Become a Creator today!Start creating today - Share your story with the world!
Start for free
00:00:00
00:00:01
Thoughtful Edge, Episode 9, Climbing the $900M Mountain with Ravi Parikh
47 Plays1 year ago
Join us for an inspiring interview with tech visionary Ravi Parikh, CEO at Airplance, who scaled Heap to a $900M empire! Discover the secrets of fostering collaboration, driving growth, and mastering the art of building a successful SaaS company. From Heap's early product-market fit to the innovative journey of Airplane, learn invaluable lessons from Ravi on uniting teams, adapting to industry changes, and navigating the road to riches. Gain insights on overcoming challenges, embracing creativity, and leveraging mentorship to accelerate success. Tune in to unlock the blueprint for constructing a corporate powerhouse in the ever-evolving world of technology.
#saas #techvisionary #collaboration #thoughtfuledge #heap #airplane #industryinsights #scalingsuccess #RaviParikh
Recommended
Transcript
Ravi Parikh's Entrepreneurial Journey
00:00:00
Speaker
Hello. Hello. Welcome to Thoughtful Edge. And today we'll be having a conversation about bridging the gap and how to do collaboration and lessons from building a billion dollar company with Ravi Parikh. Welcome Ravi. Thank you for having me. Sure. Thank you for coming. So Ravi has a computer science degree and he has an engineering background.
00:00:25
Speaker
And now he's an enthusiastic entrepreneur and one of the founders of the big attack platform which provides insights. And the name of the platform is Heap. And we are going to talk about his experience there and maybe about his current ventures, Val. He's the founder of Airplane and might be having good insights for us to tell about his experience and achievements there.
00:00:53
Speaker
So at HIP, he was able to grow up the company and team up to 200 people, if I'm not mistaken, and achieve the evaluation of almost 1 billion. So can you share some key lessons you've learned while doing that and while scaling the company and while doing all that great stuff at HIP?
00:01:18
Speaker
Yeah, it's a good question. As you mentioned, I co-founded Heap in 2013, along with a friend of mine named Mateen, and then helped grow the company to about 200 people over the next seven years. I left the company in 2020, but the company is still around. It's now about 400 people, so it's roughly doubled since I left. Yeah, it's worth close to a billion dollars, not quite a billion. It's worth $960 million, but
00:01:43
Speaker
kind of in the same ballpark. There's a lot of lessons learned along the way. Obviously, spending seven years anywhere, you're going to learn a lot. I would say that some of the key lessons that really were some of my biggest takeaways were, I think number one was on the product side, all the best decisions we ended up making on the product side were
00:02:03
Speaker
due to having a deep understanding of our users. And all of the worst decisions we ended up making were when we started trying to be too strategic and trying to be too... Overthinking. Yeah, overthinking things and feeling like we were smarter than our users. We knew better than they did what they needed.
00:02:22
Speaker
I think ultimately, your users can't tell you what to build, but your users can tell you what their problems are. And I think really understanding those problems and believing those problems and having a deep understanding of them is step one to building any great product, no matter what it is. And so that was really the core driver of all the best things that we ever did at the company.
Hiring for Success
00:02:42
Speaker
I'd say that's lesson number one.
00:02:44
Speaker
I think lesson number two is just never being a rush to hire. I think ultimately, once a company hits a certain level of scale, its success or failure is just based on the people in the company. You as a founder cannot do everything past maybe the first couple of years.
00:03:01
Speaker
Ultimately, if you hire great people and you onboard them well and set a good culture, then they're going to be effective. And if you don't, they're not. And I think some of the worst decisions we made were when we felt a lot of pressure to really fill a role really fast and then ended up compromising on some of our values or some of the things that we really cared about because we thought, Oh, well, you know, we really need to fill this role today. Um, I think the thing I've learned is, you know, as, as painful as it can be to have a key role open for a long time, it's much more painful to the wrong person in that role.
00:03:31
Speaker
So I'd say those are the two biggest takeaways that I've tried to bring with me to my next role.
Founding Heap: Solving Personal Pain Points
00:03:37
Speaker
There's obviously lots of other small lessons along the way, but those are two big things that come to mind. Awesome. Yeah. The wrong hiring decisions that might definitely hurt than anything else, because ultimately, just the people who will do the job, who will make the company successful. And what we see now, the market is just
00:04:00
Speaker
reconveyed the message once again like all the big companies they were hiring like crazy for the several years and then they realized oh there is no demand anymore for that kind of for that amount of people so we are gonna let them go and then
00:04:18
Speaker
And I'm just even more curious now, like you said that you started that company with one of your friends at that time. And why did you decide to go to the ad tech? Since I'm an ad tech, like I like the industry, I like everything what is going on here.
00:04:37
Speaker
And what was your preference? Yeah. I mean, the reason we started the company, it was really, so Mateen, my co-founder, was the one who actually came up with the idea and ultimately ended up starting Keep First before actually bringing me on as sort of a late co-founder a couple months into him, starting to build it out.
00:04:53
Speaker
Um, but he, uh, came up with the idea. He basically, um, it was from his own personal pain point. So he used to be a product manager at Facebook. And, um, he was the idea behind heap is to sort of like automatically collect client side data from like, um, websites, mobile apps, that kind of stuff. So you don't have to sort of spend a lot of time instrumenting, um, tracking code manually. And so he came up with that idea because he had sort of experienced that pain point over and over and over again at his previous company at, at Facebook.
00:05:22
Speaker
And so it was a very natural idea for him. And that's kind of where we sort of started building Heap from. And he pitched me on the idea when he had just started building it. And sort of the idea made a lot of sense to me. I've sort of implemented analytics tools in the past and had a lot of pain around that. So it just made a lot of sense once he kind of like gave me the idea and we started working on it together. So yeah.
00:05:42
Speaker
Yeah, I also found the industry quite fascinating with all the data and user behavior.
Transition from Technical to Business Strategy
00:05:50
Speaker
It's interesting to learn and do something useful for the people as well. Because I found a great failure in the ad tech and advertisements and marketing. I don't know, whenever I go to some website or to the social media and they target me with some useful stuff, why not?
00:06:12
Speaker
Okay, and as we are both coming from like technical backgrounds, computer science degrees and everything, sometimes it's a struggle for us to balance our
00:06:26
Speaker
like technical knowledge, a deep understanding of how computer works. And then when we move on to some business roles, we're struggling to find the balance between leaving the go-to-market team and still having this temptation to be as deep as possible in all those small technical details and stuff. And all the technologies that are coming out of the framework
00:06:55
Speaker
So what was your learning on that area there? Yeah, that's a good question. Yeah, so both me and my co-founder, we were both engineers by background. So neither of us had a go-to-market background, really. And scaling any SaaS company requires balancing both. Most SaaS companies don't sell themselves. You have to actually build a go-to-market team to take the product to market in an equally important part of building the company.
00:07:23
Speaker
And so as the company grew first year or two years, we're really just me and my co-founder writing code and talking to customers. We weren't really doing much else. But once we started getting off the ground and started generating revenue and stuff like that, it became pretty important pretty quickly to build out a sales team, to build a marketing team, build out customer success, all that stuff.
00:07:40
Speaker
And so I took on that challenge. My co-founder lead more towards managing product and engineering. At the end of the day, the company was his idea. It was his sort of product brainchild. So he kind of led that side and I kind of love to go to market side. And really what ended up happening was I think one thing we both did well was we were both pretty self-aware about what we were good at and what we were not so good at. And so.
00:08:04
Speaker
I knew I was not a go-to-market expert. I knew I didn't know anything about sales or marketing or customer success or any of that stuff. I mean, even some of the really basic terminology was just new to me. And so we were just very, I think, humble about asking for help. And so we were very lucky. We went through Y Combinator with our company, and YC is full of other B2B companies that are at various other stages. And so we would talk to our friends who worked at these other companies and just get their sense of like, hey, how did you get your first
00:08:31
Speaker
customers how do you get off the ground how do you hire first sales people what is that profile look like you know what should we do and so you can't outsource all your decision making up to make some first positions yourself but i think a lot of. Technical founders make the mistake of trying to solve too much from first principles like i think.
00:08:47
Speaker
I built a great product for first principles. People really like it. So I must be able to engineer a whole go-to-market machine. And the thing is that there is a lot of cultural knowledge in Silicon Valley about how to do it. We were actually very shameless about just, let's ask five other companies how to do it. Four of them said, do X. One of them said, do Y. Probably X is the correct decision. So we would do that for a lot of things. And that was quite helpful to us in terms of getting those first few sales hires correct, getting customer success process correct, all that kind of stuff.
00:09:15
Speaker
And so we leaned on that a lot, and then we just hired really good people and gave them a lot of ownership and a lot of trust. And so if I hired someone who's really good at sales and has a good background, I'm not going to tell them, here's
Fostering Collaboration in Teams
00:09:26
Speaker
how you should sell. I'll tell them, hey, here's what's worked for me. But you should integrate that knowledge into your own best practice of how you know how to sell.
00:09:32
Speaker
And so I think by doing that, we ended up having really good success. And so nothing super complicated, really just deferring to people who are better at these things than me and my co-founder were and just like setting the high level direction correctly. And that worked pretty well for a while. So yeah, I think that was more or less the way we approached that challenge. Awesome. Awesome. Yeah. The way you trust people and the way they trust you and you're more capable of building.
00:09:58
Speaker
those more reliable relationship and especially if you trust their expertise, if you know that they can do good themselves, you just give them general direction and it usually works out well. By the way, what was the moment when you realized that, oh, we cannot do that, everything else else, we need to seek for any kind of assistance? What was the moment that lead you to that decision?
00:10:23
Speaker
Yeah, I mean honestly it wasn't any one moment it was more just like as the company grew gradually became kind of obvious that there was a lot of parts about building this company that we weren't quite aware of how to do effectively and like yeah we can sort of speculate about how we might do it but
00:10:39
Speaker
The nice thing about having gone through Y Combinator and living in San Francisco at the time and being surrounded by other SaaS companies is just our friends were trying to deal with the same challenges. And we would just hear these stories from them, people who are a little bit further along, that kind of stuff. So yeah, I think that being surrounded by that culture of people who have similar set of challenges that you do is not to be underestimated. I think that knowledge kind of seeps in and helps you kind of craft the way you're building your company better.
00:11:06
Speaker
Yeah, knowledge sips in, right? Yeah. Yeah, it's the environment. It's definitely like environment and the battery, especially. Yeah, it's definitely my makes you like things differently as well.
00:11:25
Speaker
And if we are moving on and thinking more about growing and growing the team and growing the company, the collaboration is an essence here. And what strategies did you implement to foster that across the company as Hebrew and maybe used any specific tools or techniques that you found effective?
00:11:50
Speaker
Yeah, I mean, I would say like, honestly, it's something that we did. Okay, I think we could definitely have done a lot better job of it in retrospect. Ultimately, I think the thing that I found over time as being one of the best ways to foster collaboration was just spending time with each other. And that sounds kind of obvious. But the thing that I've noticed is you grow is, you know, let's say, you know, let's say you're a 10 person company, you naturally spend a lot of time with each other, it's hard not to.
00:12:17
Speaker
But if you are a 50%, 80%, 100% company, everybody's calendar is totally full. If you look at any executive, any VP of sales, VP of marketing, CTO, everybody's calendar is totally full with meetings, internal meetings, external meetings, recruiting, sales, whatever it might be. And as a result,
Navigating Enterprise Market Challenges
00:12:37
Speaker
it's very easy to neglect
00:12:39
Speaker
the one-on-one time spent with each other. And so one thing that I actually felt was really effective at Heap, around when I was leaving, we hired a COO to take over my role. He actually eventually became the CEO of the company. My co-founder left as well. And
00:12:54
Speaker
One of the first things he did was when he came on board, he said, look, the executive team is full of really talented people, but it's kind of dysfunctional because everyone is off on their own island. Product is over here. Marketing is over here. Sales is over here. One of the first things he did was he had everybody who was leading each of these teams
00:13:12
Speaker
Um, just to a one, like a standup every single day for half an hour. And this felt really expensive at the time. We're like, Hey, you're taking all the highly paid execs in the company and you're taking 30 minutes out of their day, every single day in the morning. That's time that could have gone towards building the business, all this kind of stuff. So people kind of were annoyed by it. But the thing we noticed is even after just a few weeks of doing that, the level of communication and empathy that people had with each other just grew a lot. And the sort of level of misalignment between teams went way down.
00:13:41
Speaker
And so just forcing that time, that frequency of time being spent together across departments was really effective once we got to a certain level of scale. And so that's, I think, something it's easy to not do. The easiest thing to cut from your schedule is one-on-ones with people. But you have to do them.
00:13:59
Speaker
You have to respect that time. It's stuff that is really invaluable, especially for cross-departmental collaboration. I haven't found more of a shortcut solution than that to foster collaboration. Right. But how other way you'll be close to people, how you will build this bond to them and
00:14:21
Speaker
Yeah, this is available advice, like to truly agree with you. And as you exactly said, it's easy not to do that. It's easy to forget or neglect or whatever. People will care about their stuff themselves. They don't need to be there. It's not true. Yeah. And as Heap moved up markets with its product offering,
00:14:46
Speaker
What's the main challenges you encountered and how you overcame them? Yeah, it's a good question. So Yaheap started out as very much selling to startups. We were a self-serve product. You could pay for $50 a month on a credit card. That was kind of the earliest price point. And then we slowly moved for those prices over time and eventually became purely an enterprise, or most of our revenue came from enterprise. There still is a free tier and a self-serve tier for Heap, but the majority of the revenue comes from people paying six figures or seven figures for the product.
00:15:15
Speaker
so i would say a lot of challenges on the way moving up market tons and tons of stuff that it took for us to raise our deal size from a few thousand dollars to a few hundred thousand dollars over many years i would say like some of the biggest challenges we face i think number one we had a lot of competition and
00:15:34
Speaker
there was always pricing pressure on every single deal. And so it would be from direct competitors. And also we're an analytics product. We would compete with Google Analytics, which is free. And so Google Analytics even has an enterprise offering, but it's very, very cheap. And so as a result, it was very hard to sort of compete with people who are like, well, we already use Google Analytics, and we pay nothing for it. So we're going to create a new budget for your thing. That was always a big challenge.
00:15:56
Speaker
Ultimately, we had to just get really, really crisp about the ROI that our product generated. And so we had a lot of case studies with customers. We had very precise viewpoint on how our automation of tracking code would save engineering time directly, allow you to accelerate your product roadmap, allow you to do more testing and iteration.
00:16:17
Speaker
We had really crisp stories around that. We trained our sales team really well to be able to tell those stories and to be able to contextualize that ROI in the customer's view. And so when it ultimately came down to it and we asked for $250,000, it should be hopefully obvious we're going to save the company 10 or 50x that amount in terms of revenue, uplift, or cost savings.
00:16:38
Speaker
at the sort of size of companies we were talking to. And so I think if your ROI case is very clear, if you're driving differentiated real value, it's really hard to prove those things for a software product. But if you can, and you can back that up with real data points and real case studies of customers who have done it, there's not really any substitute for that.
00:16:55
Speaker
Very, very few products in software have obvious ROI. Most products, you have to kind of prove it indirectly. And so you have to build up those case studies over time. There's no substitute for having real-life customers who will attest to the savings that your product has driven for their companies or the revenue uplift that they've driven through it. So that's ultimately what you've got to do. There's no real substitute for it. And then you could train your team really well to be able to articulate those stories.
00:17:22
Speaker
So you actually, you're always emphasizing on ROI that you're bringing and then it was the differentiator for you, right?
00:17:32
Speaker
Yeah, for sure. It was tying that all right back to the product differentiation. So, you know, it's one thing to be like, Oh, analytics is helpful, but it's weird to prove that heaps specifically is going to help you in ways that are unique. You know, if we're competing head to head against, uh, Pendo or amplitude or Google analytics, and they're coming in at half the price. Um, but we can say, well, this, this set of unique offerings that we have is going to provide. Outsize value that even paying double for heap is should still be seen as a bargain.
Heap vs. Competitors
00:18:00
Speaker
That's really what we tried to prove to people.
00:18:03
Speaker
Did you have a moment by the way that you've realized that your competitors started to noticing you and like paying attention to what you do and maybe trying to replicate some of the stuff?
00:18:18
Speaker
Yeah, we definitely had that. So when Heap started, it was 2013. Obviously, we were very small on day one. Our main competitor was a company called Mixpanel, which is still around. I think they're doing pretty well. But they were really big. They were five years older than Heap, and they'd been around for a while already. They were kind of the new analytics tool that was doing really well.
00:18:42
Speaker
And so every single customer that we talked to was either using Mixpanel already or comparing us to them or things like that. And Mixpanel didn't really care about us, of course, because we were a new startup. But three, four years in, when we had started maybe stealing a little bit of market share from them and growing a little bit, they'd started noticing us more. They actually came out with a competitive feature. So the Heap's main innovation was, Mixpanel was a powerful product, but it took a lot of work to implement. You have to basically write a bunch of code to implement Mixpanel.
00:19:10
Speaker
Whereas the way HEAP works is you install HEAP once, it takes you five minutes to install it, and it'll automatically track all that stuff that Mixpanel took manual work to do. So it was basically more of a low-code, no-code approach to analytics compared to our competitors. And so Mixpanel noticed that, and I think around 2016 or 17, I think, they rolled out a feature they called AutoTrack, which was their competitive offering to us. And they even sent out their CEO sent out a tweet that day saying, oh, we're coming for HEAP or something like that.
00:19:38
Speaker
Ultimately, it ended up being that day when they rolled out that feature was our number one day in terms of signups. Because when Mixpanel announced that Mixpanel was still so much bigger than us, there was a lot of this attention. Like, why did they build this feature? Oh, it's to compete with this product called Heap. Oh, what's Heap? Let me go check that out. So we actually got a lot of attention from that. And I think the really interesting thing is Heap was built with this feature in mind from day one. And it was a fundamental differentiator. It was core to how the entire product functioned. Whereas Mixpanel had tacked it on after the fact.
00:20:07
Speaker
And as a result, it never really worked. From what we heard from mixed panel customers, there were some clever ideas in there. I'm sure they had done a good job building it, but it wasn't woven into the product at the same deep level that we had done so. And so it never really got real adoption among their customers. They didn't really understand how to make it successful.
00:20:23
Speaker
And ultimately, they shut down the feature like two years later. So it never ended up being a threat. It was mostly helpful to us. It was validation that what we were doing was the right approach to analytics. And it ultimately ended up being more of a positive than a negative to us. So yeah, people definitely notice. I think that's really common for startups. You come up with a cool new innovative thing. By the time you're big enough that people take notice of you, you're already really far along. And you really not just build that differentiator. You've built 20 other things on top of that differentiator.
SaaS Market Evolution
00:20:50
Speaker
It's hard for your competitors to really internalize that.
00:20:53
Speaker
So yeah, right. And they also made a mistake of mentioning you as like, yeah, something they try to go after. Yeah, yeah, that's funny. And, you know, like fast paced world, everything is changing with that tremendous speed that we cannot just keep up with everything. And
00:21:15
Speaker
Now you work at different companies, like you're building your plane, which is apparently another successful startup, which is taking up. And like those experiences, they are almost 10 years apart. So when you're working at Heap, it was like,
00:21:34
Speaker
2013, 14, 15. And now we are in 2020s. And do you see any difference between those kind of errors when you build a company that time and what is happening right now? Is there any differences, approaches, or how you can compare those times?
00:21:59
Speaker
Yeah, it's really different. I think some things are the same, some things are very different. I think the things that are the same, having a deep understanding of your customers, there's no substitute for that. Having to really understand the customer problem you're solving, there's no substitute for that. But I think a lot has changed. I think when we started Heap in 2013, it was not
00:22:19
Speaker
SaaS was not a super lucrative industry in Silicon Valley. People were still more interested in funding mobile apps and marketplaces and a lot of consumer stuff rather than B2B. And so when we started Heap, there were only three public companies that were SaaS companies that were trading above $10 billion in market cap. And they were Salesforce, ServiceNow, and Workday. And even those Workday and ServiceNow were not that big. They were just a little bit bigger than $10 billion. Salesforce was the only really big one there.
00:22:48
Speaker
And so people saw B2B as a way to, you know, build, you know, okay sized companies, but not really consistently get to huge outcomes that would return your whole VC fund. So very few VCs were like B2B focused in the way that they are today.
00:23:05
Speaker
And cloud adoption was growing. Clearly, it was a trend everyone was aware of, but it wasn't yet obvious that that was where all of computing would move to. And so there were still huge on-prem vendors like Cloudera and stuff like that that were quite successful. And so that was the environment in which he started. And as a result, it was
00:23:25
Speaker
It made fundraising a lot harder, but it also made company building a lot easier because you didn't have 50 competitors springing up overnight every time you had any interesting ideas. And so today, 10 years later, we're in an era where there have been so many multi-billion dollar SaaS companies, $10 billion, $50 billion SaaS companies haven't been created over the last 10 years.
00:23:46
Speaker
It's become basically the dominant vertical in Silicon Valley. It's probably the number one place where VCs have made money over the last 10 years. I mean, I don't know the exact stats, but that's what it feels like, at least from my perspective. And so any decent SaaS idea can get funded these days. And any decent SaaS idea is going to have at least 10 other people who have similar insights. And so with Airplane, we had a pretty interesting innovative idea, but there are four or five other companies I know of
00:24:12
Speaker
that have had very similar insights. I don't think they copied us. Maybe a couple of them did, but I think a lot of them probably came to those same insights independently of us because they had experienced the same pain points we had. You just have to be a lot more crisp about building a differentiated, really valuable product and about having a differentiation value compound very quickly.
Rapid Product Development in Modern SaaS
00:24:33
Speaker
The SaaS companies that do well today, I think it seems like the ones that
00:24:37
Speaker
like ship very fast. Not that it's ever bad. It was always the case that building product quickly and shipping quickly is important, but it's more important than ever to be able to take a small edge and compound it very quickly. And so you look at a company like Rippling, for example. Rippling has been very successful over the last few years. And part of their success has been they've built a massive platform very, very quickly. They've
00:25:01
Speaker
built something that does HR and IT and payroll and benefits and basically every aspect of back office for a company with just sort of one offering rather than having to buy five or six different vendors. And so that to me is like, DEAL is another example of coming, DEAL.com.
00:25:21
Speaker
They're also sort of like international payroll. They started out with just sort of hiring contractors internationally, but they quickly added full time. They added the ability to do benefits, all this kind of stuff. They added lots of country offerings very quickly. So I think that speed is much more of a key value now and a key differentiator now than it was 10 years ago because you just have so many more competitors that you're trying to stay ahead of. So I think it's one thing that's changed. Yeah.
00:25:46
Speaker
And he mentioned that speed, which is apparently crucial nowadays, and how this speed is achieved, what people started to do differently compared to what we had 10 years ago compared to now. Where this speed is coming from?
00:26:06
Speaker
Yeah. I mean, ultimately there's not any easy answer. I think you just have to be really good at building product very, very quickly. I think, I think in some cases, speed comes from having a clear understanding of, of customer pain point. I mean, I keep coming back to that, but it's like, uh, I think about products. There are certain products that I use. I don't want to call it specific products, but there are certain products that I use day to day and have used for 10 years and they don't seem like they've changed at all. You know, they seem like the same. They had six, 10, eight years ago.
00:26:36
Speaker
I don't think it's because of people there aren't doing anything i'm sure they're actually shipping features left and right but i'm everything they build isn't for me and so as a result i don't end up seeing any of the new innovation.
00:26:47
Speaker
Whereas I think if you have a clear understanding of your user, then every new thing you build should ideally be something your user will see and think, wow, that actually makes this part even more valuable for me. And so if you have a crisp understanding of who your users are, and every two weeks you're releasing yet another thing that they like, then they will perceive that as being very speedy and very efficient. I think about a product that I've used for a few years that I really like. It's called Superhuman. It's like an email client that is fairly popular, I think, these days. And the thing I've noticed is
00:27:17
Speaker
that of all the products I use day-to-day, it's one of the ones that I've noticed the most change in over time. They're very good at telling you that new features have shipped, but also when those new features ship, it's very clear to you where you think, wow, this is actually something I want to use. Even if I wasn't thinking I needed it, now that it's there, I'm going to start using it.
00:27:36
Speaker
And so that's like a, I think creating that feeling your users is not easy. Linear is another example. It's a task tracking software we use to sort of do our project management and all that kind of stuff as a company. They're sort of a competitor to Asana and Jira. And compared to sort of other task tracking features of tools I've used in the past, you know, every one to two months, they'll come out with something where I'll think, wow, it was actually going to make us more efficient as a business. And so I think that like, you know, I don't think their engineers write code any faster than anyone else, but I think they
00:28:03
Speaker
their founders and their product people and their engineers have such a crisp understanding of what the value they're trying to drive is, that everything they do build feels high value. And that gives you this feeling of speed and compounding effect. So the technology itself may not play that huge role in speeding this up, but you think it's just the mindset that the product team needs to embrace, right?
00:28:29
Speaker
I think so. And look, this is all easier said than done. It's really, every company wants to say they understand their users really well and that they build the right things. Every company thinks that. It's not simple by any means, but yet you see a difference. You see a difference between when I use linear versus when I use Asana, not to pick on them in particular. I've used both products for long periods of time. And there's only one of those two products where I felt like everything they built was really moving that product forward.
00:28:54
Speaker
Um, and it wasn't due to a lack of effort necessarily on one side or the other. It was more, um, there's a crispness of understanding there that that's just really hard to replicate. And so I think that's much more crucial in 2023 today than it was 10 years ago, 10 years ago. I think, um, you still needed that, but I think you could get away with saying you had a disruptive innovation from day one and you could sort of just rely on that differentiation lasting from a lot longer period of time. Um, today you can't rely on that as much.
AI's Future Impact on SaaS and Work
00:29:22
Speaker
need to be as fast as possible. Yeah, you know, that's competitive markets. Okay. And this way I'm moving on to the future. And like with all that AI and it's crazy, new technologies that are popping up every day. What the future of SaaS and tech industry in general, the future of airplane as well. Like how do you envision that?
00:29:49
Speaker
Yeah. Honestly, right now is maybe one of the hardest times in the last decade to predict what's going to happen next because of this sort of AI wave. I think there are a lot of split opinions on what's going to happen. I think that right now, a lot of the AI stuff is a little bit overhyped. I think it's really real and genuinely valuable, but
00:30:14
Speaker
I think people are trying to use a lot of these large language models in ways that they're not really meant for right now. I think what they're really good, they're very good at summarization. They're very good at generating content. They're very good at generating ideas. They're not very good at things where you need high levels of precision and accuracy.
00:30:30
Speaker
And so if you end up having to use them in ways where they're autonomously generating some content that's going to be used somehow without a human review step, it's not going to work, whereas people are trying to use it that way. However, that doesn't mean that the next version won't be good enough to do that.
00:30:46
Speaker
And so I think this is the thing that makes this stuff really hard to predict, which is that you can see what they're capable of today. You can see what they're good at and what they're not good at today. You can't quite tell what that means a year from now or two years from now or three years from now because the rate at which GPT-4 versus three versus two, that rate of progress has been so fast.
00:31:07
Speaker
My personal opinion, and this is purely an opinion from someone who's not an expert, is I think you will see these models become faster, more accurate, require less and less oversight. And I think what will end up happening is, let's call it five to 10 years from now, building a situation where the role of any sort of knowledge worker, whether you're a software engineer or a product manager or a content person or whatever, is going to be massively different, where you're not
00:31:31
Speaker
you're sort of relying on a lot of these tools to do a lot of the generation and you're kind of like applying that human edge to sort of curate things and edit things and stuff like that. So I think that's the direction things are trending. And I think that will mean like platforms like airplane work really well. The way airplane works, we basically build a
00:31:49
Speaker
a platform for creating internal tools. Let me give you a set of building blocks to sort of build software on top of. Right now, people write code into airplane, like they write code directly and an airplane takes advantage of that code to sort of assemble things for them. That code can get sort of like written eventually by large language models and AI and things like that in a more efficient way, but you still need the building blocks at the root of it. So I think you'll see developer tools and stuff like that that do a really good job of encapsulating key concepts with like the right primitives being successful.
00:32:17
Speaker
Um, whereas things that are more on the UI layer where you're saying like, Hey, the way we're going to add innovation to people is by building like a innovative new drag and drop system for building apps or something like that. I think that'll be less valuable in the future because the way people interface with software is going to change so much. So I think like that infrastructure layer will be more and more valuable and more durable. I think that more like.
00:32:36
Speaker
User layer will be like less terrible and will change more but again like I say all this from a position of like very little knowledge because I think we're at this inflection point where it's really hard to predict the future even the companies that make these AI models have no idea how they're gonna evolve you know even open AI themselves no idea what gpt-5 will be able to do or not able to do and all that kind of stuff so I think it's a an interesting time for sure yeah interesting but however in your opinion
00:33:03
Speaker
Do you think AI will take over of all of our jobs in the future? So, I mean, on a long enough time scale, I don't see why not. I think, you know, I don't know if that'll happen in five years or 10 years or 20 years or 50 years, but ultimately, I don't think there's anything fundamentally impossible about the ability for a computer to do anything we do.
00:33:25
Speaker
at the end of the day, my job, the jobs of all the people who work at airplane at all the companies I've worked at are more or less like you're exercising your decision making abilities, right? You're saying, you know, what, how should I build this feature? What should it do? Who is it for? You're synthesizing information and using that synthesized information to sort of create something. And ultimately that is what these AI models seem to be okay at. They're not great at it yet, but you can see the trend
00:33:53
Speaker
of how they will eventually get better and better and better at these things, how they'll be able to integrate more and more and more context, how they'll be faster at thinking about these things, how they'll be able to think about these things at a higher level. I don't see any reason why that trend would stop. So yeah, at some point, I think that changes the relationship of how people build companies and software and do their jobs and all that kind of stuff.
Solving Real Problems: Key to Success
00:34:14
Speaker
Yeah, yeah, probably the only thing is left for us is just learn it as quickly as possible and improving our efficiencies as well using all those tools. I use chat GPT often, like it creates for idea generation. For example, when I know that I need to write an article, for example, or write something, and there is a blank page in front of you and like, ah,
00:34:38
Speaker
or should they start, even though I have some ideas, but it's sometimes it's hard even just to start typing, start typing the first like sentence, the first paragraph or whatever. Okay, and as we...
00:34:55
Speaker
almost at time. And wrapping up, I have the last question for you. Based on your experience with both Heap and Airplane, what advice would you give to aspiring tech entrepreneurs looking to build successful SaaS companies?
00:35:09
Speaker
Yeah. Number one piece of advice that I'd keep coming back to is really understand the problem you're trying to solve. I've said that a few times in a few different ways. I think most SaaS companies are built by technologists, they're built by engineers, people who have really strong technical skills. Heap and Airplane are no different. Me and my co-founder in both cases, both at Heap and at Airplane, we're both engineers by background and we like building cool things.
00:35:32
Speaker
The danger, if you love building cool things, is that you build something that's cool but has no actual application. I think today, that's honestly more dangerous today than it's ever been because today, you have AI models that are so much fun to play with. They're so cool to build on top of. Every day, you go on a Twitter or Reddit or whatever, and you see cool new demos of cool new products that look really magical. And ultimately, you can build all the demos you want to, but if there's no actual
00:35:59
Speaker
key business problem that you're solving, then all you're doing is playing with technology. You're not actually building a business. And so I think ultimately tie it back to what's the problem you're trying to solve. Both Heap and Airplane came from personal pain points. Heap was started by my co-founder had a personal pain point that he wanted to solve. Airplane, both me and my co-founder here started the company because of things we experienced at our previous companies. And so we knew there was at least one person ourselves who needed this product.
00:36:23
Speaker
Many companies don't even start with that. Many companies start with zero people who actually need the product, and then they're searching for a problem to go solve. So I'd say start with a real problem, ideally a personal problem, but even if it's not a personal problem, have a real person in mind when you're building that company. If it turns out that AI or whatever cool new technology is key to solving that problem, then use it. That's great. But if it's not, then maybe choose a better problem. Okay. Sounds good.
00:36:53
Speaker
So we had a great conversation with Ravi Prich today. Thank you for coming and wish you all the best. Thank you. Thank you so much for having me. Thanks.