a lot of custom images and backgrounds, animated backgrounds, and right now, like, you know, we we've had the YouTube integration around for awhile and then we also added dynamic audio. So you can add ads and stuff like that to your audio, but that stuff doesn't get sync to YouTube.

Yeah. Yeah.

And I honestly forget what we, what the problem was.

I don't know that we ever, like, it was some kind of weird timing issue. Like I don't know that we ever completely tracked down what the issue

was. It was, I think we did figure it out and it was really. I don't, I don't remember.

Maybe I just expunged it from my brain. Cause I didn't

want to know. I mean, that's the other point is. Uh, underneath every magical feature is a developer that might be stressed out about it. That it's all going to break

for. The magic is actually someone who's manually running a bunch of things every hour. Yeah. Yeah. It was not the case for us, but

it's like, people love that hot water when they turn on the taps, but the plumber knows what's really holding everything together, down in the base.

Right. There's some guy with a blowtorch on a pipe, eating the water.

I'm somebody that gets up every morning and hits the pipe without wrench. Get to go. And again, yeah. Well, I mean, I think the other thought I had when you were saying that John was sometimes in product development, like you can have ideas, like we've had this, this loosely held philosophy called wait and see. Right. Like, should we do this? Well, let's wait and see. And sometimes that's a reaction to the market. Like, let's just see what the market does.

There's definitely. Yeah, there's that, there's definitely one piece that we've been working on in the background that. Having dynamically generated images for us. We kind of need that, but we haven't built it out yet. So once we have that in place, we can kind of use that technology for a couple of different things, which would be nice.

Well, and even like, uh, transcripts are now, uh, an, uh, an object in transistor and maybe in the future, it'd be like, it's actually not that hard to add, uh, transcripts on the video, you know, when we render it or, uh, so there's all this other stuff that comes to get built up at the same time. I think we should talk about a little bit about podcasts website stuff. Yeah. You guys okay. To change gears and anything else, Jason or Helen that you felt like we should be talking about?

I still think we should have a bouncing transistor FM logo in the background for the YouTube

videos. Yeah. Like the bouncing DVD logo. Yep. I agree. I agree. One of the things we talked about. After John and I did our founder retreat was this idea of how do we become the most recommended podcast hosting option? And I think like John, I mean you, that website, the podcast websites we have, so. That's one of the benefits, right? And this is something you did when you were building simple cast as well.

wasn't. It was, it was, it was built for the cartoons, humanity podcasts that they did the good news podcast. And then we just sort of just stuck with it and expanded on it a little bit. And it really hasn't changed since 2020, I mean it's or 2019, or no, 2017, late 2017.

Think about how many of those websites have been built. Yeah. Like, including, like when you think about in our industry, We don't have super high churn, but you know, let's say it's whatever 3% churn that means there's like, it could be 6, 7, 8 thousands, 8,000 versions of that website have been built at some point. It's crazy to think about. And it

definitely, I mean, it it's showing its age and people request things all the time, but we just haven't added really anything. . Justin: Yeah. And we have people like doing crazy hacks. I mean, in some ways it was, it was cool because you got to see how far people could take that simple design you're dead, like, right. And then you can, you can post in custom CSS. I mean, it's basically like a MySpace page. Right. You could kind of do whatever you wanted to with CSMs.

Yeah. Which was cool, but it was time to, to. Build something else. So, uh, maybe Jason, why don't you describe how you and John started working on this? So we said, okay, we want. Build a new system, and this has always been too big of a project for us. So like, we're always like nervous about it. Cause it's basically we're building a CMS and then we have to figure out how we're going to template it and all that. So how did you and John start working on that? I think we

both kind of been thinking about it the same way. Like there there's like a lot of interrelated stuff that we want to build. Um, John had thought about using liquid, um, which is it's a templating engine that Shopify uses, um, to let people sort of build their own sites or purchased sites off the shelf and roll them out. But, um, like rather than kind of revamp our existing. Websites.

technologically though, it's gone through a couple of different thought processes. Like I think we w we kind of. Went back and forth on this a lot, Jason and I about how to actually build this thing and like serve up the websites. And I don't know if you want to speak about that, Jason, but your initial ideas for like a standalone, a standalone app that runs the website templates from like kind of. Infrastructure.

Yeah, well, a couple of different iterations. Like the very first thing that, um, that I was thinking about doing is, is just rolling out sort of having an engine that, that builds a static website and that just like lives somewhere. Um, I always sort of moved on from there on to, um, a completely separate application from chances for FM, um,

Why did you move on from that idea? Yeah, that's a good

question. Um, I think it was really just around, I think the main problem was, um, SSL, right gen. Um,

yeah, it was. How do we do that and get everyone to have their own custom domain with SSL without having to like kind of rebuild what we do or we would have had. And mess with our DNS, I think a little too

much. Yeah. Got it. Okay. I mean,

there's certain, there's certainly something compelling about it. Um, like if you just have static files that you're serving, you can, you can serve them extremely quickly. Um, but yeah, the nuts and bolts around it, we're just going to be too complicated. I think a

lot. I think a lot of the discussion too was like, how do we build this in a way that when we roll it out, customers won't notice they want us to do anything. They want to have to change their DNS. They want to fill out. I think if we had gone a different route, people would have had to basically update their DNS settings in transistor or in their, in their, you know, wherever they buy their domains. And it's just trying to avoid that pain for customers, I think was kind of a big

deal. Yeah. Yeah. That's a good point once you get that set up. Okay. So how did, so then how did you go on from there? You decide you're not going to go to the state. Generate a route. What, what did you do next? There

was, there was still sort of this, this idea of like the, that the websites would like live on this separate thing. Um, and I'm, uh, I'm a big fan of go and sort of often looking for places. Did you have it. But it made a lot of sense because there, there is a liquid engine implementation and go that's really, really fast. Um, just sort of on a whim. I, I built out like a little like proof of concept one day to, to try that out. And it worked pretty well.

um, and you try to its initial implementation and go, is, is that, but is that the receiver application you built or is that the next thing?

Um, yeah, so this is the receiver. So this actually worked out really well because we wound up using, um, this go tool as, uh, like a development tool rather than the actual. Um, website itself or the actual like web server itself, um, which is, it's sort of an easier, like it's a command line tool. So it makes it a little difficult, um, for a lot of people to use, but it it's easier to implement for us then to build some kind of like editing front end.

Yeah. You just connect it to your account API, Helen. You got it set up. Right? Did you try

it? Yeah, I've tried it. Yeah. Um, then, uh, kind of looking at the receiver and looking at different liquid tucks and just kind of, um, Getting a good understanding of like the kind of structure of what it is that we're building and how it's going to work.

in the future. Yeah. Yeah. I think this is all the, also the advantage of having a team that, where everybody has some technical knowledge, you know, For you to be able to set that up for me to be able to set it up, right. Like, and actually get it going and go through all the, I went through all the dumb questions with Jason. Uh, but you know, having, and you've worked with liquid in the past when you were working with other companies, right?

uses it. Um, I've used Shopify in the past, um, and kind of having an understanding of how their theme structure works. Um, well, now it's kind of interesting to kind of be part of that and just to be able to. Hopefully share that with our customers and future. So we can kind of give them a better understanding, educate them, right. Health articles, and documentation on that as well. So yeah, kind of teaching me at the same

time.

I was just going to say that that's a lot of the win with, um, cause there's a lot of different routes. We could have gone for creating templates, but using, um, using liquid and doing things in a similar way to how Shopify did them. Sort of immediately makes this accessible to however many hundreds of thousands of people have, have like built themes on Shopify.

Um, yeah, we ended up rebuilding our current or the classic 10th that we call it, which was the original template we launched with. We rebuilt that. Liquid and rolled that out a couple of weeks ago. And you know, the idea is if we did it right, and nobody would notice, and I don't think anybody noticed, did anyone

notice Helen? No. Don't think so. Oh, there we go. Again,

it almost didn't feel like it happened. Like we, we, we sorta, we did like all this work. We were like rebuilt basically like the engine and internals of a car. And then months later, like flip the switch and then heard nothing and

nothing.

It's kind of like maybe a couple of questions about, um, different individual elements within the website, but nobody said, have you completely replaced my entire

website? Yeah.

So then the idea, then the idea is we'll, we'll keep slowly building out additional templates. Um, I have one I built it's already in use by our website for build your SAS, but that's it. And

then we'll have to see if anyone notices.

Yeah. W we have a couple people. Or at least one person lined up to build a new template and we'll just go from there and, you know, add a handful of templates that people can choose from and switch between and customize in different ways. And, uh, eventually I think the idea is to potentially open it up for people to create their own templates. Uh, however, however they want to, and just use our liquid system, but we'll see where

that ends up. Yeah. And then we, Jason, you also added localization. Uh, which is like different language you can, it's actually pretty cool. I didn't realize you're going to do it this way, but you can just add like, um, uh, um, what do you call, what do you call an appendage to a URL? Uh, so why don't you explain how that works? Cause it's really neat and it's already live like it. We already have a few

languages. That's like one of the huge wins. Well, there's, there's like a few things that we get out of putting everything into. These templates. Um, so we have this like translation filter that you can put on anything. So you just kind of like name a thing. So like the label has episodes and then you've got that in French and Spanish and German and whatever. Other languages we have. Um, yeah.

us. Yeah. I just did the French translation it's live, which is so cool. Like. Helen, you must see this in support. Like there's people ask about that all the time. There's a French podcast, but all of our labels were in English. Right.

And also the fact that our support staff. Helps to communicate in other languages, but yeah, websites aren't. So it's definitely been kind of a deal breaker for some, um, customers in the past that they, you know, they want to release a website that they may use for their company. That's primarily in the Netherlands. And, you know, obviously it's, um, we can only provide English only navigation.

Yeah. Well, we knocked that out pretty quick. Jason, Jason built that really quick. I don't even think we weren't even really planning

on doing that again. This is the advantage of having other people here. Yeah. I, I also like how we do, we, we do that weekly meeting where Helen kind of leads us through all of the. Things she's noticed in support because I think it can, it can get easy to just forget about how people our customers are feeling and reacting. Like it's, it's easy to have blinders of like, you know, I make an English podcast. And so why would anyone need labels that are in other languages?

trends and patterns emerge over time as well. Like it's the same thing keeps coming up week after week. Um, kind of convinces us there's enough. Um, the month to make that change, especially if this kind of like a big amount of work involved, um, we've kind of got to be convinced that it's, um, with making that change.

Yeah, totally. Um, what, yeah, what else do we want to say about websites? We're getting close. Right. Getting close. You're you're building a theme picker right now, John.

Yeah, we're getting close. Uh, I would say the new theme should be available. I don't know, maybe next week, next week.

Nice. How are we doing in terms of our yearly plan? Outline? Let's just say click here. So we set our first. What's going to be on podcast V1, and then the cool-down was going to be February 14th, 2015. Did you guys do a cool-down February 14th to the 25th?

We forgot to cool down.

He didn't, he just kept grinding and then cycle two was supposed to be February 28th to April 8th, and we thought we'd still be working on podcast websites and maybe the embed player. We definitely haven't gotten there yet. Um, so we're, we're, we're pretty much on we're, we're kind of where we thought we'd be.

Yeah. I think we're doing pretty well. I think, I mean, we should probably get through all the websites. So if we really wanted to do.

In these two cycles of work, I think so we're

just talking about it yesterday and like that theme picker and each theme can be configured a little bit differently. Like they'll have different colors or some toggles or whatever, but once, once that's out there with a new theme, that's sort of, that's the whole thing. Like that's new websites, new website themes.

Yeah. There's another, uh, big part of this, which we haven't really talked about. It. Dealing with like free accounts for people,

but oh yeah. Yeah. Let's save that for the next episode. We do. Cause I think, I mean, going back to like, what's driving this, obviously we want to make the experience better for customers. We want to make the experience better for us like supporting these old templates that were never really meant to be. Distributed at this scale. And ultimately we want to have more people using transistor and yeah, we want to bring in

new customers. That was the, that was the definitely the big driving force. Yeah.

Yeah. And so this is like, this is the, this is what's going to draw people in, you know, the website is one of the things that people look at, like what does the built-in website look like? And I mean, I think sometimes it's hard. It's hard to forget.

And it's cool that we can just kind of add in new templates as we want. If we find a designer who's like, I'm really excited about building a podcast website, then we can hire them and, you know, have a new template that probably looks quite a bit different than from what we already have. Yeah.

Well, especially three customers, like if we have customers signing up and they're like, you know, that's one thing Helen you could watch out for is if there's like, That is designing client and they sign up and they're like asking questions. That's the, you know, that's the time to say, well, Hey, let's talk about maybe you building something because, uh, having people that are podcasters themselves, they're putting a lot on the line.

I'm really looking forward to seeing how. Like take it and run with it as well, because it's always been kind of impressive to see what people have done with the existing templates, um, with kind of even a minimal amount of customization. So, um, yeah, I think it's going to be interesting to kind of see what. Create with it really? And it'd be good to keep an eye on

what people have built in. Yeah, for sure. All right. We're at about the 43 minute mark. So I think we're going to wind things down before we read out our patron on supporters. Uh, if you have questions. If you have questions for Jason and Helen, if you want to, you know, you want to know what's it really like working with a couple of old pieces of coal, like John and I, uh, tweet us get ahold of us somehow. I'll put all of our Twitter handles in the, in the show notes, except for John John's.

Yeah. I'm really glad we finally got through this. Yeah, we've talked about it before. Um, this is fun and it's been, I don't know. I really enjoy the team we built so far. It's really fun working with everybody. I think we're getting a lot of, a lot of good work done with a very small team. Yeah. And I know, I know initially Justin and I were both like, uh, hesitant to hire anyone full-time.

Yeah, I agree. So, John, why don't we read out our good Patrion supporters? Uh, yeah. Do the shout outs.

Uh, thanks to everyone as I was for supporting us in Patrion, we have Marcel filet from, we are bold dot AAF, Alex Payne, bill condo, Anton Zoran from prod camp.com Mitt. Harris county from the intro to CRM podcast. Olay Coolic Ethan Gunderson, Chris Willow ward Sandler from member space, Russell Brown from photo.com. Noah. Praill Colin gray, Austin Loveless, Michael, sit for Paul Jarvis and Jack Ellis from fathom.

Thanks everyone for listening. Tell your friends about the show. Recommend an episode.