Skip to content
PodSubs
ShowsPeopleGuestsJobsNewsEvents
More
BlogPodCred
Menu
ShowsPeopleGuestsJobsNewsEventsBlogPodCredNetworksSign in
Join nowSign in
  • Shows
  • People
  • Jobs
  • News
  • More
    • Guests
    • Events
    • Blog
    • PodCred
    • Networks
    Join nowSign in

Smart Bear Live

Smart Bear Live artwork

E14 w/ Ben Orenstein: How to Price Your SaaS for Long-Term Growth

Nov 15, 20261 hr 10 minAudio

Rails developer/evangelist Chris Oliver has never done outbound sales: “I think I’ve just been afraid of it.” His startup (Hatchbox) deploys apps on servers the customer owns, and its best accounts run more than 30 apps each, yet he has not even emailed them. How does he start selling to accounts that size? Selling means spending the day talking to people, and that, he says, “doesn’t feel as productive.” (And surely not as fun?) Cohost Ben Orenstein and I dig in: A candid working session including: - Claude doing Hatchbox’s job? Chris: “Probably close enough” - can AI one-shot your niche SaaS yet? - the 24/7 support Chris gives for almost nothing - Jason’s price test: worst case, “they all leave” - Ben dares Chris to raise his prices ME ME ME!!! Subscribe on YouTube Subscribe via Podcast 📘 Get my book 📘 My bio + 20 years of articles Twitter | LinkedIn STARRING CHRIS OLIVER Hatchbox GoRails Jumpstart Pro Remote Ruby podcast Twitter | LinkedIn WITH COHOST BEN ORENSTEIN Tuple Hackers Incorporated podcast Twitter | LinkedIn REFERENCED IN THIS EPISODE “Hidden Multipliers”, Chapter 11 “Stack Overflow Developer Survey 2025” “Stripe Projects” - Stripe Aaron Francis TIMESTAMPS (00:00) - Rails hosting stalled, then came AI (02:15) - From GoRails tutorial to hobbyist host (06:46) - Hobbyists are the tail, not the head (09:18) - Why he avoids selling to big accounts (16:44) - A strategy in your own shape (23:24) - Right business? Can Claude replace it? (26:47) - Hatchbox as the precision tool AI uses (32:25) - Go ambitious, go niche, or be a brand (34:56) - A Rails and AI YouTube channel (40:08) - Will AI one-shot small dev tools? (46:14) - Rails isn't shrinking: teach again (50:12) - 24/7 hosting support at $20 a month (55:38) - Why would AI agents pick Hatchbox? (01:02:59) - Influencer, or charge for support? (01:07:02) - Tough it out? Raise prices as the test

Audio and video play straight from the show's own hosting. PodSubs stores and serves no media.

Transcript

From the show.

Jason Cohen (00:00) The star of today's episode of Smart Bear Live is one of the most prominent educators in the field of Ruby on Rails applications. He also built an application hosting product for Ruby on Rails, but despite his visibility, he hasn't been able to get the hosting company growing the way he wants. And now he's facing the impact of AI, who isn't.

Chris Oliver (00:18) Hatchbox…

Chris Oliver (00:18) It's kind of like, our life raft in a sense, but isn't necessarily like the right product, I guess.

Chris Oliver (00:24) It's ended up where it's primarily, I think, hobbyists using it to deploy things.

Jason Cohen (00:29) Don't give you any money. They quit a lot.

Jason Cohen (00:31) And that's tough.

Chris Oliver (00:33) It was always a nightmare amount of work… But the least amount of revenue…

Jason Cohen (00:36) Because…

Jason Cohen (00:37) That's a bad business model.

Jason Cohen (00:39) What's your heart telling you right now?

Chris Oliver (00:40) Just to figure out, okay…

Chris Oliver (00:42) How do we fix the business model if we did try to do that. Um…

Jason Cohen (00:46) So today we are going to help him sort out who this product is for and what it should cost. And who is this we, you might ask? I'm Jason Cohen, known online as Smart Bear, and among other things I'm the founder of WP Engine, one of the most prominent managed hosting platforms in the world that specializes in WordPress. We have over 200,000 customers. We host more than a million websites. So in this workshop, I'm bringing the perspective of managed hosting. And I also just wrote a book called Hidden Multipliers, which has all my best ideas about how to grow your business, but you can look at that later, because joining me now to help our star founder.

Jason Cohen (01:19) Is the great Ben Orenstein. Which is remote pair programming tool for software engineers. When he started Tuple, everyone said that developers don't pay for tools like that. And yet Ben proved that for the right product. They actually will pay for it. Ben is an expert in building products that developers want to buy, which is also what the star of the show needs to do.

Jason Cohen (01:21) Co-founded Tuple…

Chris Oliver (01:39) All right… Does it make sense to do this for another year…

Ben Orenstein (01:41) Seeing if you can find that fit again seems promising to me.

Jason Cohen (01:45) So let's now turn to the star of the show, Chris Oliver, with his company Hatchbox.

Jason Cohen (01:52) So, Chris, thank you for being vulnerable and opening up your business and even yourself. This is a scary thing to do. And so it's brave of you to do it anyway. And it's generous to the listeners. So we all thank you for that.

Chris Oliver (02:03) Yeah, thanks for having me. Uh you gotta be vulnerable if you want to grow and learn and get better, so why not?

Jason Cohen (02:09) We'll see why not. Soon. No, I don't know. Okay. Hopefully it won't be that bad. Okay, so Hatchbox handles deployment and server management for Ruby on Rails and Node apps. And so that means provisioning servers, deploys, SSL certs, databases, all this kind of stuff. And it does that in a cloud account that the customer already owns instead of locking them into a platform like Heroku does. But that also means you're not making money on infrastructure, right? So how did you come to that decision?

Chris Oliver (02:38) You know, uh it actually started because the most popular tutorial I had for GoRails was deploying your Rails app. And it was just like some notes I wrote for myself, and somebody was like, Hey, you should check out Laravel Forge. They basically like turned that tutorial into a product, and I was like, Oh, we could do that, you know? And a lot of that was I would set it up a server, come back six months later to go fix something, or you know, a package would break and I'd forget everything and need to come back and like relearn it all.

Chris Oliver (03:12) So that's kind of where Hatchbox started was just taking our most popular tutorial, turn it into a product, um, but it lived as a side project forever pretty much was a side project until like last year when AI started to take over uh the education space and kind of destroy that. So we were like, well, you know, it makes some money. We just need to figure out how to grow it could become our primary source of income going forward. So it's kind of how it got there.

Chris Oliver (03:44) It was always a nightmare amount of work to keep up with everything. And so we never did a lot of marketing for it because it was the most amount of work but the least amount of revenue out of the stuff we sold. So we just kind of put it on the back burner until it became like this has potential because AI means more people want to deploy stuff, but it's also very expensive to deploy all your applications to Heroku or something, all very often um so that was kind of our approach was like well…

Chris Oliver (04:10) Where you have to pay for every single um you know dyno and database and background job and whatever for those we kind of offer a different approach where you know we set up your server you can run 10 side projects on one server and save money you're probably not deploying those applications…

Jason Cohen (04:37) You keep seeing we and our approach.

Jason Cohen (04:39) Who's we?

Chris Oliver (04:40) Uh I have one employee, Collin. Um so he joined me like…

Chris Oliver (04:45) Four or five years ago um doing screencasts and stuff like that. Uh so yeah, like Hatchbox has grown enough where I could live on it by myself. It's just getting close to that point where we can both live off of it, but be ramen and profitable and not…

Jason Cohen (05:02) Yeah. So just to be clear, besides screencasts, what does he do?

Chris Oliver (05:07) Um maintains the hosting Hatchbox service with me.

Jason Cohen (05:11) Oh, so he's your DevOps.

Chris Oliver (05:13) Um we both do it, yeah.

Jason Cohen (05:15) Yeah, okay. So you're both engineers and he does screencasts and who else does marketing or sales or something?

Chris Oliver (05:21) Us. The two of us, it.

Jason Cohen (05:22) Both of you do, okay.

Chris Oliver (05:23) Yep. Which we don't do enough. That's you know, we are developers first, so we have the tendency to just add new features or whatever instead of marketing and that sort of thing.

Jason Cohen (05:32) Okay, and so the answer to the original question, which is why don't you resell the infrastructure? The answer is because you had a tutorial where obviously you weren't selling infrastructure because it was a tutorial. And was it that you just never rethought whether the product was right? It was just a productized tutorial, or you decided after the fact, no, this is an advantage because as you just said, um, oh yeah, at Heroku dings you for every dyno or whatever, every application, let's just say. Um your idea is I won't ding you for every application and by you maintaining the infrastructure you've minimized costs and therefore you're targeting people who are extremely cost sensitive and also you get the least amount of money.

Jason Cohen (06:10) Is that the idea?

Chris Oliver (06:11) That's what happened, yeah. Um we do have some like you know, for example, Heroku has I think US and Europe is the two main regions. We've got people in Australia that want to use a local, you know, server host. Uh so we've been able to offer that or Hetzner or whatever offers really competitive pricing for servers. It's ended up where it's primarily, I think, hobbyists using it to deploy things. So we kind of have the lowest end of the market, which is, you know, not the best for pricing and everything like that.

Jason Cohen (06:46) Do hobbyists use Tuple or is that like you don't even do pair programming until you're serious developers at a real company?

Ben Orenstein (06:53) We have a pretty big range. We do have some like small teams of like one to two people.

Jason Cohen (06:58) One? How are you team of one and use Tuple?

Ben Orenstein (07:01) You know, I was actually telling this story to a friend the other day, which is like when we like our ideal customer is a freelancer with lots of clients and they'll pair with the clients and that'll be great for spreading it around. Uh and we'll get a lot of like marketing that way. And then as soon as we launched, people started signing up with teams of 10, 20, 30, and I like, Oh, yeah, obviously this is the customer. Uh and so the freelancer thing uh was always a small part of the business. Like most of the usage and revenue is on like big teams of, you know, a thousand plus developers.

Ben Orenstein (07:05) First started, for some reason the vision in my head was…

Jason Cohen (07:33) Yeah, so I mean when you just say it out loud of like, I think, okay,

Jason Cohen (07:36) Yeah, we're targeting hobbyists with no money…

Chris Oliver (07:40) Yep.

Jason Cohen (07:41) that's obvious, right? Like what is the discussion here?

Chris Oliver (07:45) Yeah, or I think number of like servers that our accounts have is like two servers, so I assume that some people are, you know, running a single app with production on one server and staging on another, but we have like half of our customers have one server. And then there's the long tail where there's I think our best customers, like 32 servers. So we do have some that are like agencies.

Jason Cohen (08:14) That's not the long tail. Long tail means lots of N. You're talking about the least of your customers are big.

Chris Oliver (08:20) Yes. Um yeah, like when you look at it, there's a handful that have maybe say five servers like in a load balance cluster and they've got more of a you know production traffic, but then we have agencies too that are a…

Jason Cohen (08:37) It's very funny to me that you've got the… Like you don't realize that the head are the people with five servers and 30 servers. That's the head.

Jason Cohen (08:38) Me that you've got the telescope turned around the other way.

Chris Oliver (08:47) Yeah.

Jason Cohen (08:48) Then you have a long tail of hobbyists that are probably barely profitable, very hard to acquire profitably because they don't give you any money. They quit a lot. That's the long tail.

Chris Oliver (09:00) Yeah, yeah, I was just looking at it from the graph of like, you know, number of accounts with this many servers and seeing like yes, there's one that has 32…

Jason Cohen (09:11) Focused on? Obviously you've been focused on the hobbyists.

Jason Cohen (09:14) We have to go get those other customers and we deserve to.

Ben Orenstein (09:18) Chris, do you want to go do that? My hunch is you haven't done it because you actually don't want to.

Chris Oliver (09:22) Mm, yeah. Yeah, I I've never really done like outbound sales kind of stuff. Um so I think I've just been afraid of it. I think I've not been comfortable, so I've just not done it. I think I would enjoy it if I did though. Like I enjoy the conversations with people and it just feels scary to send a bunch of cold emails or whatever and yet I know that like if I go to a conference uh about Rails or whatever, like I enjoy talking to everybody there about what they're working on and stuff.

Chris Oliver (09:54) So i in theory it's similar enough.

Ben Orenstein (09:58) Yeah, I mean you jumped up to outbound sales…

Ben Orenstein (10:00) Is maybe one way to go get a bunch of customers, but it's not the only way. Like at my high level view on this is like you had a um inbound thing that worked, which is you could make screencasts and educational things that you knew how to do and presumably liked doing, and that generated inbound for the hosting business, and it kinda all just worked for a number of years. And now that channel does not work. And so there needs to be a new one of some kind. And outbound sales is one option, I think.

Ben Orenstein (10:32) My guess is like in a world of increasing development by agents and increasing autonomy, outbound sales is maybe actually not gonna be the thing. Um, like the question that pops into my head is like why would my agent pick Hatchbox?

Chris Oliver (10:47) Mm-hmm.

Ben Orenstein (10:47) Um and I actually just asked it before this. Like I pulled up one of my many threads on one of my many applications that I actually do have deployed and said like hey should we use this? And it was like eh probably not. We're on Railway. I think it's 15 bucks a month total or 20 bucks for this app. This would be 15 just for the deployment piece, and then we'd still need a VPS. And then now we're taking on backups and it's all these things, although it makes it easier to do that. But basically it was like, no. Um and that is actually a hobby project for me. Like this is not for Tuple stuff, this is just like side hustle kind of things.

Ben Orenstein (11:20) Um so I do think the broader question is how do you get a new source of something, like new source of customers. Um, but then also it's a new world now, so things are weird. You probably would grow by going out and emailing people and doing the traditional thing. I think you probably would pick up some customers, probably more than you were getting now, but I wouldn't assume that that's like the answer.

Chris Oliver (11:42) Yeah, yeah, I agree. And that's maybe partly been why I haven't done it, because I am skeptical of like, is it gonna be worth the effort? Or whatever. Heroku's prices are still really high compared to, you know, Render railway.

Jason Cohen (11:56) You keep talking about high prices. I think this is part of the issue.

Chris Oliver (12:00) Mm-hmm.

Jason Cohen (12:00) Because if you keep fixating on the price, you're going to only talk to people who care about price. Those are not the people you want to pivot into. So I think what you need to start doing instead is ask, who are the people that still want their own infrastructure with some kind of tool to make that easier that are not price sensitive?

Chris Oliver (12:19) Mm.

Jason Cohen (12:19) That's a different Venn diagram, and it's the one you need.

Chris Oliver (12:20) Yep.

Jason Cohen (12:22) And so who is that? Have you thought about that yet?

Jason Cohen (12:25) Tell me about the very biggest customer with the 32 servers. What are they like? What kind of company is it? What are they doing?

Chris Oliver (12:31) They're an agency building apps for their clients… I talked to them recently and I think they have their clients sign up for infrastructure, but I bet it would be better if they could just, you know, bill and have margin on top and then they take care of everything and their clients don't have to know about,

Jason Cohen (12:49) Right.

Chris Oliver (12:49) you know, any of it.

Jason Cohen (12:51) And then who's the second biggest customer? What do they do?

Chris Oliver (12:54) Uh I don't know that off the top of my head, but I'm gonna guess that they're in a very similar boat. Cause any of our users that have 10 plus servers almost certainly are, you know, doing something similar, consulting or whatever.

Jason Cohen (13:07) And and why weren't you curious enough about wh who they are and what they're doing that you would want to talk to them?

Chris Oliver (13:12) Uh I am. I just haven't sent an email yet to everybody, so that's my

Jason Cohen (13:17) And why not?

Chris Oliver (13:19) Uh we just hadn't broken down our metrics and stuff for the longest…

Jason Cohen (13:23) Oh, you know who your second biggest customer is. You don't need a big metrics breakdown to know that.

Chris Oliver (13:28) Well, yes.

Jason Cohen (13:29) Yeah, so you could email them

Chris Oliver (13:31) Yes.

Jason Cohen (13:31) now, but you chose not to. What are you worried about?

Chris Oliver (13:35) Oh, that's a good question. We've just been worried about delivering, you know, top notch, reliable stuff. And so that's been our like, you know…

Jason Cohen (13:45) So what you obviously don't want to switch modes and sell.

Chris Oliver (13:50) I think I think it's yeah, just being stubborn and I know that I need to, but just comfortable doing the same thing.

Jason Cohen (13:58) You just don't want to. Like it's just not what you like. You like writing code and you don't like, you know.

Chris Oliver (14:02) Yeah, I mean that's probably definitely part of it. And but that's changing too is you know, now with AI it's like you don't have to look at a lot of the code. You want to review it and…

Jason Cohen (14:12) No, no, AI doesn't change whether you like to code or whether you like to dig into customers, find out who they are, and then change all your marketing. AI hasn't changed that.

Chris Oliver (14:20) Yeah. Well, I it's given me more time to do that, which has…

Jason Cohen (14:25) You already had the time to send one email to your second biggest customer. That's not what's happening.

Chris Oliver (14:31) Yeah, fair. Yeah.

Jason Cohen (14:33) This is the tough love part because by the way, I'm the same way. You know how Dharmesh Shah's are literally afraid of the phone? He's talked about this on stage at the HubSpot events in front of 20,000 people. So I'm not selling a secret here. He's afraid to pick up the phone, the CTO of HubSpot. Great founder, one of the greatest founders, right, that we all love. I'm the same way. The idea of like I have to do this is like I have to steel myself of this crap. I so don't want to do it. So I'm with ya. Of course it's better to be in Visual Studio and do something fun. I'm just trying to dig into like what is actually happening because these excuses are of course not the real answer but I don't blame you.

Jason Cohen (15:07) I mean uh you know, I'm just saying we're just trying to get to it.

Chris Oliver (15:10) Yeah, it's interesting because all I've done for the past like 12 years running this and I feel like it's been really hard in the last year to realize like I can't do that anymore. I have to change like the more important stuff is not writing code for fixing this stuff in the business, I've got to go do all of that. And I think have really gotten in the same habits of ah I feel most productive when I write code or ship a feature or fix a bug or whatever.

Chris Oliver (15:45) And spending my day talking to people doesn't feel as productive…

Chris Oliver (15:51) I've like I guess been I don't know, lucky's not the right word, but like I've just been able to make stuff and build stuff and then that turned into a business…

Jason Cohen (16:01) Not exactly what you've been able to do is make great educational content and that turned into a great business. And then as that market has shifted through no fault of yours, of course, you've realized, oh, okay, I have a product business. It's not a great business yet. And I don't mean that because if it's this size or has this many customers, that that's not what I mean. What I mean is the only reason we're talking today is because you're saying it's not growing the way I want. It's not yet a great business by your own definition, whatever that is. And it could just be what you said. I just want to be profitable enough that me and one employee make good money and I'm good.

Jason Cohen (16:34) And I would say awesome, I love it. Well then it's not there yet. And that's cause it's still kind of a hobby, like you were saying.

Ben Orenstein (16:44) I'm curious I don't know uh maybe I'm maybe this question's for Jason, I don't know. Um part of me is skeptical about the idea of saying to Chris like you should be a different kind of person. Like… And so you should just like tough it out and do it because you want this business to work. And so like you should have a day that maybe doesn't appeal to you that much. But suck it up, you're the founder, you gotta do it, otherwise your business might fail. That might be trueish. I wonder if there's like another path in here, which is like making things. You like making content and maybe code, it seems like. If in fact your ambitions are basically a business that supports roughly two people, maybe we can just find a thing where it's like you spend 90% of your time making content or code just differently than you are now, because it stopped working. And that generates enough inbound to this business that it works.

Ben Orenstein (16:54) You should be willing to like change your to me there's like change Chris or change the strategy, I think. Like we could say, yes, doing product development and customer calls and maybe outbound sales or something. Uh we think that generally works.

Jason Cohen (17:39) So I totally agree. I totally agree that in fact, oh look, I'm gonna no over here. I'm gonna but I'll push the book a little bit. Chapter 11 in there is exactly about this. You have to make a strategy that's in the shape of yourself. It's the only thing where you'll push through the hard times because it's you. You'll naturally be good at it just because this is what you are. It's what you're made to do, you might say. Um and when you're in that zone even when there's competitors who are well funded and blah, blah, blah, uh, you'll still be better at them at least in a certain way that some people will appreciate because it's you.

Jason Cohen (18:14) Um and so I 100% agree. So I do think it is useful to at least outline what let's call it option A, not because it's the best or first option, but just the option we talked about, option A, which is hey, you need to get people that pay more. That means you need to understand who is that, why is that, what are the characteristics of that kind of customer because it's like your top three and not like the other 200. And so what's that like? And so you have to discover. And then you have to target and yes, probably sell.

Jason Cohen (18:47) That doesn't necessarily mean you're banging the phone on outbound, but some kind of sales, some kind of conversations. And you said you like geeking out about the stuff. So to me, that sale is probably a geek out style sale where m you are geeking out to do the sale. That's okay. There there's a kind of sales that I used to do at Smart Bear, by the way, because I sold to Intuit, Intel, Microsoft, EA Games, Qualcomm. These were my customers. They're all enterprise customers, and they were buying enterprise-sized deals. Uh Adobe was a couple million dollars. Like these are enterprise-sized things for real.

Jason Cohen (19:19) And yet I'm doing the sales and I'm doing it like you're kind of like what I imagine m many engineers who learn just enough about sales to be dangerous, quote unquote, do, which is I mostly geek out. I make fun of normal salespeople. Like I'd go in and say, okay, I've got 40 slides and then I'll give you and then I'll slide into the demo. And you could see people going, Ugh. I was like, I'm just kidding, I have no slides at all. Um if you have any questions about the company, happy to answer. But let me just show you the demo. If this strikes your fancy, let's keep talking. And if you're like, I don't want any of this, then okay. Um that's the kind of way I would sell.

Jason Cohen (19:50) Just very honest, very geeky, um, but confident and you know, but the other part of that is sales is a project management job. Where are they in the buying process? You c stay on top of them, you're moving them through. Well, what else would it take to get you to finish? Oh, you're looking at this competitor. Let me talk about that. And you're like pressing and pushing and project managing it to a close. That's the difference between just geeking out, which is what you were talking about, and sales, even a geeky style sales, okay, but you still need that project management to get the sale. So it doesn't have to be pushy, doesn't have to be like used car salesman, doesn't have to be any of that stuff.

Jason Cohen (20:23) I am concerned even with other options, but we'll get back to it. Uh we don't have to dwell. We'll get back to it, but I'm concerned that like we do still need a market that has money that like if they're only gonna give you 20 bucks a month and the cancellation rate's really high, like can you do anything and make that work? Maybe. Like I'm not closing the door, I'm just saying that doesn't sound terrific. So I'm hoping that it's a mixture of a company you want to build and okay, maybe it's not huge companies, but like still it's something nice, like maybe agencies who build on behalf of clients, so that you're geeking out with the agency not trying to sell the end user.

Jason Cohen (20:57) So it's still geeking out, but in a better way. Also, another problem is it sounds like you hired another one of you, and that's tough. Because now you're not two puzzle pieces that connect, and now you're covering the things a business needs. Instead, you've overlapped and you still have a gap. Now, I'm not saying go fire this person. And they're probably gonna listen to this, and I'm… I'm just saying that's another choice, back to Ben's point. That's another strategic choice you've made that may limit what your strategies can be.

Jason Cohen (21:01) Um I'm hoping we can get to that. But let's get to what Ben said.

Chris Oliver (21:32) It made sense at the time with education being our primary business because everything I taught him was content to, you know, sell. So it was such a good fit there. This is a…

Chris Oliver (21:45) Different situation though.

Jason Cohen (21:47) So Ben, I mean this is Tuple, right? So you want to lead us into this option B, which is be yourself and whether that's PLG or whatever it is, like this feels more like Tuple territory, yeah?

Ben Orenstein (21:58) Uh yeah, I would say so. Um I mean I enjoyed some of the stuff that we're preaching to you about like doing customer development and talking to people and like we won our first 30,000 in business by like me getting pre commitments um on like calls of people. So we started with like doing the real sales thing and when the product started to work, um, it became a lot of uh yeah, just make the product better and uh because it was better than the other options, people referred their friends to it. Like this is why I like t selling to developers is because they do talk to other developers and they actually seem to like recommending things.

Ben Orenstein (22:34) So yeah, I don't know that helps you. I was thinking as we were just talking, I have some of the proclivities that we're sort of pushing you towards, but like if you don't, then who cares? Like if you hate your life. Whatever, like it's it doesn't matter if the strategy is good for you. Um but I I've been doing some digging while Jason was talking and I'm like, I was like, is there a content strategy that would lead to customers for Hatchbox? And I don't know that content is the path right now. Like when I talk to LLMs about hey, how do I deploy a hobby project?

Ben Orenstein (23:08) You don't come up. Um and it basically points me at like free open source tools on top of $5 VPSs if I want the super low end of the market and the railways and the renders and the flies for like the like ho Heroku experience. Um and so I have like, do we change Chris? Do we change Chris's strategy or like do we change the business? Like, I hope this doesn't just blow up the show, but like I wonder if this is like the business you wanna be in. If we found you a content strategy or like an inbound strategy that somehow matched your natural state, maybe we get to where you wanna go.

Ben Orenstein (23:41) But also wonder about this as a business in general, I guess, as like, is it good? Do you like it? Do you want to be here?

Chris Oliver (23:47) I've wondered that too, because yeah, the we're way closer to the free open source end of the spectrum, but we don't have the say Kubernetes expertise to go build an actual competitor to uh a Heroku, you know, like that requires way more DevOps knowledge and Kubernetes land than we have. Those companies have raised $100m or whatever to go build render and railway and stuff. So we're yeah, we're definitely in a weird spot. And I guess part of that is, you know, we looked at our products that we had built that worked for us in the past and this was the one that was the most stable of them.

Chris Oliver (24:24) And so we've just like, okay, I guess we should probably invest in it. Uh, 'cause the other ones are declining pretty fast. And so it's kind of like, you know, our life raft in a sense, but isn't necessarily like the right product, I guess. You…

Jason Cohen (24:38) Yeah, I mean the education and content is declining because what?

Chris Oliver (24:42) We just don't get any traffic…

Jason Cohen (24:43) No, I know, because why? Why aren't people looking at that anymore?

Chris Oliver (24:47) It's easier for you to go ask Claude to go write the code for you.

Jason Cohen (24:50) Right. That's what I thought you were gonna say. But can't you ask Claude to deploy Rails to AWS and configure AWS and like I mean also for the same reason can't it do Hatchbox already or close enough?

Chris Oliver (25:04) Probably close enough, yeah. There's a whole bunch of like a anything new it doesn't know, but we keep up to date on that. But it's a small value prop at that point, because it will be able to read the source code for whatever new stuff is there or whatever and catch up.

Jason Cohen (25:21) In what way is Hatchbox doing something that I can't do Claude?

Chris Oliver (25:35) It's the maintenance stuff I think is the biggest pain point for people. They like can have Claude get it set up, but then they gotta spend the time with Claude if something goes wrong.

Jason Cohen (25:44) Something like what?

Chris Oliver (25:45) Le the new update to Ruby or you know, a Rails security fix or something that.

Jason Cohen (25:52) And then they have to change their infrastructure because of a security fix?

Chris Oliver (25:55) Yeah, yeah. There's a big security ImageMagick or libvips uh security thing that just happened in Rails and so you had to upgrade your server to the latest version so you could have the latest libvips because it wasn't, you know, used securely.

Jason Cohen (26:09) When you say upgrade your server to the latest version, you mean like Linux or packages or something?

Chris Oliver (26:14) Yeah. Ubuntu Linux. There are things that happen like that fairly regularly that we can save time for people with, but it's probably not enough of a value add in that sense to make it a product worth paying for. I guess some people see it as well, it's 15 bucks a month… If it saves me an hour, you know, saves me plenty. But is there enough of those people out there? I don't think so. Probably not. And this gets easier and easier as LLMs get better. So our value prop probably continues shrinking over time.

Jason Cohen (26:47) Thing with LLMs is of course they can figure things out, but when it comes to precision, they're bad.

Chris Oliver (26:53) Fair, yeah.

Jason Cohen (26:53) Right? And so if the idea is, well with Hatchbox, there's all these details. Like you just said, a security patch comes out, then that day you need to know exactly what to do. Or you want Hatchbox to know exactly what to do, but you know, that day you need to do something very specific. That's a bad use case for an LLM. Because current stuff it's worse at? I mean you can go search around, but like it's worse, right? Even knowing something just happened that you needed to do? Like, you you're not asking every day is there a something right right? Like so just knowing that there is something current is valuable.

Jason Cohen (27:25) Doing it exactly right is valuable. So I can see that. I wonder, since people obviously are using AI more and more, I wonder if some of the value of Hatchbox is, given that you want to use AI to do these things, comma, but you need all this precision and currency, and you know, there's all these things that LLMs are actually bad at. What you want is your AI that talks Hatchbox to do the stuff, or also, Hatchbox doing some things on its own, like upgrading the servers or whatever, right? And it is that marriage that makes you all powerful.

Jason Cohen (27:58) The one reason I like this a lot is if my skills in Claude or my CLAUDE.md, etc. etc., you know, or whatever, right? Like scripts, skills etc. depends on Hatchbox. Then how do I rip you out?

Chris Oliver (28:13) Yeah…

Jason Cohen (28:13) I don't think I do. Like once you're in, I think you're in. I mean you could be ripped out, right? But like, I don't know, right? So does that mean MCP? Maybe a CLI for sure. Your own skills that connect to that API, blah blah. Now I don't know offhand whether that sounds good to a developer, but it's modern. It uses AI as a tailwind, a reason to buy Hatchbox, as opposed to AI being a disruptor of why I don't need Hatchbox anymore. And if all my stuff is configured around Hatchbox, it'll be hard to kick you out and retention should be high.

Jason Cohen (28:47) I'm also wondering how that strikes Ben who's not in the Ruby on Rails world as I am not, but does that sound more like something that works in the dev world nowadays?

Chris Oliver (28:55) Uh just for some context, one of our best customers, Josh Pigford's been using us over railway and stuff because now that we have an MCP, he's been deploying side projects and ideas and whatever. So he's really been promoting us as one of our, you know, best customers for that reason since we got the MCPs like

Jason Cohen (29:18) Mm-hmm.

Chris Oliver (29:19) pretty much wrapped it into his process like that. So that's a…

Jason Cohen (29:23) So criticizing you on what I'm about to say. I'm just pointing out on your homepage the big text says the Ruby PaaS that deploys on every git push. Now, to me, that's Heroku's byline.

Chris Oliver (29:36) Right.

Jason Cohen (29:37) Heroku… That's the whole exciting thing about Heroku from the very beginning is that their innovation was deployed on Gitpush, right? So first of all, that's not what's unique about you. Second of all, you're not a PaaS. Because still need the infrastructure part. You orchestrate infrastructure. That's not a PaaS anyway. It's kind of not what it is. And this new idea with AI, it's not there at all. The point is you are using AI for Rails. So here's an analogy. You use AI for Rails, but you don't use LLMs for GitHub because the diffs and checking it in and going to CICD that has to just be right. You don't use LLMs for that. That'd be the terrible idea. You don't use LLMs to actually manage files. Like that'd be really stupid. You have to use tools for that. So similarly for managing the deploys, doing staging, doing blue-greens, doing backups all this rails specific and even immediate stuff. You would never use AI for that. So Hatchbox is the way to use AI and yet all this stuff that requires precision or timeliness happens perfectly. Now what I just said is not what you'd put in an H1 because it's crazy. But like it's the it would be the message and it's like absolutely completely different than your current message, which is I'm Heroku, which you're also not. That message, I don't know, how does that strike you, Ben? Like GitHub is for version control, you need Hatchbox for deployment and stuff because so that it integrates with AI.

Jason Cohen (29:38) Is the Ruby PaaS that deploys every Git push. In fact…

Ben Orenstein (31:00) I think I like it better than the current positioning. I'm somewhat skeptical about this like need of like, oh you can't let an LLM manage deployments or something. Like it doesn't do that out of the box. Um or like we need some extra code that we've written around this. That seems probably not true to me. Mean I like the sound of what is Josh Pigford doing and why does he like it so much?

Jason Cohen (31:21) Mm-hmm.

Ben Orenstein (31:21) And is that a better lane than you're in right now? Um I guess it's probably yes, although I f knowing what I know about Josh, I think he i he is sort of one of those price sensitive people. He's like, I'm gonna make 50 apps, I don't wanna pay per app, so I'm gonna use cheap VPSs and put stuff on here. But um and when we get back to that question of like what's the content that would maybe convert your customers these days, is it like Josh Pigford live streams? Is it you doing the same thing which is like let's make 10 apps and like oh notice that like I don't need to through our MCP you can like buy a domain name and deploy the thing and set up backups and whatever and this is all just happening and I'm not doing any of it manually because I don't want to.

Ben Orenstein (31:58) Um uh could be at least, I think, perhaps a stronger position than you're in today.

Chris Oliver (32:04) Mm-hmm. Yeah. Yeah, I think that would definitely help for now, but not a big change in the long run too. Cause I guess you could probably make the same argument of you can use LLMs with render railway the same way, you know. So it's not anything we do that's unique or whatever that's…

Ben Orenstein (32:25) That feels like the problem. It's like you either need to like become dramatically more ambitious and try to do something way more full servicey, which I don't think you seem to want to that much. Or be like, we're gonna find the little niche inside this giant world and we're gonna be amazing at it because most people won't do it.

Jason Cohen (32:42) Yeah, I agree with that. A good general rule is in a in any business of any size, um if you're doing one big transformational thing, that's pretty much all you have time for. So if for example you say, okay, we're gonna change our posture of go-to-market and it's gonna be more sales oriented. That's already, and again, I'm not saying that's a good idea, but if you were to do that, that's gonna consume you.

Jason Cohen (33:08) And the idea of also completely r changing how the platform works, you just don't have the time. You said yourself, even if all you did was to platform change, you don't have the time to do it. So you certainly don't have time to do both, right? So, like you could totally transform the platform. But you better sell the way you normally sell, which is to say just product led growth or whatever. Right? Not really selling. Or you could just choose to sell the product that you have. But doing both, you definitely don't have time to do that. That's kind of what Ben was saying, or part of what Ben was saying just now. I agree that the idea of trying to do everything, not only are you not equipped to do that, meaning like you said, like with this team and funding and stuff like we're not gonna compete with companies that have been doing this for years we'll just be a worse one of those and they're not even that expensive.

Jason Cohen (33:52) So what's the point of… So that seems to me that seems like a non-starter if the idea is to compete with them on product. But another thing you could compete with them on is brand. Because if they're the big baddies, and Heroku's definitely the big baddie, and you're the, you know, support me. I'm the pirate ship. I'm the one you should do. I'm the one you've learned from half your life. Let's do this. No, I don't have features that they don't have. I have just the features you need, and it's me. It's Chris. Is that effective? Will that build the business you want?

Jason Cohen (34:24) Will that appeal outside of hobbyists? I don't know. I'm just saying it's a path where the product is not what you're distinguished on. It's this other feeling of why they're here. The Josh Pigfords of the world probably respond to that.

Chris Oliver (34:38) Mm-hmm.

Jason Cohen (34:38) That doesn't mean it's a good market, but they probably do.

Chris Oliver (34:41) And I feel like that's probably how we've gotten to where we are with it.

Jason Cohen (34:47) Mm-hmm.

Chris Oliver (34:48) And I just don't know how far that can go. And I my hunch is it's not that much farther.

Jason Cohen (34:56) Well, you say that so if I went to your YouTube channel, is it right now showing people how to use AI and Ruby on Rails and Hatchbox altogether?

Chris Oliver (35:04) Uh we've been doing that more. We've been making less content because nobody has been watching it because YouTube doesn't recommend it…

Jason Cohen (35:12) So if you were gonna execute this strategy, let's call this option C now, because option A is you go sell to agencies with what you have and people who want custom infrastructure. Option B is we're the AI Ruby on Rails where you're using Hatchbox for the precision and timely aspect. Although you said maybe the other ones are doing that too, but okay, that's still option B. So option C is it's all about you then the deal would be your root YouTube channel would be growing in viewers, even with LLMs. Because if you're showing them how to use AI to do this, then they are watching.

Jason Cohen (35:42) Like the idea that like people don't watch educational videos anymore because of AI, that's not true. You don't think people watch YouTube videos about how to use AI? They surely do. So that isn't true. They may not be looking at old style, but they are looking at content about programming with AI. Yes, they are. And so if you had this strategy of like, no no no, we're the ones who educated you and with reciprocity you're gonna use Hatchbox. If that's the strategy, which definitely leans into your strengths. No doubt.

Jason Cohen (36:14) Then what I would want to hear you say is, at least if you were to make this decision, okay, we're gonna do videos every other day. They're gonna be really good. They're gonna be on every hot topic of AI that ever happens. We're gonna show people how to do all kinds of stuff. So that again, uh we're gonna be like the channel for Ruby on Rails, how to Ruby on Rails and AI and all the latest stuff. And every time it's gonna be, and don't forget, you know, we're the ones that run Hatchbox. Uh know, the best way to support us is to um deploy through us. Everything you hear today, you know, everything in every one of these videos, that's what we support.

Jason Cohen (36:46) And uh let's go. That is a strategy for sure. That does lean into your strengths for sure.

Chris Oliver (36:51) Yeah. Yeah, that makes sense and could be a decent option. I don't feel like we need to grow to uh you know, have 50 employees or anything, so we'd be happy making a decent living and is staying a small team of two or whatever and that I think would serve that pretty well.

Ben Orenstein (37:10) You said like YouTube stopped recommending your videos to people and so no one's watching them. And I wonder about that. I clicked through some of your videos and just kinda like skimmed them briefly and they look a lot like they used to look.

Ben Orenstein (37:22) Like RailsCasts

Chris Oliver (37:24) Yep.

Ben Orenstein (37:24) kind of peak twenty eighteen ish, maybe even earlier, of like little guy's face in the corner, maybe and some code on the screen and like here's how to use with options. Here's how to use action cable 12.12. And like it's an eight minute video on how to do this thing. And I think that format is probably kinda dead. I think you stopped getting recommended videos because you weren't making things that people wanted to watch.

Chris Oliver (37:45) Yeah.

Ben Orenstein (37:46) I think the algorithm just noticed…

Ben Orenstein (37:47) That like this isn't it anymore. Sorry. Um and it used to be, uh, when like knowledge was hard to acquire and watching someone do it was valuable and we had our hands on the keyboards and needed to rip like you know how to use that stuff and it was useful to watch a demonstration, for example. And so I think there's probably a like my question is sort of for you is maybe like what actually do people watch now? Maybe some experimentation of like try 10 different things, none of which are like what you've been doing, and just see is there something that you like making that people do want to watch that kinda hits?

Ben Orenstein (38:21) Like of the 10, do two do pretty well? Like, oh interesting, that format actually now feels fresh and feels worth doing in this world. I'm almost positive it should match how people are writing code now. Not like in Vim, uh, but in a harness, right?

Chris Oliver (38:36) Yep.

Ben Orenstein (38:37) Um and so you could generate opinions about how to do this well and thoughts. Like people love talking about skills and like which harness are you using and why is that one better and like what tasks is fable better at than Astra and this kind of thing. And so like I think there is plenty of content bandwidth or something. Like people I think want to watch things along these lines. Like I would maybe go look at the top Twitch streamers perhaps. Like who's actually growing content wise that developers are watching and like what are they doing? And is there something kinda like that fits you really well that you would enjoy making?

Ben Orenstein (39:07) I think the algorithm is giving you a message. And you should listen to it.

Ben Orenstein (39:11) Is that how you're building these days, Chris?

Chris Oliver (39:13) Like what do you mean? Um…

Ben Orenstein (39:15) How do you ship code these days if you're making new features on Hatchbox? Is it with like a Claude Code or similar or…

Chris Oliver (39:19) Mm-hmm. yeah. We'll go through and, you know, basically plan it out like you know, if you're whatever it is, like help me find any of the edge cases I didn't think of. That's one of the most valuable things I think. Using L Ms, like all the stuff you didn't realize were gonna be a problem. Um, so we'll start there and, you know, have it go sketch it out and then review it, refine it, break it down. But yeah, been making videos of that process. We've been like, you know, doing the same thing we've always done is like showing the actual code that we produce,

Ben Orenstein (39:53) Mm-hmm.

Chris Oliver (39:53) which isn't the thing people are interested anymore in…

Ben Orenstein (39:56) Yeah, a little bit true.

Chris Oliver (39:58) Yep.

Ben Orenstein (39:58) This seems like a promising thing to explore for me on the content side. Which is like sort of just update yourself for the times and see if that works better. Um and then the other question I wrote down was what do developers hate doing right now? Because I think when you initially made Hatchbox, you were solving like an annoying problem. And the pitch was like, yeah, it's hard to do deployments on your own infrastructure and it's annoying to manage the backups and a few other things and provision SSL and whatnot. And so yeah, paying a little bit per month to have someone to do that for you is kind of nice and saves you some hassle.

Ben Orenstein (40:34) But I suspect that's really not true anymore. I suspect it's not actually a hassle. And so I think an interesting prompt f is to sort of think about what still is. Like what is a hassle and kind of annoying and doesn't sound fun to developers to like fix.

Chris Oliver (40:50) Mm-hmm…

Ben Orenstein (40:51) I think we kind of with Tuple got a little lucky that we ended up in this space, which is like developers find it kind of annoying to like share screens on Zoom and meet, and like it's annoying to pair program, it's hard to do remote control. Um and also it's kind of an annoying app to write and maintain because it has like deep integrations into the operating system and it uses real-time WebRTC and there's C involved and all this. I may be on Jason's show in a couple months because with every passing month the models get smarter and it becomes easier and easier to one shot make me Tuple. So this doesn't feel like a you know, don't I've mastered this or that this is a long term durable necessarily.

Ben Orenstein (41:23) But something where it's like it sounds like a slog. It is a slog. It's not a one-shot open source thing or like a one-shot private s solution for this away. It's maybe at least a couple days or a couple weeks um that maybe is like your niche for like niche dev tool useful thing that's like easier to just pay for than to do yourself and maintain and whatnot

Chris Oliver (41:45) Yeah. Yeah.

Ben Orenstein (41:46) uh…

Chris Oliver (41:47) I'm not sure if I have any good answers off the…

Ben Orenstein (41:49) Yeah, no me either. That's just meant as a little thing to stick in your brain. And mine too for when we are one-shottable.

Jason Cohen (41:56) I don't know

Jason Cohen (41:58) who is gonna one-shot something that's important to them that just works? Who's gonna one-shot stuff where it's 20 bucks a month, even a hundred bucks a month, and it works? And the answer is some people. And not most people. So if I have a little throwaway website, more and more people are one-shotting that. So if I was in the one-page website business, I'd be more worried. But the CMS people are not getting one-shotted. There's just too much stuff going on there. Um or for example in Tuple there's integrations to various IDEs.

Jason Cohen (42:29) There's the server-side stuff, there's all the latency stuff. You can try to one shot it in and we all know that the one-shot version of Tuple will be a very bad user experience. And trying to get that tuned is a job. The Josh Pigfords of the world say, but that's fun, and they do it anyway. Okay. But that's super rare. So I just don't believe in this whole like one shot means all software has no value. In fact, even now, uh with AI everything AI does, almost all SaaS stuff that's public is growing. Now I know their Wall Street valuations came down, but guess what? They've come back up. Because there was a little scare and then came back up.

Jason Cohen (43:01) But that's valuations. What about their revenue? Oh yeah, that's still going up 30% year over year, like clockwork. Okay. So where's this thing? Now you could say, but in six months this is one-shotta, and a year this is one-okay, that's fine. Um none of us know what's gonna happen in five years, so I don't see why that's an interesting conversation at all. I don't think in the next year any of this shit's getting one-shotted by almost anybody. So to me it's like just a non-worry. If two years ago I asked you what you should do, you'd say, I'm doing everything right. So that just… Two years ago. You still don't know. So I'm not worried about that.

Jason Cohen (43:31) Shows you don't know what's gonna happen in two years. You didn't know what was gonna happen…

Ben Orenstein (43:36) I'm not worried about like one shot exactly, but this value proposition of it's too annoying to create code and it's too annoying to maintain that code. And so if someone offers you a fairly cheap monthly cost to do it for you, it makes more sense to focus on your core mission and not get distracted. I think that premise has gotten a lot less true when I basically can do this more or less for free ish now. And so it's not like yeah, I if I were a professional content like blogger or whatever, if I was a giant WP Engine customer, I'm not gonna be like, make me WP Engine.

Ben Orenstein (44:08) Um but if I was reviewing my set of subscriptions and there's like a niche bootstrap SaaS that I'm a little money to, uh, I might be like, oh, I wonder if this is just uh a thing I could just basically one-shot these days. Um and I think there's just more and more things in that blast area.

Jason Cohen (44:22) For 20 bucks a month you can't one-shot things like, hey, there's a zero-day patch for Rails and we already solved that for…

Jason Cohen (44:28) For 20 bucks a month, you cannot one-shot that.

Ben Orenstein (44:33) I mean I could have Claude once a day look for all the zero days and patch things. I could have it do it every five minutes. Like it's…

Jason Cohen (44:37) You could do many things, but I'm saying almost no one wants to do that and then manage that. And then their security team says, so you patch all the zero day things? And you're like, yeah, I have something what running on my local machine to do that? And that that's not acceptable in a corporate environment. That has to go somewhere. Well, who's managing that? Do they actually want another process to monitor and manage and audit? No, for 20 bucks a month? A terrible trade-off. I would never want to do that. I'd always pay some security company to just do that.

Ben Orenstein (45:06) I think there's a blast radius.

Jason Cohen (45:07) That's why the way I put it was there are people for whom they're like, oh good, I like doing that anyways. All right. So that's a bad target market. But there's also most people who don't feel that way.

Ben Orenstein (45:18) Hm, mm-hmm.

Jason Cohen (45:18) So that's the good news, is the Venn diagram of people who don't feel that way is big. But so from your target market, you should probably select from that group. Because I that more and more people will feel that way. I'm saying those are the kind of people that are bad target markets for probably any SaaS company. Fortunately, it's small. Now will it be small in five years? I don't know. Let's talk about what we actually know. You know, right now it's really small.

Jason Cohen (45:43) There's very few Josh Pickfords. I love Josh Pickford just to be super clear. Love it. Right? But that's not even 1% of people like that. So from a market perspective, just not relevant. From a marketing perspective, highly relevant. Because Josh telling people to get Hatchbox is great. So marketing? Hell yeah. Market? No. No, I want the big company that has 200 Rails developers and they're like, oh buy this thing. And they're not worried about Heroku's cost. Right. One of the things that happened earlier is you were like, I don't know if the Rails market is decreasing in size.

Jason Cohen (46:17) That's a good question. Because to Ben's point, do you even want to be in this business? This calls the whole business into question, even though it is your strength and it is your expertise. So I went and looked at the Stack Overflow um survey, which you know, you could say whether how scientific it is and how biased it is, but it is tens of thousands of developers and they've done it for multiple years. So I went and pulled up 23, 24, and 25. The 26 uh one is still going on, so we don't know. And in 2023, 6.2% use Rails. And in 2025, 6.4% use Rails.

Jason Cohen (46:50) I would guess that's within error and whatever. So let's just say it's not going down. Plus, I know there's more engineers now than ever. A number of software engineers, in absolute numbers, obviously it's going up, yeah? There's not fewer software engineers now than there was two years ago, right? So if the percentage is, let's just say constant and the number that's a percentage of is growing. That means there's more Rails developers now than ever. Now, does that mean it's the healthiest place? No, like the most number of developers is JavaScript and Python and stuff like that.

Jason Cohen (47:24) Okay, you know… Fair enough. But then again you want to talk about being super saturated with stuff. And hard to get attention. Oh my God. So here you're in a niche which a decent number of developers in millions of developers are in. Millions. And that number does not seem to be going down to me. And you do have a special position in that world. So to me, that's pretty good. Especially with your goals. If your goal was to have 100 million in annual recurring revenue, then I would say, well, you know, you'd better dominate if you're gonna do that in real, right?

Jason Cohen (47:55) But that's not the goal. And so that's not important. And so I think for your goals the market is absolutely big and healthy enough. Just in sheer size. Like that we have to pick the right thing, but just in sheer size. Of like, is there enough people doing this? Yes, millions are doing it. Check plus.

Chris Oliver (48:10) Yep. Yeah, I think that's fair. Mm-hmm.

Jason Cohen (48:13) So that's nice that we can put that to bed. And we can say, yes, you should be in Rails. And to Ben's point, yes, you should be doing educational content. So actually, if you did start splitting this up a little bit between you and is it Collin? Is that what you said?

Chris Oliver (48:26) Collin, yep.

Jason Cohen (48:27) Right? Maybe Collin should be working on the tech and you should be doing nothing but this outward facing side. Education, the videos, what videos work, how do you know that? That means topics, that means style. Gonna have to try different things to find out. You need to put the videos out every day, every other day. Like you need to be doing this.

Chris Oliver (48:45) Mm-hmm.

Jason Cohen (48:45) Right? Because the more iterations, the better you get, the more you learn, blah da. How do you tie that into hatch Hatchbox? Interesting question. So all this stuff, and that's what you just get up and go to bed thinking about that.

Chris Oliver (48:57) Yeah.

Jason Cohen (48:58) Right?

Chris Oliver (48:58) Yeah, I think that makes a lot of sense. It's like everybody else has also been discouraged making content, so the volume of it has gone down overall. So then if I was actually to come back, full steam ahead,

Jason Cohen (49:13) Mm-hmm.

Chris Oliver (49:14) who else are you gonna watch? It's gotta be me if nobody else is doing it. So that makes a lot of sense to change the style and all that, but like reinvest when nobody else is doing it is also a unique thing we could do that, you know, hopefully would stand out more or whatever.

Jason Cohen (49:30) And just to be clear, you're not changing you, right? But you are changing things like format or topics and whatnot that you are putting yourself into. Because to Ben's point, if you try to change you, like now what's the point? But you're taking dozen years and your style and everything and keeping that part. Does that feel better to you, Ben? Is this more like what you were thinking or what?

Ben Orenstein (49:50) Yeah. I think that feels better. I think it's got a shot. I would still say I think the basic pipeline of like you have a large audience of developers who trust you and think you're good and that turns into customers on a product that you sell them is probably still a good like basic blueprint for a business like this. So restarting that first part, which is what worked before, and seeing if you can find that fit again seems promising to me.

Jason Cohen (50:12) Here's a really scary question. Would you make more money making these videos and sending people to Hatchbox? By making more money, I'm including all the work to maintain Hatchbox. Or would you make more money getting a very expensive sponsorship from one of these other competing products so that you're the number one Rails channel on YouTube and they sponsor you not just buying ads, but sponsor you in some sort of general sense?

Chris Oliver (50:35) Mm-hmm.

Jason Cohen (50:36) Which one would actually make more money? I mean today we know the answer because you're not the most popular channel, you know, you haven't actually executed this yet. And so perhaps this is a rhetorical question. But it's important because hosting products are really the good news of hosting products is retention is high because it's a big pain to move. Um, especially if it's low cost like yours, because there's not a big value in moving or one-shotting it, because it's not like you're saving that much money, you know. Um the bad news is it's a twenty-four-seven thing. Oh my god.

Jason Cohen (51:08) It just never ever ends. You know, of course now we have, you know, literally hundreds of people who can manage the stuff twenty four seven. But even when we had, say, 50 people, there was still like, well I guess I can't go to this birthday party tonight 'cause I have to stay sober because I'm the one on call in case something you know, like, it doesn't stop, right? And it's hard when there's only two people you can't spread that out very much. I know you know this 'cause you're living it. And yet you're not charging enough to pay for that because you're not making money off the infrastructure. So you still have this penalty of it's real time, is the way I used to call it, but you don't have the money that would go with supporting a real-time thing.

Jason Cohen (51:44) So when the Amazon server goes down, they're gonna call you, but you're not making money off that Amazon server. It's an asymmetric in the wrong direction kind of a proposition. Whereas if you were charging them as like a managed thing, which again, I'm not saying should because we already said why that's probably a bad idea, but then at least you're getting paid so you can afford to do things that are real time. Even if it's a small team. I mean at WP Engine, it was just me and then it was just a couple people. But we were charging enough to where I could hire people and so we could do it, you know, given a number of customers. I worry that asymmetry that you might even call it a financial asymmetry is a fundamental business model problem where you have taken on board a product that has this real time component, which is expensive and hard and so forth.

Jason Cohen (52:27) And not be getting the revenue needed to um support that.

Chris Oliver (52:31) Yeah.

Chris Oliver (52:32) Mm. Yep. Yeah, and people have suggested, you know, we should charge for support or whatever and it's like, we can, but then we still can't deliver we gotta sleep at some point and…

Jason Cohen (52:45) No, no, no, because if you charge for support, then what you do is you hire people in different time zones and now everyone sleeps at a normal time.

Chris Oliver (52:51) Yeah, it was just like, we'll need a fair bit of customers paying for support to afford people to do that,

Jason Cohen (52:58) Yes.

Chris Oliver (52:59) you know, which we uh probably don't have considering they're here for the cheapest option, you know.

Jason Cohen (53:05) Well, we're back to that. And yes, I agree.

Chris Oliver (53:07) Yep.

Jason Cohen (53:09) You can't escape the thing that says someone who has no money and never wants to spend money but and yet wants tech support 24-7 and also 24-7 uptime. You can't escape…

Chris Oliver (53:21) No.

Jason Cohen (53:21) You could escape it through huge scale, like shared hosts do. But that's not what you want to do. So given the constraints like that, which again I think are is a good constraint. I don't mean constraint in a negative way. I just mean it's a decision, which again, good. Very good. Every decision, some doors get closed and some open up wider. That's good. Gonna be small and I'm gonna charge people almost nothing, and they want in return… I would say that is a terrible strategy because that's just internally inconsistent. That's a bad business model.

Jason Cohen (53:42) Given that constraint…

Jason Cohen (53:48) 24-7 support, 24-7 uptime.

Chris Oliver (53:57) Yep, agreed.

Jason Cohen (53:58) And you're kind of there right now. So to me, something's gotta give. And so we've been exploring some of these options. You could sell to larger agencies or two agencies period until or the larger customers and actually do get paid and so they still have these high expectations but now you can afford to do it. That's option A. It doesn't necessarily match how you want to work. But at 20 bucks a month, they gotta stop asking for support and you can't answer the phone 24 center. Like to me, like you need to decide what the product is. That can't be the product for 20 bucks a month, it seems to me.

Chris Oliver (54:26) A lot of the complaints with railway or whatever is they're all very generic and they don't care about rails. And people are looking for something that really does and obviously we do, so the more we can go in that direction.

Jason Cohen (54:38) What do you mean railway doesn't care about rails? They're called railway.

Chris Oliver (54:42) Well, I don't know if that's related uh on purpose or not, but they're all very generic. You know, they'll all have their default stuff they do when they detect a Rails app, but they don't go above and beyond. Some of 'em, you know, Heroku won't support old Ruby versions and it's up to you to work around those things. And they're not…

Chris Oliver (55:03) Going over and above to tailor it to Rails…

Jason Cohen (55:06) I see. But you obviously do, right?

Chris Oliver (55:07) Which I think you're getting at we can, you know…

Jason Cohen (55:10) And is it clear? Maybe that needs to be clear too on the homepage.

Chris Oliver (55:13) Mm-hmm.

Chris Oliver (55:15) Yeah, I think that makes a lot of sense.

Jason Cohen (55:17) It definitely can be clear in your own videos, because you can just point that out. You can just point out like, well, by the way, you know, Railway doesn't even do this. They like they're too busy, you know, hosting Astro or something. So they're not, you know.

Chris Oliver (55:27) Yeah, and their bigger customer base will be other frameworks or whatever…

Jason Cohen (55:31) Can earn the right to throw shade in a quote unquote nice way on your videos, right? Yeah, yeah.

Chris Oliver (55:37) Mm-hmm.

Jason Cohen (55:38) Ben, I think you seem like your attitude is this is the best idea so far, but you're reticent still.

Ben Orenstein (55:45) Yeah.

Jason Cohen (55:46) Is that why? Is it because Rails? Is it because you don't think his heart's in it? Or like what is holding you back from going like hell yeah, this is totally it? Or is it just simply because Chris is like, yes, I could do that, and Chris doesn't seem so excited either.

Ben Orenstein (55:46) Oh.

Ben Orenstein (55:59) Um… I was looking at the Stripe projects page, which is a new product that Stripe made where you can provision uh services from the command line. Like Stripe projects add uh Railway, Stripe Project add Vercel. Which is a an interesting pain point, which is one I've had, which is like, oh, I need a uh database for this thing hosted somewhere or I want uh to be able to send text messages. And it's annoying for me to go manually sign up for Twilio and then put the API key over here and whatnot. So Stripe has this like nice answer for this. And I just sort of noticed that like the word deploy doesn't show up anywhere in the list of customers they have there so far or the services that they offer so far.

Ben Orenstein (56:35) They're offering third party companies in this list and there's like 60 of them. And none of them really look like Hatchbox exactly, which makes me kind of feel like you are not in the path of like the agent future currently. You are in the path of like person lives on and like signing up for stuff, entering credit card, whatever. Um and I see you have an MCP server, so that's like great, that feels like, you know, kinda bare bones, like at least make it agent easy for the agents to use you. But I'm pretty uh agent pilled as in like humans managing agents to build software at a fairly high level of abstraction is like probably true today and just gets more true. Uh and so anything where you're not on that happy path feels kind of wrong to me. And so I think possibly this like content strategy where you win developers mind share by showing them things and then they like think of you when they need to go sign up for a thing is like decent but like I'm not gonna try to predict five years but it doesn't feel to me like the future of software development most likely. And to me the question is more like why would agents pick you… And that's a hard question, but I that to me feels like the right one for like a longer term perspective on this.

Ben Orenstein (56:48) Manually thinking about what server this thing…

Jason Cohen (57:42) But surely there's agents that deploy things.

Ben Orenstein (57:45) Yeah. Mine do.

Jason Cohen (57:46) There's gotta be a way. I know that like OpenAI and I think Anthropic MCP libraries and there's been people on Twitter more and more saying, Oh my gosh, I put my thing into there and now I'm just getting new customers signing up. I mean paying customers every day like Clockwork, holy crap, this is the channel. So maybe it's Stripe and maybe it's not Stripe, but I it sounds like that's happening in some of these agent libraries.

Jason Cohen (57:50) Also have…

Ben Orenstein (58:10) I think that's right. Yeah. I think if you don't make a thing that agents want to use, you're probably not in dev infrastructure land, you're probably hosed.

Jason Cohen (58:18) Okay.

Ben Orenstein (58:18) Short to medium term is my sense of it. Um and so one question that came to mind was like what is annoying and painful about developing Hatchbox now, even wielding agents as best you can, that is maybe a pain point or something that like maybe there's a product in there potentially? Like I think a lot of times like the thing that you solve as part of solving something else can be an interesting product. But if you say like I'm gonna stick with the Hatchbox lane it feels like it's gotta be a thing that for some reason agents are like, oh in this case, this is the choice. I'm gonna do this. And I think it's a hard question to have a good answer…

Chris Oliver (58:51) Yeah.

Jason Cohen (58:51) How are you using agents today?

Chris Oliver (58:53) I'm using it to develop pretty much everything. Deploy this app to Hatchbox and it…

Jason Cohen (58:59) No, no, no. So my question is, how are you using agents today in your software development?

Chris Oliver (59:05) Building everything. Like…

Jason Cohen (59:07) When you say that, do you mean you're using Claude Code?

Chris Oliver (59:10) Mm, mhm.

Jason Cohen (59:11) Or do you mean you're building agents and then orchestrating the agents to build the code?

Chris Oliver (59:16) Um yeah, not orchestrating as much as I could be.

Jason Cohen (59:22) So name an agent that you're using right now to develop code.

Chris Oliver (59:26) We're mostly using Claude Code itself and…

Jason Cohen (59:29) Right. So name one agent that you're using to develop code right now.

Chris Oliver (59:34) Not really, yeah.

Jason Cohen (59:35) The answer is you're not using agents right now.

Ben Orenstein (59:37) Wait what? How is Claude Code not an agent? Isn't that an agent?

Jason Cohen (59:41) Uh no, it is not.

Ben Orenstein (59:42) Okay.

Jason Cohen (59:43) It's a harness that can do things like invoke agents in parallel. Agents, in fact, one of their defining characteristics is they can't launch more agents.

Ben Orenstein (59:51) Okay.

Jason Cohen (59:51) Like if you go into Claude Code, you have a directory called agents. You have a directory called skills. The things that are in the directory called agents cannot launch agents in parallel. It's one of their things because the harness doesn't let them do that. Now if you want to call the whole thing an agent, okay fine. But my point is it's not what you were saying, Ben. You're saying what do agents want? You didn't mean what does the Claude Code harness want. You meant what you said, right? Which is I'm orchestrating agents to do things, yes?

Ben Orenstein (01:00:15) I think of Claude Code as the harness and inside is an agent called Claude. Um and what does Claude want to reach for when Claude wants to solve problems? Is my question.

Jason Cohen (01:00:25) So I what I'm getting at is if the new strategy is to be the thing that agents want to use and to be agent first or something like that, I mean that's maybe putting it too strongly, but maybe it's not. One step is you've got to be that yourself. Because quote unquote you're the customer, but just because how do you understand any of this crap unless you're using it? I mean, like I live in Claude Code. Every time I see something on Codex, I don't even look at it. I don't have time like that's another thing to have to learn. And I'm sure it's better than Claude I don't know. So in that same vein it's like, okay, so if the idea is it's all AI all the time, it's being found by agents, like that's what you've got to live to stay on top of what the hell it is.

Jason Cohen (01:01:04) So again, this is not a criticism. I'm just saying like if the strategy is this is about AI. Everything here is about AI. It's gonna be interfacing with us directly with AI. Agents are then interfacing by AI. How to market to agents and the other half is how to market to people. Like so if that's the attitude, okay, then like I guess also getting used to what it means to dev with agents. And so when you see videos of people going like I have five agents going at all times doing things and this is my code review agent and this is my this my deploy agent which is what it means to develop with agents.

Jason Cohen (01:01:36) Um because they're all specialized in doing stuff and in parallel and you know maybe Claude Code is orchestrating it or maybe eight N is orchestrating it and a CI C D whatever. That becomes your new world you don't have to be an expert tomorrow, but like that becomes your world that also on your own engineering side, you do so that you're completely steeped in it. So that when you say, yes, Hatchbox is gonna be correct in this world, you start understanding what that world is. Does that sound exciting to do? Brave New World of AI and agents?

Chris Oliver (01:02:06) Yeah, I mean this stuff's super fun. Like the amount of tedious stuff that we used to do now becomes you don't even think about it, 'cause anything that comes up that you're like, Oh, there's this stupid bug and I know it's gonna be annoying to fix, spin off something to go do it, you know…

Jason Cohen (01:02:21) Yeah, I know why AI is good for coding. Do you want to pivot this whole company that way…

Jason Cohen (01:02:27) Or not?

Chris Oliver (01:02:28) Hmm. I mean, it makes sense and it's a…

Jason Cohen (01:02:30) I know it makes sense. Do you want to do it?

Chris Oliver (01:02:33) I think so, yeah. I mean it's a blast. Like we've just had a great time doing it ourselves and I guess, you know, sharing how we do it, all the direction stuff, keeping up to date's hard, but you know, that is a challenge. Um, so that part sounds like frustrating or whatever, but it totally could be a blast to go pivot to really embracing that… I think it all comes back to like Hatchbox itself as the product is such a strange place in the market that I still wonder, you know, if you'd be better off doing that with Railway, render, whatever.

Chris Oliver (01:03:14) And then maybe the sponsorship things, a better direction for it, where it's like we still make all the AI content, but use a better platform underneath instead of our own do that. I don't know.

Jason Cohen (01:03:24) Now I'm confused because few minutes ago you told me about how they're bad for Rails because Rails is just one thing of many and you're better. So what do you mean they're better?

Chris Oliver (01:03:32) But well, in the sense that not for Rails…

Jason Cohen (01:03:35) Rails is the only sense that matters.

Chris Oliver (01:03:37) Okay, well, that's what I'm getting at. It was like if we ignore the servers and our constraints and only focused on Rails, if we found a partner that wanted to embrace that, then we could make the content for it and help them do the polish or whatever, build that externally. Um, potentially that's a good way to do it as well. An option D or whatever that's kind of similar but isn't built on our own software product there potentially.

Chris Oliver (01:03:59) To grow as a…

Jason Cohen (01:04:10) So option D is become an influencer.

Chris Oliver (01:04:13) I guess.

Jason Cohen (01:04:15) Yeah, that's an option.

Chris Oliver (01:04:15) Could be, yeah. I mean it's similar to where we started with education and stuff.

Jason Cohen (01:04:20) It's the same. It's just you're you would have pivoted into what people want to see now

Chris Oliver (01:04:24) Mm-hmm.

Jason Cohen (01:04:24) and how people make money doing that now.

Chris Oliver (01:04:27) Yeah. Yeah. It's not getting paid directly like it used to be. Sponsorships or whatever make or the better business model currently at least.

Jason Cohen (01:04:38) You'd say kissy bye bye to Collin and all your customers. You'd send your customers to wherever. And uh But maybe you get a sponsorship for X dollars and a little bit more money from YouTube and that's what you'd do.

Chris Oliver (01:04:49) Yeah. Yeah, I don't know 'cause I like building stuff, so I wouldn't feel as fulfilled if I lost that. 'cause that was how I felt when I was just making videos. I needed something to fuel, needed to build stuff to fuel the videos. If I didn't have that, it just didn't work. I burned out. So yeah, in that sense it can't be the option. It could be part of the solution for me…

Jason Cohen (01:05:13) I still think one of the unsolved things is the business model is currently the costs or expectations of customers and what they pay for is incorrect.

Jason Cohen (01:05:22) I think you need to solve that. You either need to not get the phone call when anything breaks and still charge the low amount, or you do charge enough to where you can afford to do that. You claim, uh, well, I couldn't afford to hire those people, but I would claim you're already doing it. You're already fulfilling that promise. You're just not charging for it. Why don't you charge for it? So yes, charge for support, yes, charge for like twenty four seven, you know, response time. Otherwise, it's a uh you know one business day response time. And so the $20 dollar a month gets you a 48 business hour response time, and the $100 a month gets you a 24-hour response time, and the $1k a month, I'm just making that up, but you know, whatever, is a real-time response time type of thing.

Jason Cohen (01:06:05) And segment your customers. How much you actually try to get the bigger ones and in what way, I don't know. But the ones that do need to be giving you enough money that it makes sense.

Chris Oliver (01:06:15) Mhm.

Jason Cohen (01:06:16) And you could say, I don't want anyone to do that. Okay, fine, then don't do that for anybody. Stop it You know, like you've got to decide like to me this is a broken uh business model financially. So you gotta fix that somehow.

Chris Oliver (01:06:28) Yep, yep, agreed.

Jason Cohen (01:06:31) Ben any final ideas here or thoughts?

Ben Orenstein (01:06:35) Have you talked to Aaron Francis about dev marketing stuff recently?

Chris Oliver (01:06:38) Not recently.

Ben Orenstein (01:06:39) 'Cause I he's my North Star on that kind of thing. I think he does it better than pretty much anybody.

Chris Oliver (01:06:43) He's phenomenal, yeah.

Ben Orenstein (01:06:44) And so if you haven't had a conversation and picked his brain about this stuff, I think it's worth it.

Chris Oliver (01:06:48) I will do that, yeah.

Jason Cohen (01:06:50) Nice.

Chris Oliver (01:06:52) Yeah.

Jason Cohen (01:06:53) Okay, uh I think we said it all.

Chris Oliver (01:06:56) I appreciate it.

Jason Cohen (01:06:57) You probably need time to think and also talk to Collin about it?

Chris Oliver (01:07:01) Yeah.

Jason Cohen (01:07:02) What's your heart telling you right now?

Chris Oliver (01:07:04) You know, I think part of me wants to tough it out, f learn, figure out how we can fix things and…

Jason Cohen (01:07:13) What does tough it out mean?

Chris Oliver (01:07:14) Just to figure out, okay…

Chris Oliver (01:07:15) How do we fix the business model if we did try to do that. Um you know. Uh part of me wants to see if I have what it takes to do that. Um…

Jason Cohen (01:07:25) What is that?

Chris Oliver (01:07:27) To figure out can we turn it into something…

Jason Cohen (01:07:31) I don't know what that means. We had four options of what you could turn it into.

Chris Oliver (01:07:34) Yes. Well, in the sense of like options A and B, like do I want to continue with the product given all its kind of flaws and everything? Or do I want to go really down a different direction? I guess is the thing most on my mind of is this really worth fighting through to figure it out, or should we, you know, say it was a good try and it got decently far, but is never gonna be really to the point we want.

Chris Oliver (01:08:07) Um…

Jason Cohen (01:08:07) If you were making 10 times the revenue with the same customers

Chris Oliver (01:08:11) Yeah.

Jason Cohen (01:08:11) right now,

Chris Oliver (01:08:12) Yeah.

Jason Cohen (01:08:13) would you still be asking whether the product's good or not, or would you be happy?

Chris Oliver (01:08:17) I'd be happy with it…

Jason Cohen (01:08:19) All right. So rather than just shutting it down, why don't you change your prices? And the worst case is they all leave and then you got your answer. But the other case is they don't all leave and now it's the situation you just said.

Chris Oliver (01:08:30) Yeah.

Jason Cohen (01:08:31) So in other words, you say either this is the business model because it would actually work, or if that won't work, then the business model is simply broken. I'm not interested. That's what you could do if that's how you really felt.

Chris Oliver (01:08:41) Yeah. Yeah, and I think we need to decide if

Jason Cohen (01:08:45) Yeah.

Chris Oliver (01:08:45) what that looks like and if we want to try it or whatever. Cause yeah, I feel like we're just at the crossroads of all right, should we just try something new completely? We've been at this for a long time it's really hard and doesn't quite line up 'cause we just don't have uh, you know, the right product and the right market and whatever. Um, all the stuff we discussed. So yeah, it's that's kind of where I'm at on it. It's like deciding does it make sense to do this for another year we would definitely need to know uh let's raise the prices and see let's test it and if it doesn't go well then there's our answer or whatever.

Ben Orenstein (01:09:31) Sounds fun. I dare you to do it.

Open the transcript file

Hosts and Guests

Nobody is credited on this episode yet.

Credit the people on this episode

Be the first to rate this episode

More episodes

All episodes
  1. E13 w/ Ruben Gamez: How to Restart Growth in a Stagnant SaaS Business

    Nov 8, 20261 hr 9 minAudio

    Michael Koper says Nusii, his proposal software, is simpler than its rivals, but every one of those rivals say that too. Revenue has been flat for a few years now at around $10k per month. How does he stand out? Buyers...

    Credit the people on this episode

  2. E12 w/ Tibo Maker: How can a $40K MRR SaaS be losing 20% of Its Customers

    Nov 1, 20261 hr 21 minAudio

    20% of Post Bridge's customers leave every month -- massive churn that puts a hard ceiling on growth. Jack Friks’s tool sends posts to many social platforms from one dashboard, and he built it in public on X, where his...

    Credit the people on this episode

  3. E11 w/ Mariya Delano: How to Fix a SaaS Trial-to-Paid Conversion Problem

    Oct 24, 20261 hr 12 minAudio

    About half of Userlist’s paying customers never fully set up the email automation they paid for. Jane Portman and Benedikt Deicke built Userlist over nine years for sophisticated SaaS marketers, but its onboarding is...

    Credit the people on this episode

  4. E10 w/ Jessica Malnik: Plateaued at six figures, but still hungry

    Oct 18, 20261 hr 11 minAudio

    Brett Lindenberg’s startup Mindstamp puts six figures of profit in his pocket every year, helping businesses create interactive videos for training and sales. But growth has leveled off, and he’s having trouble changing...

    Credit the people on this episode

  5. E09 w/ Jon Sturgeon: AI changed his agency's model. Should he grow or sell?

    Oct 4, 20261 hr 54 minAudio

    Mario Peshev reduced his agency from 60 to 30 employees thanks to his own AI innovations. But almost nobody there is working on landing new clients. Pouring money into sales means building a machine he has never had. He...

    Credit the people on this episode