Build Your SaaS

Justin is worried about bandwidth costs and Jon is worried about getting hit by a bus.

Show Notes

This week, Jon and Justin are wrestling with some heavy startup founder topics:
  • What should your SaaS company's gross margins be?
  • How much is bandwidth going to cost us? How does it affect our margins?
  • Example: [Account A] does 495,000 downloads in a month, average file size is 8 MB, that’s 3960 GB of bandwidth in a month. If bandwidth is $0.01 per GB, that’s $39.60 just for bandwidth for that one account. And that’s if we can get it that cheap. Otherwise, we’d pay $198 on AWS, or Azure. If we used Google Cloud it’d be $316.
  • The burden of keeping our service up rests on Jon's shoulders. How can we reduce the risk?
  • How can we increase platform resiliency, automation, and improve our internal customer service tools?

★ Thanks to our sponsors:

  • Alitu.com: Alitu removes all the tech headaches associated with producing your podcast. They take care of processing, editing & publishing your podcast. Go to Alitu.com and check out their video demo.
  • Podcastinsights.com: If you want to get into podcasting, check out Podcast Insights. They’re a great resource for folks who want to start and grow a podcast. Podcast equipment guides, how to make money. Join millions of readers learning all about how to start, grow, and monetize a podcast at podcastinsights.com.
  • GetRewardful.com: Guess what? Affiliates is providing almost 30% of our revenue. Rewardful makes it easy for SaaS companies to start an affiliate program on Stripe.

Show notes:

You can reply to this podcast here:

Thanks to our monthly supporters:

Thanks to our monthly supporters
  • Pascal from sharpen.page
  • Rewardful.com
  • Greg Park
  • Mitchell Davis from RecruitKit.com.au
  • Marcel Fahle, wearebold.af
  • Bill Condo (@mavrck)
  • Ward from MemberSpace.com
  • Evandro Sasse
  • Austin Loveless
  • Michael Sitver
  • Dan Buda
  • Colin Gray
  • Dave Giunta

Want to start a podcast on Transistor? Justin has a special coupon for you: get 15% off your first year of hosting: transistor.fm/justin
★ Support this podcast on Patreon ★

Creators & Guests

Host
Jon Buda
Co-founder of Transistor.fm
Host
Justin Jackson
Co-founder of Transistor.fm
Editor
Chris Enns
Owner of Lemon Productions

What is Build Your SaaS?

Interested in building your own SaaS company? Follow the journey of Transistor.fm as they bootstrap a podcast hosting startup.

Justin:

Hey, folks. If you are thinking about starting a podcast in 2019, it is not too late. In fact, our friends at podcast insights dotcom have a course for you. Just go to that website podcast insights.com and scroll down on the left hand side. You'll see a free how to start a podcast course.

Justin:

It's 10 days. They'll email you a brand new, tutorial every day. It'll cover the equipment you'll need, how to pick a topic and how to launch your course. Go check it out and let them know we sent you.

Jon:

Hey, everyone. Welcome to Build Your SaaS. This is the behind the scenes story of building a web app in 2019. I'm John Buda, a software engineer.

Justin:

And I'm Justin Jackson. I'm a product and marketing guy. Follow along as we build transistor.fm. We are really becoming pros at doing those intros.

Jon:

We are.

Justin:

It it's remarkable how, like, just doing something consistently, it it starts to feel a lot more smooth, doesn't it?

Jon:

It sure does. What are we at now? 46 episodes?

Justin:

Yeah. This is episode 46.

Jon:

What kind of tea are you drinking?

Justin:

Oh, that's a good question. It's ginger. I'm in my office, and it's whatever was left over.

Jon:

Oh, I'm also drinking a spicy ginger.

Justin:

Oh, wow.

Jon:

How about that?

Justin:

Man, that's that's like

Jon:

Full of food. It's good for your digestion.

Justin:

Did you eat

Jon:

did you eat did

Justin:

you eat at a buffet as well today? Okay, that was just me. We have this we went swimming with the kids at the water park. And we have this tradition that after we go swimming I don't know why we do this. We go to an alley you you can eat buffet.

Jon:

Okay. I mean, traditionally, I was I grew up swimming, and swimming made me hungry a lot. But I'm I'm I'm glad they didn't have a buffet in the water park because that sounds like it would be a disaster.

Justin:

Could you imagine? That's the worst business idea ever. Anyway, so here we are. We're recording on a Sunday night, 7 o'clock.

Jon:

It's a little late. Yeah.

Justin:

And, which would be 9 o'clock your time.

Jon:

Yeah.

Justin:

And we I thought, you know, we got a lot of good feedback from that episode we did, 2 episodes ago where we called it Monday morning therapy. But, the idea is that we would share kinda what we're working on, what we're thinking about, what we're wrestling with. Each of us would get to share that thing, and, then we would talk about it. And so I thought I would kick things off because this week, I have been I've been in the weeds on bandwidth costs. And, this is interesting for me because I'm coming to it late.

Justin:

So

Jon:

Yeah.

Justin:

You've run a podcast company before. Yep. You have also, I'm I'm assuming you've just done more of this kind of stuff, but I had no idea. Actually, until I thought about it, I've I've never really been concerned about bandwidth. And storage, yes.

Justin:

I've been concerned about storage. I'm always running out of storage. How much storage do I need? All that kind of stuff. Number of visits.

Justin:

So typically for, like, WP Engine, you get you are, build based on the number of visits. And I know that there's bandwidth limits, for example, with my Internet. Like, if we if the kids watch too much Netflix, then it you know? Is there? Well, we've never hit it, but it's it's quite high.

Jon:

It's usually high. I think most consumers never really have to worry about bandwidth. I mean, phones have limits to cell phones usually, but I think

Justin:

Yes. But and, actually, now you're gonna have to confirm this for me because, and I'm sorry if I sound like an idiot. Now what I'm understanding is that a customer comes to transistor.fm, they upload their MP threes, and we store those MP threes. But when those m p threes get downloaded by the millions of devices around the world, there's a cost for transferring that file from our server to the individual devices around the world.

Jon:

Correct. Yeah. Yeah. There's generally no cost for sending our files to our storage, which is s 3 at the moment. So, generally, uploading files are free, downloading files are not.

Jon:

And they yeah.

Justin:

So this is so interesting. And my guess is then that our cloud service provider, they have to pay, like, the cable company for Bandwidth? Who are they paying?

Jon:

I believe they they sort of there are these major, like, backbone companies that that just sell, like, bulk bandwidth.

Justin:

Okay.

Jon:

I mean, these these providers have, you know, major server platforms in large server farms across the world, and each of those buildings just has, like, fiber coming into it. Yeah. That are hooked into various backbones that all hook into the Internet, and then there are companies that run these backbones. They don't really pay they don't pay for bandwidth because they're the ones building the infrastructure to actually allow data to move around.

Justin:

So And are what's the cost? Are you just paying to use their pipe? Like, is that and what is the those infrastructure companies, what are their costs? Like, electricity or the

Jon:

Oh, yeah. It's probably a lot of electricity, and I think all that infrastructure and laying those cables and pipes are probably not cheap. May you know, maintaining them and

Justin:

It would be it would actually be interesting to know how much, like, actual cost is associated with sending 1 megabyte over one of those pipes? Like, what is the if you dig down you know how you Elon Musk is like Elon Musk's whole thing with rockets is that he was he was, like, well, how much does it cost? And they would be, like, well, it's, like, $10,000,000,000 to go to the moon. And then he would be, like, okay. Well, if I price it out myself, if I go direct, how much does it cost?

Justin:

Right. Because that's the frame of mind I'm in. Yeah.

Jon:

So are you saying we need to build a server farm and buy pipes?

Justin:

Well, I'm I don't know yet. I I mean, Elon Musk had to, like, go and buy things from Russians and do all sorts of stuff. So we, you know, we might have to get we might have to go and and do some crazy things. But okay. So let's go back to kind of the work I did.

Justin:

And in some ways, I think I just as a cofounder, I just had to do this work myself to really dig into it and figure it out. And, some of it might have been frustrating for you because I was just replicating

Jon:

No. No. No. I I I never went into that much detail though about, like, how much one of our shows would cost and what, you know, how much profit we'd make on each show. Mhmm.

Jon:

I I sort of came at it from a slightly different perspective.

Justin:

Okay. So let's first, let's talk about so what kicked this off for me is I had an investor email me, and and he said, I'm hearing from a few folks, you know, second, third hand that as things scale up, bandwidth costs for podcast hosts can mean materially different margins compared to different software companies. And he say he said, I'm hearing numbers like 30 to 40% of MRR just for bandwidth costs. Now if you go back to our archives, we've discussed this before.

Jon:

Mhmm.

Justin:

Yeah.

Jon:

I think I think we knew from the get go that bandwidth would be our number one cost.

Justin:

But I had not really dug into it yet. This email, it it made me think, okay. I know that things can get costly, but really that much? And and what he's talking about there is if we pull in a $100 of MRR, that's just like gross revenue coming in. We've gotta pay salaries.

Justin:

We've gotta pay hosting costs. We've gotta pay marketing. We've gotta pay customer support. We've gotta pay for all of our software. We've gotta pay for we're a remote company, so we have to pay for a retreat or 2 every year.

Justin:

You know, there's all these things. And typically, what's great about software software as a service compared to opening up a donut shop is that your margins are much higher. And, you know, I went online and looked at, you know, what should the gross margins for a SaaS company be? And LeadEdge Capital says, you know, they should be 60 to 80% with the primary cost of goods sold being network and delivery costs, which is what we're talking about today

Jon:

Yep.

Justin:

As well as services, personnel, you know, maintenance training, implementation. Although and and they say, you know, but as things scale up, you should be definitely above the 80% mark. And then I went to Quora, and they say, you know, in the software industry, gross margins have always ranged between 80 to 95%. So margins above 80%. Gross margins above 80% seems to be the general rule of thumb.

Justin:

And so then I started calculating things. And it's amazing because, you know, you think Amazon Web Services what is it? Amazon Web Services is are they 2¢ per gigabyte or 4¢? Now I can't remember.

Jon:

I'm not sure. They're it's expensive.

Justin:

Yeah. So here's I'll give you an example. Here's a really popular show. 495,000 downloads in a month. So 495,000 times one of their podcast audio files gets downloaded.

Justin:

The average file size is 8 megs, which is actually quite small.

Jon:

It is. Yeah.

Justin:

So that's 3,960 gigabytes of bandwidth in a month. See how quickly that added up? Like, it's like, oh, this is so small. Okay. But 3,960, gigabytes.

Justin:

Is that correct? Oh, and by the way, I I know some geeks are gonna get after me for this. What I'm doing is I'm multiplying 495,000 by 8 megabytes. That gives us 3,960,000, and then I divide it by a1000. Now I had some geeks kinda take me aside into a dark alley and say, hey, you idiot.

Justin:

It's a a gigabyte isn't a 1,000 megabytes. It's 1,024. Right?

Jon:

Yeah. 1,024.

Justin:

And so I was like, oh, man. Why why did I have that idea? But then I talked to all these cloud providers, and they all use a 1,000 meg gigabytes. That's they they round down.

Jon:

Interesting.

Justin:

So, I mean, it it's probably because sales and marketing folks.

Jon:

They don't they don't wanna have to explain the history of computer science and why that number is that

Justin:

way. Yes.

Jon:

Yeah. Probably.

Justin:

So I divide it 3,000,000,960,000 by a 1000. That gives us 3,960 gigabytes. Does the math

Jon:

Which is roughly 4 terabytes.

Justin:

Roughly 4 terabytes. Correct. Because you also do a 1,000 gigabytes and terabyte even though that's not roughly that's not even correct. Okay. So the cheapest bandwidth I've been able to find so far is 1¢ per gigabyte.

Justin:

Now that's not what we're paying right now, but 1¢ per gigabyte. That seems so cheap, doesn't it? Just 1¢.

Jon:

Seems cheap. How does Who

Justin:

can't afford a cent? You know? But that just for that one show, that's $39.60 just for Bandwidth. That's not for storage. That's not for anything else that we don't get paid.

Justin:

We don't so Yeah. You you folks can see at home, you can see the problem here. In order to get margins of 80%, like, we'd have to charge that that customer, well, definitely more than $99, which is our highest, tier. Right? So and and by the way, that's if we can get it that cheap.

Justin:

What we'd pay on Amazon Web Services for that account, A $198. Same as same with Azure. And on Google Cloud, it would be $316. So this really kinda blew my mind.

Jon:

Yeah. It adds up quick, and that's yeah. I it's a thing I ran into in the past with s 3. It's a thing we've started to run into as we've grown with transistor. Mhmm.

Jon:

And it's obviously really good to think about it now, and I'm really glad you kind of did all those calculations.

Justin:

And, by the way, in the past, I have been the kind of guy that hides from numbers and reality. That one of the this this thing has been in the back of our mind since we started our partnership. We knew, like, we've got notes. What is our biggest threat? It says right there, bandwidth.

Justin:

We the first thing you'd said to me when when you as soon as I signed the document and you had my soul, you turned to me and you said, Justin

Jon:

In a in a deep voice.

Justin:

Smoke came out of your eyes.

Jon:

Dark days are ahead. I gotta talk

Justin:

to you about bandwidth. And so we we've known, but it's been in the back of my head, and I have not crunched the numbers. So, folks, if you are out there and you have not yet crunched the numbers, this is an important step. And, really, it shows that our profitability rests on 2 variables in terms of costs, average file size and average number of downloads per month. So Right.

Justin:

If you have a show like ours, which is about 45 minutes long, That's about, I think we our average file size is 35 to 45 megs, something like that. Now that's a fairly large file. That's fine as long as this show doesn't get a 100,000 downloads in a month. Right?

Jon:

Right.

Justin:

If this show gets a 100,000 downloads in a month, I had a friend send me his numbers. And this is a friend that I was like, man, I would love to have him on Transistor. Like, he would just be a great brand ambassador. We're always talking about it. Like, I should get you on the platform.

Justin:

I saw his numbers, and I'm like, oh, shit. We can't Yeah. We can't afford it. So a 100,000 monthly downloads. He's doing 5,388 gigabytes per month.

Justin:

Cheapest bandwidth we could get, and we haven't been able to figure out how to make this bandwidth work for us yet. But it's 1¢ a gigabyte. That's $53.88. But currently, I think our cost is 2¢ per gigabyte. So just to host his show sorry.

Justin:

Not to host the show. Just for the bandwidth for his show in 1 month is a $107 76¢. How many people do you know are paying well, even a $100, but to get our margins Yeah. Like, we would have to

Jon:

Like,

Justin:

let Is

Jon:

that $500 a month

Justin:

or something? Like, it ah, jeez. Let's look. I got it. I've got my little calculator.

Justin:

I'm gonna just pop it in. Oh, man. So, yeah, if if we wanted margins of Did

Jon:

you have do you have your is that your TI 83?

Justin:

Yeah. I got my TI 83 here. Okay. So just to get oh, man. To get to 80% margins, but just that's just with bandwidth.

Jon:

Yeah.

Justin:

We would have to charge him $549 a month.

Jon:

Yeah. I don't think he'd I don't feel like he wouldn't be up for that.

Justin:

We should

Jon:

I I

Justin:

already asked him, and he said no. So jokingly, but I, you know, no one's gonna pay that. Well, no one's gonna pay that if it's just bandwidth because so this is the interesting discussion is that what is valuable for the customer? Now we have customers that have emailed us and have said, you are charging too little. And And I'm like, what are you talking about?

Justin:

Go and look at their downloads. They're doing, like, a 100 downloads per month. What Mhmm. What's the deal? And we've talked about this before, but some of these folks are are running private podcasts or they're running, podcasts for organizations.

Justin:

Their primary metric for success is not, thousands or hundreds of thousands of downloads. It's just that their members get the audio in some sort of serialized way.

Jason:

Right.

Justin:

And they're saying, man, you should charge more than this because we can host all of our shows on your $19 plan, and we're getting so much value. We get great customer support. You know, for them, they don't care about, you know, being the next Joe Rogan.

Jon:

Yeah. For them, it's a great deal. Yeah. They're not they're not going to blow our bandwidth costs out of the water.

Justin:

Yes. And you can see Zedcast, Zeddashcast, they recently closed down because they got a really popular podcast on the platform, and it it bankrupted the company. Yeah. So this is a big issue. And, and so you can see those two variables, our target customer and our pricing strategy moving forward have to address those two things on the cost side.

Justin:

Now on the customer side, on the demand side, they might not care about either of those things. Right?

Jon:

Right.

Justin:

But we it's something we need to consider when we're we're figuring all this stuff out. You know, we have a lot of shows that are in that thousands of downloads per month, and those are a lot more manageable. Like

Jon:

Mhmm.

Justin:

If it's, you know, if you're doing 2 or 3000 downloads a month, that's great. It's great for you and it's great for us. But it's once you get into those big numbers doing, like, over a 1000000 downloads in a year, man.

Jon:

Yeah. I mean yeah, there's a lot I think there's a lot to dig into with that.

Justin:

And we have some popular shows. We we have some users that I begged to transfer over early on.

Jon:

Yeah.

Justin:

And now, I mean, I love them, but I'm looking at, wow. This is costly, which by the way, I I hate to call this out publicly, but I'm gonna do it anyway. But because Anchor wrote a a blog post about how hosting should be free and that it doesn't cost anything, and they were just talking about storage costs.

Jon:

Yeah. It doesn't yeah.

Justin:

Sure. They did not talk about this. Storage costs are

Jon:

very cheap. They're very cheap. It's in the Even on it even on s 3. Very cheap. It's like we pay, like, a couple dollars a month.

Jon:

It's really cheap.

Justin:

Yes. That's see, that's of course. So yeah. That it's it's cheap to store them, but it's not cheap to transfer them from our server to your phone. And by the way, like, we're happy to do it, and we part of me got fired up about this.

Justin:

Part of me was, like, a bit, like actually, I I don't think I was ever discouraged. I was just, like, man. Okay. This is a problem we need to solve.

Jon:

And it and it is solvable, and I think there are a lot of options, thankfully.

Justin:

Yeah.

Jon:

And I and, like, your math is great. I don't think it necessarily covers everything, though. Like, it's accurate with what you had, but I think there are a number of of options we have that I think would make it a lot cheaper than 1¢ per gigabyte.

Justin:

Yeah. And and so we've looked into a few options. Do we wanna talk about any of that?

Jon:

Yeah. Sure.

Justin:

I don't wanna go too long because I wanna get to your your stuff here. But Yeah. Yeah. Just because I know people are gonna write in.

Jon:

Yeah.

Justin:

And by the way, we love it when you write in. It was nice getting all that feedback. We had a a lot of people say that they could identify with the way you were feeling, the last time we chatted. So keep keep writing us in if this resonates with any of you or if you wanna help us crowd crowd crowdsource the the solution for this. Specifically, actually, if if any of you work at Cloudflare or know somebody that does, if you could have them get a hold of us.

Justin:

Who's this? Who's the CEO of Cloudflare? If he's listening or she's listening, I want to Matthew Prince. Okay. Matthew, if you're listening, give me a call.

Justin:

Yeah. So we're we're trying to figure some stuff out, and there might be some creative ways of doing it. Maybe one thing is to talk about because you you said you get what you pay for with

Jon:

You get yeah. You get what you pay for. So I I come at this sort of from the performance and price scenario. Like, obviously, ultimate performance is you sign up for, let's say, you sign up for a CDN, which is a content delivery network that is you know, there's hundreds of these things out there. You you have one origin point, which, let's say, is s 3, And then you sign up for a CDN that pulls your media files from s 3 once and then distributes it with a within their network globally.

Jon:

Mhmm. And then all of the media is served from their CDN. That's obviously, like, ultimate performance, ultimate, like, resiliency to downtime. But, generally, those get fairly costly, and they're generally more than 1 cent per gigabyte. Although, occasionally, you can have plans where you pay a set amount upfront per month for x amount of terabytes of data

Justin:

Yeah.

Jon:

That are that are lower than 1 separate gigabyte, and then any overage is x amount of cents per gigabyte. So as you grow with those types of plans, generally, you can negotiate cheaper prices Mhmm. And probably get it below, like, 1¢ per gigabyte because and we're not talking about, like, you know, transistor's not serving, like, a couple terabytes a month. Like, we're up to, like, 15 or 20 a month terabytes of data, and it'll keep growing. And, you know, we'll we'll break gonna get into the hundreds of terabytes a month.

Justin:

How many are we doing right now?

Jon:

It's about 15 or so between 15 and 20. Okay. Grow I mean, it's growing every month. Wow. So that's like ultimate CDN, full CDN, ultimate performance, very high price.

Jon:

Now there's, like, other scenarios we can take that are pretty good performance

Justin:

Mhmm.

Jon:

For a much better price. And I think when we talk about performance for podcasts, like, we're not live streaming data. We're not like, latency does not matter. Really what matters is there's, like, a pretty decent download speed, and you can handle a lot of concurrent connections. Yes.

Jon:

Because because these podcast players will be downloading, shows pretty much at the same time when the feed is updated. Mhmm. So, you know, you'll you'll be downloading your your audio in the background, and everyone's gonna be pulling from your servers at generally the same time. So there's dedicated servers that you can purchase, which are just like you get your own box that sits in some server location somewhere. And those generally have basically unmetered bandwidth, which means it's as much as you can pipe through it as you want.

Jon:

But that's limited by the speed of the network connection, that that box has. Mhmm. So let's say you get, like, a dedicated box that has unmetered bandwidth, but it has a, like, a 1 gigabyte per going gigabit per second connection, like, from its location to outside.

Justin:

Mhmm.

Jon:

So, like, if you just fully pack that pipe with data, it's only you're only gonna be able to send out x amount of terabytes a month anyway. So you're you can, like, totally load up the pipe with data and overload it. But you can also usually upgrade those to, like, 10 gigabyte per second connections. So it's technically unlimited, but there is a cap. There's, like, a not really a physical cap, but, like, you can really only put so much data through this thing per second.

Jon:

So that's really, like, the cheapest way to go. And so that I think this and we've tried this with a with one dedicated host, and it was just, like, not a good like, that that one was not good quality. Like, that's where I kinda got into, like, you get what you pay for. Mhmm.

Justin:

Because what happened? Explain to people what happened.

Jon:

It's like it it's like I I have a friend of mine, Joe, who who's helped us, with a few things for our, bandwidth. So he helped me set the server up and helped automate a few things. He's like he does a lot of DevOps work. From the get go, he was, like, kinda noticing that, like, there were some issues with it. And I forget how much this even cost a month, but it was technically unmetered.

Jon:

And so we set it up, and he installed this, you know, server software that he likes. And then I think we used that for a couple months, and it worked, like, it works pretty well. Like, speed was fine. There were no real big issues. But then the server just, like, crashed.

Jon:

I think one of the hard drives failed or something. And for whatever reason, like, we couldn't reboot it. We couldn't get into it. We couldn't, like, SSH in and do anything. Mhmm.

Jon:

And I emailed support, and they were like, yeah. We do I we don't know. Like, you can't do anything. We're like, okay. So I guess we'll just cancel the account.

Jon:

And then, you know, luckily, we have this, like, failover where you we can just switch it back to s 3 real quick. That won't interrupt service. And I think at that point, it wasn't a big deal because we weren't serving, like, terabytes of data. But so that's the stuff that can happen. Right?

Jon:

Like, you find these deals on the Internet. You're like, oh, this is a great deal. This is cheap bandwidth, but but they're not they don't care about you. It's just like

Justin:

Yeah. Yeah. And this is also the problem with with crowdsourcing, which we're still happy to do. But the the problem with crowdsourcing some of these problems is that folks will find a solution, but they don't have to live with it. Like, we're the ones that have to live with it.

Justin:

And Yeah. Right now, one of our hopes for this company is that we could run it as a 2 to maybe 3 or 4 person company. And if we're gonna do that, that means we sometimes we need to spend more so that if John is in an airplane, it's okay. You know, everything's gonna be alright.

Jon:

That's a good that's a good segue to my

Justin:

thoughts. Good segue. Okay. Well, we should get we should get into that. But before we do before we do, John, guess how much this podcast cost us to deliver last month?

Justin:

$25. No. We had 12,778 downloads last month. We're thankful for every single one, dear listener. And that cost us $5.96 in Bandwidth.

Justin:

And so to help pay for those bills, we have alitoo.com. And they you know, Colin really got it right, didn't he? Because with alitoo, you just either upload your MP 3 or you record your MP 3. He just has to store it. He he he doesn't need to send it to

Jon:

We chose the wrong

Justin:

We chose the wrong vertical. But Colin, he's smart. Alit2.com. It basically helps you record and edit and then publish your podcast automatically. They've just released recording within the app, and this is a question we get all the time.

Justin:

You know, how can I make the production of the podcast easier? This is actually something I think about all the time because even even you and I, we've got, you know, we got microphones, and we have to be I can't just like it. It's still difficult to create a show, but Alitu put some of that on the rails. And, Colin's been nice enough, kind enough to support this podcast. Go check him out, alitu, alitu.com, and tell him that we sent you.

Justin:

So, John, what's, yeah, what's that's enough for me. Yeah. What's going on in your world?

Jon:

So, I mean, I've I've just been working on a lot of stuff, a lot of different features, and working with you on the bandwidth the bandwidth problem. But I recently, I've been kinda thinking a lot about what you ended the last segment with, which is, you know, what do we do if John is on a plane, or what if I'm on vacation, or, you know, whatever. I have to, like, take some time off and can't can't do anything with transistor or can't fix any bugs or things like that. Right? So, I've been thinking sort of about, like, easy wins now that we could build in that would sort of help with, I don't know, I guess, platform resiliency as I wrote it down and and things like automation and building out certain customer service tools that would help both you and I, you know, fix these problems that our customers would have.

Jon:

So so, like, you know, it there's 2 of us. There's a it's a small team. I'm more on the engineering side. You know, you've helped some of it, but, like, I know the ins and outs of the platform and and where things can go wrong. And if things can go wrong, here's what you do.

Jon:

Mhmm.

Justin:

Is this related is this related to, the Quadriga Bitcoin fiasco?

Jon:

I have no idea what you're talking about.

Justin:

Quad Quadriga, which is the which was a Canada's biggest crypto exchange.

Jon:

Uh-huh.

Justin:

The CEO this is quite sad. The CEO died.

Jon:

Oh, that one. Yeah. Okay.

Justin:

And he, had the he was the only one that knew the password to get into the encrypted store. And so the exchange lost a $150,000,000 worth of crypto.

Jon:

Oof. Yeah. I think this

Justin:

story is still developing, actually.

Jon:

Yeah. Yeah. It's it's somewhat related to that. I mean, you know, I don't have that kind of crypto, but I have none, in fact.

Justin:

But yeah.

Jon:

But let's but let's say, you know, I I got hit by a bus or something, and I was in a coma for 2 weeks, Or I don't think that'll happen, but, on vacation. Whatever. And and something horribly wrong happened with our servers or our CDN, and nobody could use our service or nobody could download podcasts. Like, obviously, our customers aren't gonna wait for a week or 2. Like, that would just kill us.

Jon:

I don't necessarily want to have to be on my laptop all the time or or have to worry about checking my phone necessarily 24 hours a day. I mean, like, you know, there's 2 of us in this start up. I sort of have resigned myself to the fact that I probably have to travel with my laptop

Justin:

Mhmm.

Jon:

All all that's regardless all the time

Justin:

Yeah.

Jon:

Which is which is fine.

Justin:

This is such a good topic.

Jon:

There's always that, like, sense in the back of your head that you're like, gotta check my email, see if there's a new support request or something's broken. And maybe it's not something Justin can fix because I it's something I have to go in and, like, do on the server or SSH in and, like, update something that isn't necessarily updatable by you. Or let's say in the case of we have a lot of requests that are like, hey. I wanna move to the annual plan. Or, hey.

Jon:

Can you transfer these shows to this other account? Because my cofounder and I are parting ways. Like, that's not something that they can do yet in the system, and it's not something that you can even do for them yet. So it basically means that I have to do that, and that's kinda where I'm thinking. What what tools can we build?

Jon:

Like, we already built in the feature where we can take over someone's account, which helps a lot.

Justin:

Yeah. The impersonation feature. Yeah. By the way, I'm gonna put in show notes, my friend Jamie wrote a great blog post on the on the not pros and cons of impersonation, but the the, you know, he talks about the benefits, but he also talks about the hidden caveats?

Gavin:

Cave

Jon:

is that Caveats? Caveats? Caveats.

Justin:

Yeah. I'll I'll put that in the show notes as well, which is s a a s dot transistor.fm/46. Yeah. I think this is such a great topic, because it right now, that burden is is mostly on you.

Jon:

Right. Right.

Justin:

And there must be other teams that have gone through this. And actually, even though when I think about the first startup I worked at, Mike was the CTO, and the other two founders were nontechnical. And, you know, like, in the early days, Mike like, the server was just underneath Mike's desk. Yeah. And, you know, the the stress he must have gone through.

Jon:

Yeah. It's, Yeah. It's it's it can be stressful.

Justin:

So I would be definitely, folks, on Twitter, you can get us on Twitter at transistorfm and let us know what you think.

Jon:

The tool the tools that we can both both use, I think, are are worth thinking about. Mhmm. That's obviously not gonna help you necessarily fix, like, major technical problems, but, like, there are certain there are certain problems that probably could be fixable by you. Mhmm. Yeah.

Jon:

Or someone else, like, even on AWS where you could log in and just, like, restart a server. Right?

Justin:

Yeah. So I think one thing that's coming to mind right now is the, the idea of you doing screen casts whenever something comes up. And the idea is every time you record 1, you're thinking this is insurance. So you you record the screen cast, and then we have a shared Dropbox folder. You just drop in Dropbox.

Justin:

Hey, Justin. This is how you restart the server if it breaks. And I could watch that, and then I might have questions. And then I might say, well, next time this happens, let me try to do that with you on the phone. But at at least to have the video documentation would be super helpful.

Justin:

The the the other kind of strength I have is I have a big network. And so, you know, there there's been times where you haven't been around, and I've had a programming question. Well, I just go out to my network. I I reach out. And if I was able to do, I mean, even the way I got started with our getting our our rails app up and running on my machine was my friend, Francois, here in Vernon just sat next to me.

Justin:

And every time I hit a a roadblock, he's like, okay. Well, what you gotta do is you know? So I I think just the documentation would be really helpful.

Jon:

Yeah. Yeah. That's true. That's a that's a very good point. Yeah.

Jon:

There's a there's a lot that I've I've done, you know, a handful of times. It's basically the same thing just for different people.

Justin:

Yes. And and I can see how, like, when you're doing it, it's just like, okay. Well, I'm just gonna do it.

Jon:

Yeah. But also yeah. But also the more the more times you do those things, you're like, oh, this is actually like a feature that we should build. Obviously, someone should be able to switch to yearly billing if they want.

Justin:

Yeah. Yeah. Yeah. Yeah. I and but I think there's still value, especially right now.

Justin:

You know, one thing I've been thinking about on my side is that I just and one of the reasons that I've been pushing myself to, you know, build little features in transistor whenever I can and get in there is I I do wanna have a better understanding of the bones. And, definitely, if we're ever gonna scale this, you know, we think this is gonna be a 2 to 4 person company, but this could end up just being a 100 person company. I don't want all of the lessons that you learn at the beginning when those features aren't built. I don't want those all in your head. I I so even if we were gonna build some internal tools, which I think we should, at the very least, I want a video that says, okay.

Justin:

This is the old way of doing it. So if all of the abstraction gets stripped away, this is where it is. And, you know, I think there's benefits also if we ever sell this company, if we ever you know, there's all sorts of other reasons that might be helpful to have. And if we had just this treasure trove of of John just going, okay. Hey.

Justin:

I'm just doing this right now, and, I'm gonna keep those assets somewhere. Yeah. And I I've actually started doing this myself with, you know, here's how I'm here's how I'm solving this problem for our customers right now. I I think I started a little Dropbox folder called something like archive customer something. But that's one idea that kinda came to me.

Jon:

Yeah. I like that. I think yeah. I think you've got you had better habits of of, like, recording screencasts when you're doing things. Probably probably just from your, you know, history of podcasting.

Justin:

Yeah. I I I it it it's kind of like a habit you get into. Oh, yeah. Like, the other day, a customer asked me. They said, I'm trying to create an Alexa skill from a podcast.

Jon:

Mhmm. And

Justin:

it's not working. And I said, no. That's interesting. Well, I've never tried that. So I'm just going to hit record, and I'm gonna go try that.

Justin:

And, now I have a little screen recording in our screen recording folder called how to create an Alexa skill.

Jon:

Oh. And and

Justin:

the idea is if anyone ever needed to come in and say, okay. How does that work? How did, like, where what is that? They could they could find that. But you have some other good questions like, you know, should we focus on more internal customer service tools to help with this?

Justin:

We could have someone on contract, which I think that that's not something we should rule out.

Jon:

Right. I mean, yeah, I'm thinking, like, we could have someone sort of on standby if I'm unavailable or you or there's something that you can't do who, you know, knows the system is, like, probably someone we know, I would say, that we can trust

Justin:

Yeah.

Jon:

Who who they who can bill us just, you know, by the hour for something. Like, those that gets expensive, but, obviously, it the cost is much lower than, you know, not having a service running for a week.

Justin:

Oh, yeah. I think I think that's worth investing in for sure. We need a mercenary.

Jon:

Yeah. So I have I have a couple of friends who have worked together for a long time and have done a couple different start ups, but they're both technical. Right? So, like, they you know, it it's okay if they sort of take a vacation because either one of them could fix things. But they also had a they also had a rule of, we can't take vacations at the same time.

Jon:

We also can't really go to the bar together. Especially in the early days, they were having, like, all kinds of problems, like, keeping this thing running. Yeah. And they were like, we can't go have drinks at the same time because if we're both drunk, then it's just not gonna work. Which is interesting.

Jon:

I mean, they live in the same town, so it's a little different. But

Justin:

Yeah. Yeah. That's so good. I mean, this I'm sure other people like, Jason and David must have had to figure this out when they were doing base camp. And the and I'm sure every team has had to figure this out.

Justin:

Like, every team early start up has some person who has all of this stuff in their head. And then it's like, wait a second. We gotta get that out of your head because this is a huge risk. And so, yeah, I'm I'm curious about this and thinking and it actually personally, again, I I think I I can only get so far, but it does motivate me to wanna figure some of this stuff out. Yeah.

Justin:

I think there's something about that. And, actually, I did this at the first place I worked. Eventually, I I was able to because I would get, if I woke up early I was waking up early at, like, 5 in the morning to check customer support.

Jon:

Mhmm.

Justin:

And if something was broken, I had the ability to restart the server.

Jon:

Okay.

Justin:

And so I think there are internal, you know, routines and and systems that we could start documenting that could be helpful.

Jon:

Yeah. Absolutely. And I think, you know, out of that, like I said, it will come features or dashboard features or customer service tools that we can build. And, like, the 2 friends I mentioned that work together, they they they ended up building this web dashboard for themselves where they could honestly do, like, 80 to 90% of the stuff they needed to do from their phone. Yeah.

Jon:

And just be, like, add add 2 servers to the cluster, restart the server, do this, do that, and they just, like, hit a button on their phone and does it Yeah. Which is really cool.

Justin:

Yeah. And and you could see how that would be powerful if you ever did brings other people on, and you eventually you know, when they get to a certain level, you say, okay. We're gonna give you the black the black app. And it's like it's like the secret app that you open up, and it's like the little black book with all the passwords and all the The

Jon:

suit the suitcase, the football with the nuclear code

Justin:

or whatever. Exactly. And and then you've got some people that can help you out.

Jon:

Yeah. I mean, luckily, you know, AWS is is pretty good, and they have, you know, they have a lot of tools built in to scale up if you need it based on certain scenarios or, like, reboot servers automatically and stuff like that. So, like, a lot of those cloud services, they're pretty good with that that type of stuff.

Justin:

Yeah. This is a good topic, though. I'm glad you're bringing it up. I think I think this is, like, a good something that we definitely need to figure out and be aware of. And, I'm sure lots of other people are as well.

Jon:

Yeah. The other part of this too that I went I went down this rabbit hole of, like, how can I buy a $300 iPad and code on that?

Justin:

Yeah. And

Jon:

it's not it's not easy.

Justin:

Yeah.

Jon:

Like, it's really not you can buy keyboards for your iPad, but, like, it's really the workflow there is not great. Yes. And then I'm like, could I buy a cheap Windows laptop? A small one to carry and to travel with? Or like

Justin:

You mean just something you didn't care about. So you you you could

Jon:

Something I didn't care about that's not a 15 inch MacBook Pro that's big and heavy.

Justin:

Right? But you would hate that too. You'd be something would something would melt down, and then you'd take a shitty little laptop. The trackpad doesn't work. I think Yeah.

Justin:

Keys are falling off. And you're like, ah, shoot. And you're like, somewhere in, like, you know, Brazil in the jungle. Right.

Helen:

Fucking Wi Fi.

Justin:

Yeah. Everything would be going wrong, and you're, like, trying to hook up to the satellite phone. And

Jon:

Although those the 12 inch MacBooks are nice, but they're unless you can find a used one, they're not really cheap.

Justin:

Yeah. I think I think the idea of building in, some building in some like, you should be able to go away and not worry, and there should be a plan and capability for that. So

Jon:

That said that said, like, you know, I think time off is great, and, obviously, everyone needs it. But, like, it's still exciting. Like, it's still exciting. I mean, you know, it's we're partners in this, and it's exciting business. And I don't generally mind it, but there is something to be said for just, like, completely taking a break.

Justin:

Yeah. Unplugging. For sure. Alright, folks. We are going to, close things out here.

Justin:

I wanna talk to you about our last sponsor, and this is get rewardful.com. They have been running our affiliate program since we started, and a surprising amount of our revenue comes from affiliates right now. I think it I think it's it's not quite 30%, but it's it's getting there. And this is a channel that's worked really well for, software companies like ConvertKit, companies like WP Engine. They are running on, you know, affiliates who sometimes build, you know, big audiences and are really incentivized to get the word out about your product.

Justin:

And what's great about GetRewardful is they hook up directly to Stripe. So a lot of these this software is not good at all. It's it's really, really bad, very expensive as well. And, these folks are just really incredible. It's started by my friend, Kyle Fox, and I trust him with my life.

Justin:

So go check it out.

Jon:

I will say it was very easy to integrate.

Justin:

Yeah. It's it's and it's it's nice because it you know, it you know that, like, someone brings a customer in for you. When that gets charged in Stripe, GetRewardful just says, okay. Well, now here's, you know, your affiliates, cut. Right?

Jon:

Mhmm.

Justin:

So go check out get rewardful.com. Let them know that we sent you. The link is in the show notes as well. And, John, why don't you go through our patrons, our monthly supporters?

Jon:

Alright. Yeah. Thank you to all of our, patrons on Patreon. We have Colin Gray at elite2.com, Kyle Fox, get rewardful.com, which we just mentioned, Darby Frey, Samori Augusto, Dave Young, Brad from Canada, Kevin Markham, Sammy Schubert, Dan Erickson, Mike Walker, Adam Devander, Dave

Justin:

Junta,

Jon:

and podcast insights dot com.

Justin:

Thanks again, folks, and we will see you next week.