Rebuilding a SaaS
===

Brian: [00:00:00] You're about to hear my conversation with Ian Landsman, where we'll talk about rebuilding a SaaS from the ground up, even after over 10 years in business, the why, what, and how we'll get into it.

So I'm about to roll the interview with my friend, Ian Lansman. We recorded this one on November 3rd, 2023. And in this conversation, we talk all about Ian's business at userscape and they are building HelpSpot. Well, they've built it. They've been running it for over 10 years, but this year, as we get into 2024, they are embarking on a full rebuild.

I want them to understand the story there. It's a good one. Here's my conversation with Ian. Enjoy.

Ian returns!
---

Brian: Ian Lansman, you're back on the show. Good to connect with

you again, man. Yeah,

Track 1: back, baby.

Brian: the whole show is back. It's, it's been on hiatus for about a year and and we're bringing it back. I think your, your episode might be the first one [00:01:00] back, so, yeah.

Ian : Oh, wow. I I didn't that. That's

Brian: recorded one yesterday with with Adam Gilbert founder of of My Body Tutor, which is a super, super interesting business. Yeah.

Track 1: interesting.

Brian: It's so,

it's so cool. 'cause like he, yeah, and like, I, I just met him for the first time yesterday and we're like on the podcast and we realized that we live like 20 minutes away from each other in Connecticut.

Track 1: Isn't that funny how that happens?

Brian: amazing.

Track 1: it's crazy.

Brian: So, so Ian, I mean, you've been you've been all over the internet for well over a decade at this point.

Track 1: Two

Brian: two dec.

Oh man. Two decades.

Holy

Track 1: Crazy old.

Rebuild? Redesign? Relaunch? What is this?
---

Brian: Laravel extraordinaire. We'll, we'll get into the whole Laravel ecosystem, I think in the, in the next one here. But, you know, you're the founder of, of user scape and which, which makes HelpSpot and I thought we would dive into that. 'cause as I understand it, you're going through, what do you, what do you call this?

Is this a rebuild, a relaunch redesign?[00:02:00]

Track 1: Yeah, we're doing a full rebuild ground up. And, uh, it's still pretty early, so I've been tinkering around with this for already almost a year. But it was been much more like experimental, like kind of prototyping different things and just learning, you know, just getting caught up on technology. I definitely know really well the Laravel world, but then there's any more modern app, has a lot more JavaScript in it and things like that.

So I've been getting up to speed on what's going on over there and just kind of planning out what we want to do. So, yeah, so kind of now or into like actually constructing.

What will become the next version?

Is this a SaaS?
---

Brian: Awesome. All right. I wanna unpack this a little bit, like especially around the why and the strategy behind actually rebuilding. But can you give us like a quick recap of, of the history, like when, when did HelpSpot actually initially start? When did it launch?

Track 1: Yeah, so it launched in 2005, so we're coming up on 20 year anniversary here in 2025, which will be

Brian: Yeah.

Track 1: crazy.

Brian: And it started

like not even as a [00:03:00] SaaS, right? It was always on

Track 1: Right. Yeah. And I mean, even today it's sort of not even a SaaS. We do offer a SaaS version, but basically we just take the on-premise version and host it for people. So you get your own virtual machine, um, you have basically your own stack.

There is a shared database, but each customer has their own database on that database server. So it's not at all built like a regular SaaS app with a single monolithic database with tenant IDs and things like that. Which has its pluses and minuses, like we're sort of reaching some constraints there, which is part of what we're doing with reworking it.

But, but it does have a lot of advantages too. Like people can move back and forth between being on premise and hosted and,

You know, different accounts can get different levels of resources and things like that if we need to. And so there are some sort of interesting flexibility items there.

But

but yeah, started on premise. It was on premise for the first like 10 [00:04:00] years or so. And then we did a, and we had a partner who would host it for people, like on an old school, like PHP hosting platform. So we would just send people over to there. But then obviously the expectations became that you could just,

Brian: So

like,

Track 1: know, get it from the provider.

Brian: so, so an organization like, I don't know, let's say like it's a university or something like

that. They come like, at least in the old model or I guess still today, right? You

sell it this way with on, so, so they go through

your sales process. You, you sell them on, on implementing HelpSpot for their organization.

Track 1: Mm-Hmm.

Brian: They then need to engage their own hosting provider.

Track 1: Well, so how it worked today, like just staying in the modern era, you have a choice between hosting it yourself and or having us host it. So if we are hosting it, it's from the end user's perspective, it's like a SaaS. It's like any other SaaS. Like they just sign up on our website, they get in a few minutes, they get to log into their [00:05:00] HubSpot and they can start using it.

If you go on premise, then . Most of the time at this point, what you're really talking about is a company that has their own IT resources and infrastructure. And so that they are

Brian: So like they have their own

Track 1: install it themselves, right? They have their own servers. Now, their own servers can mean a lot of different things at this point.

Like sometimes it means they physically have servers. Sometimes it means they just have their own AWS account and they're still gonna host it on AWS just like we hosted on AWS. But, you know, within their own network on AWS

Brian: They control it. Yep.

Track 1: right? They control when they update it, they control how they back it up, they control all that stuff about it.

So, so that's usually what we're looking at. It's like, so at this point, Mo, you know, three quarters of the customers, new customers are SaaS and a quarter are on premise.

Brian: Got it. And when you did

the SaaS, have you, so when did you introduce this, the, the hosted option? Was that like

Track 1: I think it [00:06:00] was like eight or nine years ago, something like that.

Brian: Okay, got it. And, and so the SaaS version of HelpSpot, so, so you've always had the on-premise version that you're just describing,

and then the

SaaS version. Have you, have you treated that as a sep totally separate product, or is it like literally the same code base except it's hosted by you versus hosted by them?

Track 1: right? It's literally the same product. All the features are the same. Everything's the same. I mean, there are a couple slight differences, but effectively it's identical. And it's just a matter of if we're hosting it for you

or if or if you're gonna go and host it yourself on your own infrastructure.

Is this the first rebuild?
---

Brian: So it, we're recording this at the end of 2023. You are getting into a whole rebuild. Ha. Has there ever been a rebuild like this before? Like, is this version

Track 1: Not on, no, there's never been a full rebuild. This is this would be the first full rebuild. We've had some very big

Brian: like refactorings refreshing?

Track 1: refactorings. Yeah. So like we. mOved it onto [00:07:00] Laravel. At one point. That was a big one, even though now it's still not like a full Laravel app. It's like Laravel is in there and it's sort of the core, but there's still like code libraries and different things from, I mean, literally 18 years ago some of them.

Brian: it still works.

Track 1: yeah, it still works and this is why they tell you not to rebuild. Right. And yeah, so, and then another big one was when, you know, initially like UTF eight wasn't even like a thing. Like the databases didn't even like support it back at the beginning. So at one point we had a big update that moved the whole system on to be UTF eight compatible.

So that was a huge thing. 'cause we had to like literally move every piece of everyone's data. Like we basically had to rebuild their database. So even though all the code wasn't different, there was a lot of code changes. And then also there was this whole database migration. Thing that was a monster.

And yeah, so that was a big one. And we've done, we do that a lot. Like we've also done a ton of database migration as, 'cause we, on-premise customers who want to [00:08:00] go to the cloud you know, we convert them. And so,

Brian: Hmm.

Track 1: you know, if they're on MySQL already, then that's a little bit easier. But some of them are on Microsoft SQL Server,

Brian: Yeah.

Track 1: so we have to actually change it to be my SQL compatible and then import that.

So there's a few steps

Brian: So what led to the D? What, when did you decide to embark on, on a rebuild? Like when did this whole process sort of

Track 1: Um,

Brian: happen

Track 1: well, I mean I've gone through different phases with it. So about, uh, 10 or 12 years ago now I hired Taylor, who is the creator of Laravel, and basically I hired him to do the rebuild. Like that's what we were gonna do and.

Brian: years later? If we're g

We're gonna get started,

Track 1: 12 years later and we're ready to get started. He does, obviously hasn't worked here in nine years or whatever, so, yeah, so I mean, it's been around for a while.

But [00:09:00] you know, it's, it's always that trade off of obviously to go into the tank and totally rebuild. It means you're not adding as many features to the existing version. Um, and it's just a big undertaking. It's confusing for customers, all that kind of stuff. So, you know, it's like we get started on it and it's like, oh, is this the right time?

Blah, blah. Is there enough upside to it

Why rebuild? Why now?
---

Brian: I mean, that's what I wanna get into is like, all right, like before we get into the actual, like what you're building and how you plan

to roll it out and all that, but like, just the initial decision to like, okay, we, this, this, this, code base, this product has been serving us well for, you know, two decades.

Track 1: Right.

Brian: you know, why now what, what is the core problem that a rebuild should solve? What's, what, what

changed? You know?

Track 1: Yep. So I think there's a couple things here that are interesting. So [00:10:00] the initial like 10, 12 years ago when I wanted to rebuild it, right? That was, there was not any good business case to rebuild it really. That was just my desire to rebuild a case, which is like, I know so much more now. Like, I mean, version one of HubSpot, I'm like there with JavaScript Bible on my lap.

Like I don't even know how to do any of this. The code, there was no frameworks at all. So all the PHP is like my own framework essentially, right?

Brian: But I mean,

I, I feel like there's some, there are some exceptions to, to that you know, so many of us, myself included, lean toward like, just don't rebuild just because you want to.

But I think sometimes rebuilding, just to use some newer tech or a better design internally can add a bunch of speed going forward.

Like it's slow right now

while we rebuild, but the purpose is like, this

will two x or 10 x hour speed to ship going forward,

Track 1: Yeah. And I think there's even other, there's just your own satisfaction with working on it. Like this is a product you're [00:11:00] gonna work on for two decades. Like if, if it's a more fun environment to work in, you're probably gonna be happier. And then also you will ship faster and better

Brian: Attract happier team members and all that.

Track 1: right.

All that stuff. Right. So I think that that's, and you can leverage just easy, easier, more modern testing and whatever. There's like tons of benefits when you modernize. So I definitely think, and so what we've ended up doing is kind of middle grounding it where like, okay, like we. Moved it basically onto Laravel, but we didn't like, totally redo every line of code or the whole interface and things like that.

And then obviously like, I mean, it's been continuously maintained for 20 years, so it's like somebody's been, you know, there's, every month there's a release basically with bug fixes and features and whatever. So it's not like we've been just sitting around being like, oh, well we can't update it onto the newest stuff, so we're just not gonna do anything like, you know, we just do it in, you know, slightly different way than maybe a modern app would, would [00:12:00] operate in terms of a started Today app.

But in general it's been, you know, fine. And so that's what kept putting me off on it, is like, oh, like it's, there is still risk there. 'cause ultimately once you rebuild it, like, I definitely wouldn't want to rebuild it in the sense of like, just feature for feature parody.

Brian: that was another gonna be another big question is Like, Yeah.

like the benefit of not rebuilding it and maybe just doing like a front end re refactoring or something like that is like, you don't have to rebuild all the features. Um

Track 1: Right. And we've done the frontend factors, like we've probably had three or four big frontend factors over the years. Right. And that, so to me, part of it is like, if we're gonna rebuild it, then we also have to take all the other stuff we've learned. So like, yeah, we've learned about the tech, but we've also learned about how people use this tool.

And if I was building this tool again, like what would be different? What doesn't make sense the way it works now,

Brian: Yeah, like has there

been a change in the

customer base or the market,[00:13:00]

the positioning like,

Track 1: support is so much different than it was 20 years ago, right? Like now you have the emergence of ai, how does that fit in? You have widgets that are like super powerful and people want to embed them in their apps and.

Have it, you know, whether it's a chat or whether it's showing you knowledge based articles or whatever. Like there's just a lot of different functions that weren't really a thing, you know, 20 years ago. And then it's sort of tricky though too because there is this balance, because a lot of our customers, like the temptation is, okay, like an inner com style widget, right?

Like it's super advanced and it does all kinds of stuff, right? Because it's like you drop it in your SaaS app and boom, you have this whole

communication system to your, to your customers. But if we look at like HubSpot's customer base, there's actually not as many uses for that in our existing customer base.

Like most of our university customers, like we've been talking about a little bit, like they're not necessarily have a website that they might want live chat, but they don't necessarily need the like [00:14:00] Uber super messaging pushing messages into something

Brian: detect what you're trying to search for, all that

Track 1: Right. Like there's different levels there of like how much sophistication they would even want, like, um, and be willing to pay for

and things like that.

Brian: I also happen to think even for like small SaaS companies, like, like mine, like I also think that, and, and as a user of SaaS, like I think that those, those help widgets are overblown. Like they, they've gotten

to, I use them and I have one on,

on mine. But like, I, I

don't want all the advanced features. like I literally

just want a, a, a form to send a message or a simple

search box to find the article.

That's it. That's

Track 1: right. That's, and so that's kind of like my, even with this rebuilt HubSpot, like there's, it's going to be more along those lines. Like it's definitely not going to be all the way to the crazy stuff you can do with like an intercom, because at the end of the day, a lot of that stuff is, there's also this whole premise, even with the AI stuff, which like, there's definitely a lot of possibilities there for, but at the [00:15:00] same time it's like.

A lot of it is still just like the way it works now, and I think in the future it'll be different, but the way it all works now, it's still gonna be basically the way it's worked in the past for like things like chat boxes where like you have the vector database. I don't know how much people know about this, but basically there's this way you could kind of stupidly look up chunks of information and then you could feed those chunks of information to cha

GBT and it will make it sound nice.

So it will obviously be written much better than it was a year ago. If you built this, it would be kind of clunky,

right, in terms

Brian: My

Track 1: But it's still that same clunky lookup scheme, so that's where the, the answers aren't necessarily getting better per se,

when you get down

Brian: that's the thing I, I.

Track 1: specific product,

Brian: I mean,

AI in customer support tools
---

Brian: all right. My hot take sort of like in the making right now as I, as I speak, is I think that AI in customer support interactions is [00:16:00] not even close to being impressive there yet. Like, and I'm a big, I'm, I'm,

I'm super into ai. I love it. I use it every day in, in my

day-to-Day work. But

in the customer support space, maybe there's stuff I haven't seen yet, but a user, I don't want to interact with an ai. Like

if I'm at a point where I need, so, where I'm so stuck that I'm actually gonna go down to that help widget, it means I couldn't find it on my own. I couldn't figure out the interface on my own. It means I have some sort of edge case that a human is gonna need to help me figure out, you know?

Track 1: Yeah. And that's the thing. It's a, it is, it gets, you know, everybody throws out there the, the simplest scenario, right? Like, if you have a mobile app and you literally get like five types of questions, like people want a refund, people, whatever you have, like your, your really super simplistic questions like, yeah, it'll probably do an okay job at servicing that.

Although is it actually doing a better job than if like, the widget just popped up the top five questions and one of them is refund. Like, I [00:17:00] don't, I don't think it's necessarily, it's gonna be slower than that. Um, where like customer could just quickly see like, oh yeah, refund, I want that. Um, versus typing in, Hey, how do I

get a refund

Brian: But even stuff like that,

like I feel like shouldn't even be in a customer support. Like it, it should just be in the interface. If, if you're, if you're letting people like

one click refund, then just give them that button,

Track 1: Right, .

Brian: Um.

Track 1: Right. So there are, I mean, I think there are some simple use cases where like, on it's gonna take away, you know, maybe it's going to what you would call in the business, like deflect. Um, you're gonna

Brian: I could see AI

probably helping internally, like it could help the, the support

team

be more efficient, but it

Track 1: That's my more focus really too, is like, could we, um, like give you the

You know, some, so like most support tools have like the idea of like a canned response of some fashion. They're called something different sometimes, but whatever. And so like, could it, could we get it dialed in to where it's like recommending, hey, these three or one of these three is probably the right response for this scenario.

Um, and then really easy for the [00:18:00] agents, especially like new agents and things like that who maybe aren't as familiar. You have 200 can responses, like looking through them all and figuring it all out takes a long time to learn all those. It's like, well maybe, maybe there's some AI stuff there that can narrow that down a bit.

Um, things like that. And, you know, I don't really think there's some interesting stuff that's gonna happen with it, but yeah, I don't think it's all the way to like Yeah. Mean definitely. Hey, my initial thing was like, oh, like did we just get like, put outta business? Is this ? Like,

is this just gonna take over?

But

it's like no, it's

Brian: it's not

Track 1: We're, we're away.

Brian: not like you're doing chatting with PDFs like They're all getting

Track 1: Yeah. Right, right. Yeah, they're getting killed. So, um.

Brian: but all right. My, my, I guess back to my fundamental question is,

Track 1: Yeah.

Brian: let me ask it another way. What will be different in the product? How will, what's, what is

different about HelpSpot version two

Track 1: right. So, um, it'll version six, but it'll be Yes. So new, newest uh, right. Yeah. Um, this rebuilt version, so era two. Um,[00:19:00] So there's gonna be, so there's a lot of different things that have come together. Uh, I definitely can't get into all of them just yet, but I will say like there is some business elements that I think rebuilding will help us accomplish.

Um, and that's I, I at least wanna talk about. Uh, maybe back on uh, in a, and we can go through some of the business case, business side of it. But even in terms of the product itself, I think we're hit. So it's definitely still that same like, oh yeah, I would like to use all the new hotness.

'cause obviously Laravel has even more new hotness than before. And we have Livewire and we have all these great tools. Um, but . In addition to that, I do think you have things like ai, which while we just sat here saying it's not taking over Totally. Um, I do find it being useful. I think while we could certainly put it into the current code base, I think having it be a new code base kind of refactored from the beginning to

think about

Brian: Yeah. like, rather than bolting it on.

Ian : gonna have a right, rather just be bolted on.

It's like if it's more deeply either directly integrated or [00:20:00] even spots where, okay, we're not gonna do this yet, but in the future this would be a spot where like it would make sense to have some AI stuff going on. It's like now that can all be constructed and ready for that, even though we're not using it yet, but it's like there and ready.

Track 1: Um,

Ian : there'll be a lot of interesting

Track 1: like that.

Brian: Are there, are there any like,

like significant

Killing vs. reworking features
---

Brian: features that you're just not gonna include? Like, like kill killing off features in the process here?

Track 1: um. In terms of like the big features? No, like all the big stuff will still be there. Some of it's gonna be organized differently. I think it would probably be maybe a little too down in the weeds

for here, but um, so there some different concepts for sure and different organizational structures I think.

Um, for sure. But like big picture, like there'll be tickets and there'll be a knowledge base and um, you know, reporting and like, so there will be the big chunky things. I think. Uh, another that'll be really good and I think is more important than ever is like deeper ability to have integrations, like the older platforms not [00:21:00] really set up to integrate.

With other things in a modern

sassy type way.

Brian: and I feel

Track 1: do have Zapier and things like that,

but

Brian: for a help desk app

Track 1: yeah, it's definitely a key area to be able, so we have our own APIs for integrating, which people do. But again, now you see people obviously, like they just wanna like click a button and integrate with their CRM and not have to like write, even though it's a small amount of code to do it now.

And we have a pretty cool system for doing it. It's still like code that has to be written to do this integration, you know, and obviously that's a barrier. So for the bigger tools that we have a lot of customers using or could use um, it will to be like, yeah, like boom, it integrates with Salesforce, let's say,

or whatever, and

Brian: How do you how do you define your, your customers now? I mean, is it, is it like a pretty wide range of industries?

Track 1: Yeah, we're super horizontal, like, um, and it's a lot of things you might not even think of. So obviously like you have IT support uh, and then End customer support, whether you have like a software company supporting customers with a software product, or you're [00:22:00] a manufacturer and you have, you sell a product or e-commerce site and you sell products and things like that.

Um, but then lot of other use cases too, like HR departments or big, actually growing segments. Um, 'cause you know, all this email

Brian: internal support

Track 1: about benefits and things like that. Yeah. And so like they're supporting their internal staff with HR issues or maintenance departments is one.

We've another one that's been around for a long time where like people can email you and be like, Hey, the door and building C is messed up, or whatever. And this way, like that's organized for them uh, across a

Brian: pretty incredible when you like it. It's like When you already know that a, that a category is, is super huge and then you find out it's like actually a

Track 1: huger

Brian: than that.

Track 1: It's

Like anybody who deals with

Brian: there are all

Track 1: Like, yeah, there's so many use cases. Um, which is why there's like literally thousands of help desk I mean, there's legit thousands of competitors. Um, and obviously we all know Zendesk and Intercom of the big ones, but.

There's just, they're never [00:23:00] ending. And then like there are, and then even in those segments, right, like there's all these different takes on, I mean, 'cause there are dedicated HR tools

that do the email portion and they do a bunch of other stuff and there are dedicated maintenance tools and they do maintenance stuff in addition to the email.

So like, but you know, there's sometimes more expensive, for example, or whatever. The people don't want all those other features, they just want the email management. So they instead go to something like HelpSpot. So, uh, so yeah, so it's even more complicated than that. 'cause like in vertical there are dedicated tools that overlap with what a help Spott does and things

like that,

so,

Brian: So, so

Thinking about competitors
---

Brian: I guess given that, given all the different use cases, especially in all these different industries that you wouldn't even think of, do you,

how, how much do you think about, or, or, or focus on competitors? And especially like, like who do you even consider to be competitors? Like there's the obvious ones like that everyone knows the intercoms and the [00:24:00] Zendesk and all that, but like, do you even think of

them as

Track 1: Um,

Brian: or are, are there all these like use

cases for help Spott, where it's like, your customers probably don't even know

Track 1: right . Right. So that's why I don't get too caught up in the competitors. Like I have a rough idea of like, I keep an eye on the very biggest ones, just as like a sort of barometer. Like I don't even, I don't, I can't say I even give it like a ton of thought, but I will poke in on like what intercom's doing, what Zendesk is doing um, beyond like those two, uh, I don't really pay a lot of attention to anybody else. and even those, it's not so much because, we're always competing against a lot of different companies. Like, and it's a, it's a kind of product where pretty often there is a scenario where it's like, Hey, we have, you know, we're looking at seven different products and we have a committee that's evaluating and like, it's a whole thing, right?

And so we're doing demos and blah, blah, blah. And it's like, there's just so many tools. It's like, so we will sometimes get into one when it's like a customer's asking us like, Hey, does [00:25:00] it do this thing? Another tool does or whatever, and then we'll go figure that out. But

uh,

but yeah, I don't, I don't, really spend a lot of

time with it.

Brian: I mean, that's the thing. For me with competitors, like, yeah, I could do a Google search and find a bunch of competitors, but I, I don't do that.

I don't care about that. What I care about is what the customers are telling

Track 1: primarily

Brian: asking me about,

how are you different from this, or I'm switching from, from that.

And, And, like that, that's who raises to the, to the top of the list of the competitors that

Track 1: Yeah. That's pretty much my take on it too. And you know, I think it's also, you have to, it's, you know, you have to factor in like your place in the market um, your business model with it. I mean, we're definitely, you know, a lower cost provider. We're not the cheapest provider, but we are also not. Zendesk. Um, so Zendesk is, you know, gonna be $50 an agent at the pretty low end, up to hundreds of dollars an agent a month.

Uh, Salesforce has a help desk module that's the top tier is [00:26:00] $650 a user a month. So, you know, these are all these different market segments. Um, and so yeah, you gotta niche yourself in both with the feature set and the pricing model. We have some customers who are very price sensitive. We have other customers who are less price sensitive.

Um, and you have big customers that are price customers who use us 'cause I. They can get 250 users for $29,000 a year. And they're like, that's great. 'cause we don't wanna spend $290,000 a year for the task that this is, because it's probably not, in those cases, with HubSpot anyway, it's not usually like a room full of literally 250 agents.

It's more like there's maybe like 50 core agents and there's 200 sort of other people, whether they're managers or whatever, who are like not in there all the time. And they don't necessarily wanna pay $200 a user a month for a person who goes in there for an hour a month or whatever. So there's those kind of trade-offs um, that people are making.

Brian: Hmm.

Roadmap and timeframe
---

Brian: I guess as

we as we [00:27:00] start to wrap up this, this episode about the rebuilding, I,

I mean, what can you, what can you tell us about the, the, the, the project and the process of actually rebuilding it, and then, and then sort of the

plan for rolling it out. So like, the first, the first question that I would have about the process is like, how, how long it, all right.

So, I mean, once you get past the strategic decision making and say, and saying like, all right, we're gonna break ground on this and start coding or designing or whatever, like from that point to having some version ready to bring out to the world. Like, is that a year? Like, is that more, what are

Track 1: Well, you know, so it's sort of interesting. I've, I don't like to work with deadlines, , because I've worked for myself for a long time here. Right. And the product's doing fine. So there's not like this big push of like, boy, we, we need this new version because like, things are heading in the wrong direction and like, let's really crunch on it.

Um, so I, I, I

Brian: My, I, I

guess like a

part

B.

Track 1: [00:28:00] kind of my ballpark,

uh,

estimate, but,

Brian: Yeah. And

like a part

B on that is like,

Founder designing vs delegating
---

Brian: All right, you're rebuilt, you're, you're building a whole new version. So that takes a ton of resources.

What's going on on the existing product during that time, right? Like, what's the cost in terms of,

Track 1: So it's sort of interesting. This is a change I made actually um, from previous times we've tried this and I was, had, I had developers kind of doing the work on it and then I was managing and uh, I've gone a different route this time, um, where the sort of version one of the new build.

Um, and it won't be all me forever and it won't be all me all the way to the end of where it's releasable, but it is gonna be all me for right now because I feel like I'm just, this is just a totally me thing, but I'm not good at sort of thinking through the problem space. Um, and then just like writing that uh, because then I'm almost never happy with what comes out the other And I just feel like I need to be the one in there, in [00:29:00] the code. Trying out, do I like this interface for this idea? Getting the big structural things of how it should be in my head out. And then like, once that structure's there, then it's like, yes, like we can get in there and have other devs like working on the, obviously the million little features that come out of

the, you know, the, big structure of it.

But

I, I,

Brian: way with that stuff. And, and, and I, I've, no, that's another thing that I've noticed on your Twitter feed over the last several months is like, man, Ian is getting really

Track 1: I'm back in the code. I'm deep

Brian: you're, you're, you're

definitely getting back in it. Yeah.

Um, but the

I, I'm the same way with, with with stuff.

It's like I, I, I design and build most of the product myself, but then like for the initial version of stuff, but then it gets to a point where we need to ship so, so much that I do delegate, that I do spend a ton of time writing out detailed, I mean super detailed,

like specs for, for my developer to

build a new thing and it [00:30:00] A lot of times it gets to a point, it's like, man, I'm spending a whole week or more just writing

Track 1: Right.

Brian: for my developers. It's

like, I literally could, and I'm, and

I'm writing them in a checklist that's like, literally like programming logic. And I'm like, why am I not just

in the code base right now?

You know? But, but there's, there's

still the thing

where it's like, all right, I, it seems simple on paper, but I know that once I get into it, one week is gonna

Track 1: Right. And also there is the, like, you know, it depends on your setup. Like, I mean, how experienced your developers are, how long they've been with you, and know the product and all those things. Like obviously you can get shorthand for things a lot easier with somebody, you know, like, so back to your kind original question, like mainline HelpSpot is still chugging along, like I still do a little bit in there, but primarily um, there's another developer who does. And the bugs are, Yeah.

bugs are being fixed. New features are being added. So stuff's still happening there. Um, and we're, again, it's a 20-year-old product, so it's not like people are like, oh man, it's been [00:31:00] a week. Why isn't there new features and new stuff? Like, so if we are like every month shipping a solid release or every two months with like good bug fixes, a couple nights, little features, like, you know, nobody's coming to us being like, what the hell is going on?

Where's the new stuff? Like, everybody's happy, customers are happy, we're happy. So we can kind of keep that pace going. It's fine. Um, and you know, that's the developer who's worked here like nine years, right? So like, he knows deeply help spot when he, when it's time for him to work on six or, you know, what I call six, the new version.

Like, you know, he understands the philosophy of what we're trying to do, right. And how we operate. So I don't, there's a lot that's baked

in.

Um, so that helps a lot.

Brian: what's the plan

with with customers? So, so when it, when it comes time to

release it, I'm, I'm guessing that'll be like sometime in 2024.

Track 1: Yeah, I would say 2024, probably uh, maybe earlier, but

Brian: So

Rolling out the rebuild
---

Brian: Like it are, is it gonna be like, Hey everyone, we're, you're all moving over? Or is it like you can use help Spott Classic,

Track 1: we're going [00:32:00] uh, so I'm not the world's biggest DHH fan, don't have time to cover that. We don't have time to cover that here. But um, but I will say one of the things I do like that they did is you could stay on the old version of Basecamp if you want to. And that's definitely the philosophy we're gonna have um, for multiple reasons.

But, you basically we, yeah, I don't wanna move everybody. There will be these differences of how it works. There will be differences in the structure. Um, you know, we're a very small team. Just even construct, have customers with . Dozens to some of them, hundreds of gigabytes of data uh, you know, the construction of the way that to the new system.

I mean, that's a huge application on its own. So I don't necessarily wanna wait for that huge application to be built until we release this. So it's like, hey, my current thinking is there'll be like a light version of that, which is like, you want to be on the new hotness. Um, there's gonna be a way to, may maybe move your users, move your general structure, but like, maybe not all your data or maybe not other, you know, some, [00:33:00] some, or maybe it's like the basics of the data, but not like all the far reaching tentacles of the data.

You know, something where it's like we can build a simpler v one of it for people who wanna be bleeding edge and move over, like, and they're willing to take that trade off.

Um,

Brian: Hmm.

Track 1: everybody else, they can

Brian: I guess

it also really

helps, it also really helps that HubSpot has always had this on-premise version. So like, theoretically, like it's, it's like the, even if, even if you have all these you know, legacy customers who were on your quote unquote hosted version, that's like, they're basically using the on-premise version, but you've, you've installed it.

Track 1: We could run that,

Brian: so you, you, you've, you've always

had

that, like, that,

that ability to just offer it on premise. So it's like you'll, you'll continue to have customers who just

use

Track 1: Yeah. I mean, we have customers who will use on-premise uh, forever, right. And then there's, even though, but still host the customers who are on HubSpot five, which is like the current version. Um, so there's definitely not gonna be the like, all right, it's is here. [00:34:00] You're all just getting shoved over there.

Um, it will be a multi-year project to convert the people on HubSpot. Five to HubSpot six. Some of them may never go and that's fine. Um, so yeah, that will still exist and be supported you know, it obviously isn't gonna get like a lot of new features at that point. Like one six is out. You know, that will ramp down in terms of feature development, but we'll still get bug fixes.

It would definitely get any security fixes. Um, and then six would be obviously where of the new, new features and

new hotness

type stuff.

But uh, that's kind of the current

Brian: good stuff. I'm excited. I've, I've been a fan of of watching you get, get more technical it's nice to back into I mean, that's what got me started with it too. It's like,

Track 1: uh,

Brian: you know, to get back into, it's like it's 20 years. Like I need to keep it fun for me too. Like, it's like, obviously making money is fun. I like making money. That's a, that's a certain kind of fun.

But I do like coding and I do like getting in there and designing a, a, a feature and coming [00:35:00] up with a new idea for something that's gonna help the customers

Track 1: um,

Brian: you know, be better at their. Support and helping their customers. So it is really nice to kind of just be in there and then just take all I've learned.

I mean, I've learned so much in this 20 years about how to scale things and how to, where the pain points are and the support process and all that stuff. So it is useful and fun to be able to that into action,

Track 1: um,

Brian: Yeah, put

for sure.

Track 1: the interface and everything.

Brian: in

Ian, I'm, I'm excited to see this this rollout and getting it all together. It seems like a beast of a, of a project and a, and a thing for the whole business.

So, So that's, that's the rebuild. We're gonna, we're gonna keep watching, I think in our next episode. We'll, we'll talk about Laravel and all things Laravel ecosystem.