- Episode AI notes
- Early customer engagement and sales strategy is crucial for startups to gather valuable insights about customer perceptions and feedback.
- Selling a horizontal solution, like Ritool, to engineers who are skilled at building solutions from scratch has been successful due to the target marketās familiarity with platforms like React.
- Sales efforts, especially outbound sales, are essential for understanding the market reality and receiving honest feedback from customers to identify product flaws early on.
- The journey of starting Ritool stemmed from recognizing the need for more efficient tools and solutions after developing admin dashboards for side projects during college.
- Embracing contrarianism and motivation, despite facing objections and counterarguments, is important for staying focused on the vision of creating a universal product like Ritool.
- Innovation and entrepreneurship play a key role in advancing programming languages to improve engineersā productivity and overall efficiency in software development.
- Ensuring sustainability of business growth by maintaining a positive cash flow from the start can lead to long-term success and reduce dependence on venture capital.
The limitations of language in communication, especially between humans, can result in misunderstandings and inefficiencies in conveying complex thoughts and ideas. TimeĀ 0:00:00
Early Customer Engagement and Sales Strategy Early engagement with customers can provide crucial insights for startups. Sales can be underrated initially but can offer valuable information about customer perceptions of the space, the problem, and the solution. In the early stages, outbound sales efforts were essential as there was no inbound traffic. The strategy involved sending emails to potential customers. Iteration and pivoting became necessary when initial outreach yielded minimal responses.
David Hsu
That took four years. If you talk to customers earlier, you can talk to Dylan about this. I think Dylan, by his own admission again, would say, I wish we had talked to customers earlier instead of just developing the product in a garage, if you will. So I think sales is maybe underrated at the beginning of a startup because it tells you so much about how customers think about your space, how they feel about the problem, how they feel About your solution, et cetera. That is gold when youāre a pre-product market fit.
David
Were you doing all the sales? How did sales operate in early retool land?
David Hsu
Yeah. So how it worked actually was everything was outbound at the beginning because before we had launched, I mean, no one knew about us. So there was no inbound. No one would go to our website. At that point it was retool.in. It was like retooling. So no one unsurprisingly went to the website, just retooling around. And the only way we knew how to get customers was by sending emails to them thereās no other way basically so funnily enough the first time i posted retool to hacker news i think i called It something like show hn retool itās excel like with higher order primitives and that was actually the messaging that we used in outbound emails, too. And unsurprisingly, that combined with the retool and domain name, no one really replied. So after no one replied, weāre like, okay, well, maybe weāve got to pivot a few things here. Maybe letāsSelling a Horizontal Solution to Engineers Selling a horizontal solution is easier in the case of Ritool because the target market is engineers who are skilled at building solutions from scratch using platforms like React. Engineersā best alternative to Ritool is building a tool in React in-house, which further explains why selling a horizontal solution has been successful. This approach contrasts with selling vertical solutions like notion, where users may have specialized software for different tasks, making it a challenge to replace them with a horizontal tool.
David Hsu
Exactly, you know, Retool is or what is going on. Weirdly enough, though, we generally donāt have the problem of sort of the horizontal-ness of the product making it difficult to adopt.
David
In the same way that, like, I was imagining Notion, as Ben was saying that, like, oh, you can use Notion for anything. Yeah.
David Hsu
But what should I use it for yeah the reason why i think itās a bit easier for us to sell a horizontal solution is because we sell to engineers it actually if you think about sort of the batna If you will that an engineer has sort of this alternative as a retool it basically is build it from scratch via react. And if you think about React, React is a pretty horizontal platform too. Itās not verticalized for lawn care, for example. Whereas I think with Notion, itās probably more difficult because in the Notion case, maybe they are actually replacing verticalized software. So like you said, if you have a CRM, for example, and youāre mowing lawns, you could choose to use Notion for that, or you could actually choose to use a specialized CRM for that. Whereas in our world, typically the batten up basically is you build from scratch in-house. And to do that, you use a horizontal product called React or JavaScript, for example, also very horizontal. So I think weāve been fortunate, but thatās the example of something that was not an insight that we had when we first started Retool. It just so happened to be that way. And I think so many things about your business, you kind of learnUnderstanding Customer Perception through Sales Sales was seen as a way to grasp reality and understand customersā genuine thoughts about the product. Outbound sales helped reveal the true feelings of potential customers as they were brutally honest about their likes and dislikes. This approach was embraced as a means to uncover flaws early on and pivot the product accordingly, in order to align it better with customer preferences.
David
Being sold to and sales is not a natural motion here. How did you end up coming to this? Oh, yeah, weāre going to do sales.
David Hsu
Yeah, I think I thought of sales as sort of the best way to get a grip on reality, especially outbound sales. We actually did a bunch of investor intros, basically, of like, hey, can you tell your portfolio companies about Retool? Maybe they can try using it. And that actually gave us a very warped perception of reality because they were all very happy to take the call, but none of them converted. And weāre like, wow, whatās wrong here? Does their product just suck? Why is it so low converting? Whereas outbound sales, if they donāt like the product, like the filemaker person, they will tell you exactly why they donāt like it. They donāt have any allegiance to you, basically. And so sales, I think, was the best way to understand potential customers genuinely thought about you and your product. Having a firm grasp of reality was always very important to me.
David
Had you done sales before? Or you donāt seem like you were at all afraid to be told that your baby was ugly here.
David Hsu
No, I mean, you might as well find it earlier that your baby is ugly. You can figure out ways to your baby, for example. So I studied philosophy in college, and I think Iām obsessed with reality, if you will. And itās fun to find it and understand it and pivot it, if you will.
Ben Horowitz
Did you start Retool right out of college?
David Hsu
Yeah, I think maybe sixSales as a Reality Check The speaker learned that sales, especially outbound sales, provided a clear understanding of the market reality. Through investor intros and outbound sales efforts, they realized that customers were honest in their feedback during sales calls, reflecting genuine opinions about the product. This process helped them identify flaws early on and make necessary improvements, emphasizing the importance of having a realistic view of their product. The speakerās philosophy background and obsession with reality made them unafraid to receive feedback, valuing early constructive criticism to refine their ābaby,ā referring to their startup.
David Hsu
Yeah, I think I thought of sales as sort of the best way to get a grip on reality, especially outbound sales. We actually did a bunch of investor intros, basically, of like, hey, can you tell your portfolio companies about Retool? Maybe they can try using it. And that actually gave us a very warped perception of reality because they were all very happy to take the call, but none of them converted. And weāre like, wow, whatās wrong here? Does their product just suck? Why is it so low converting? Whereas outbound sales, if they donāt like the product, like the filemaker person, they will tell you exactly why they donāt like it. They donāt have any allegiance to you, basically. And so sales, I think, was the best way to understand potential customers genuinely thought about you and your product. Having a firm grasp of reality was always very important to me.
David
Had you done sales before? Or you donāt seem like you were at all afraid to be told that your baby was ugly here.
David Hsu
No, I mean, you might as well find it earlier that your baby is ugly. You can figure out ways to your baby, for example. So I studied philosophy in college, and I think Iām obsessed with reality, if you will. And itās fun to find it and understand it and pivot it, if you will.
Ben Horowitz
Did you start Retool right out of college?Seeking Better Solutions Through Innovation and Entrepreneurship The speakerās journey of starting Ritool right after college stemmed from the realization that building admin dashboards for side projects was a cumbersome task. Through a series of side projects during college, including developing a ticketing system for events like balls in the UK, they discovered the need for more efficient tools and solutions, leading to the creation of Ritool. The experience of making money from these ventures sparked an interest in entrepreneurship and innovation, driving the speakerās passion for finding better ways to solve problems.
David Hsu
No, I mean, you might as well find it earlier that your baby is ugly. You can figure out ways to your baby, for example. So I studied philosophy in college, and I think Iām obsessed with reality, if you will. And itās fun to find it and understand it and pivot it, if you will.
Ben Horowitz
Did you start Retool right out of college?
David Hsu
Yeah, I think maybe six months after college, we started Retool. And when we started Retool, it was because we ourselves had built so many internal tools before. And so a few friends and I, we had sort of worked on a bunch of other side projects, if you will. And we discovered sort of with every side project you worked on, you have to go build, like you said, the admin dashboard for it. And building the admin dashboard is so painful. Every engineer hates building the admin dashboard. And so we thought thereās got to be a better way. And so thatās why I started Redoil.
David
And these side projects, were these hustles? Were you trying to make money in college? What was happening here? Those are pretty fun ones, yeah.
David Hsu
Some of them made money. Actually, some of them made what seemed to me like tons of money back in college. I went to Oxford in the UK and in the UK, they have these sort of balls, basically. And so I built a ticketing system for these balls that actually made quite a bit of money. Ah, balls.
David
You mean like dances, parties?
David Hsu
Yeah, basically. That itself is actually a pretty interesting business where they sell maybe 500 or so tickets for maybe like 200 pounds each or something like that. And so it actually is like a reasonable sum of money that they are. SometimesTesting Computer Will and Intent One way to test if a computer has will or intent is to observe if it acts independently or has a self-preservation instinct like humans. While humans have a drive to survive, computers typically only perform programmed tasks without a will to survive. Computers can be seen as predictable statistical models based on past reactions, lacking the emotional depth and spontaneity humans possess. Testing a computer for genuine intent or desires, such as wanting ice cream, proves to be a challenging task.
David Hsu
We can spend all podcasts talking about it.
Ben Horowitz
Letās follow this thread a little bit. How could we test a computer to determine if it has will or intent? Hmm.
David Hsu
Well, so I think one path you could go down, for example, is does the computer have a will of its own? So if it exists in the world, does it do anything by itself? And so a good example is if I put you, you know, as a baby, for example, on Mars, maybe itās not too hospitable, but probably you do something about it. Maybe you crawl around, maybe you die in a bit, but you do something probably. If I put an ant there, it would probably do something. With a computer, probably not. I mean, the computer only kind of does what itās programmed to do, but I guess the computer doesnāt really sort of have an instinct for survival like a human, but you could argue that a Human is programmed to survive, if you will. But I donāt think a computer, like GBD-4 does not need to survive, I donāt think. It doesnāt have the will necessarily to do so.
Ben Horowitz
Hmm. What do you think? Itās not clear to me that Iām anything more than a statistical model. It is quite possible to predict how I will react in most situations if you have the training data of how Iāve reacted in every other situation before this. I donāt think Iām particularly surprising in the way that I go about my day if you know me. Yeah, but you probably want ice cream sometimes. Like literally experiencing the emotion or the of thought of, I want ice cream, and itās more than just, it is common for me at this time to say, I want ice cream. Yeah. This is going to be a hard thing to test. IEmbracing Contrarianism and Motivation at Y Combinator The speaker highlights the importance of solving a personal problem as the motivation behind their project, which became more evident as they shared their idea. They realized the need for a better way to develop software efficiently instead of wasting talented engineersā time. Despite being discouraged by other companies and receiving counterarguments, they persisted. One counterargument was that customer needs vary greatly, making it challenging to create a universal product. However, the speaker remained resilient and focused on their vision.
Ben Horowitz
Just like, Iām trying to solve my own problem?
David Hsu
It was mostly the latter, but then the former became more evident as we told more people about what we were working on. And I suppose maybe thatās the best merger of the two in the sense that what is obvious to you is contrarian to others or something like that. It was obvious to us, having built so many internal tools, that clearly there has got to be a better way of building all this crud software. There is no reason that we should be hiring talented engineers to go waste their time building this stuff. And yet, I vividly remember when we were in YC, so we went through YC, I think in 2017, there were multiple other companies that came up to us and said, hey, this idea is not going to work. In fact, we tried it before, and itās not going to work because itās going to be a consulting shop, because youāre going to go build this application development platform that no oneās Going to know how to use. Itās going to be really complicated. And the only way youāre going to get people to use it is if you use it yourself and then you basically become a consulting shop. And thatās what happened to us.
Ben Horowitz
And the classic pitch there is customer needs are so different from one another that itās actually very difficult to build a product that suits all of them well without ridiculous amounts Of customization. So no product opportunity.
David Hsu
Exactly. Yeah. So that was one counter argument. Another counter argument sounds actually pretty smart when you say it is, so at this point, weāre already selling to developers.Challenging the Notion of Verticalization in Development Tools The speaker believes that the traditional objections to retooling, such as the complexity of building something that caters to everyone or the allure of democratizing programming through no-code solutions, are not relevant. They argue that tools like JavaScript and React are not verticalized and are used across various industries, illustrating that a developer tool can be successful without being verticalized. The speaker also dismisses the idea of democratizing programming through no-code solutions, stating that code is the most effective way to build complex software and that trying to reinvent programming visually ultimately requires learning traditional coding languages like Python or JavaScript.
David Hsu
Something but not working on the retool idea. So, you know, they have good intent. But I think understanding sort of where theyāre coming from and where you agree or disagree with them and why, hopefully, theyāre wrong is, I think, very important. And so in our case, at least the former objection to retool, which is itās so hard to build something, you know, that sort of suits the needs of everyone. Well, if you look at JavaScript, for example, or React, for example, React is not verticalized. JavaScript is not verticalized. Programming language is not verticalized. Use them for everything. Social media companies use them. Banks use them. They all use JavaScript. And so for us, we said, well, weāre selling to the developer and weāre going to go build a developer tool that where the vertical does not matter. And so for that reason, we are sort of an objection, we understand it, but we donāt think itās relevant. For the second one, similarly. So in our case, we said, well, yes, we can see that democratizing programming sounds sexier. However, we think fundamentally, thatās actually broken. And soally, I donāt believe in no code for this reason. Because in order to go build complex pieces of software, code is the best way to get a computer to do something. Yes, I completely agree. It is so specific and so concise. And if youāre trying to reinvent programming and then in a sort of visual way and then trying to get non-engineers to learn that language, they might as well have to learn Python or JavaScript To begin with. And so, yes, I think the feedback is valid, but we had reasons for not believing it.
Ben Horowitz
This is such a good point.Incremental Advancements in Programming Languages In the last 40 years, programming has seen major advancements from punched cards to assembly and then to compiled languages which led to significant productivity enhancements. However, in the past 30 years, programming languages have largely remained stagnant, showing minimal innovation compared to other sectors. This stagnation is reflected in the productivity of engineers which seems to be trailing GDP growth. The idea behind retool is to introduce a higher level approach similar to the leap from punch cards to assembly or assembly to higher-level languages, signifying the need for further innovation in programming.
David Hsu
Really doesnāt have to be that way. And I think if you look at the history of programming over the past 30 or 40 years, for example, I think there have been some major advances, letās say 40 years ago. So going from punched cards to assembly is actually a really major advance because if youāre punching cards and you actually punch the wrong hole, you run it overnight and you come back The next day, youāre like, oh, crap. Well, okay, let me do it again. So the iteration speed is like a day. Right. Or going from assembly to compiled languages. Totally. Itās a real 10x increase in productivity. But if you look at the past 30 years or so, the languages are kind of the same. JavaScript has obviously gotten a little bit nicer for ES6, but itās actually kind of the same language as 20 years ago. If you look at C or C++, we still use C++ today. And the libraryās a little bit better, frameworkās a little bit smarter, but weāre kind of still doing the same thing. And so if you sort of graph like productivity of engineers over time, basically, itās probably actually trailing GDP growth. And so itās surprising that engineers are so inefficient, basically, and thereās been relatively little innovation in programming at all. So we hope that Retool is sort of like a much higher level way. Itās almost going from punch cards to assembly or assembly to a higher level language, basically. Thereās another jump left, basically, at least. Thatās why weāre pretty excited about Retool. Is it still true today that RetoolAdvancing from Low-Level Programming Languages Transitioning from low-level programming languages like JavaScript to higher-level programming with tools like retool and incorporating AI can lead to a significant jump in efficiency. Co-pilot enhances the programming process but is limited by the low-level nature of JavaScript. By aiming for a 10-20X improvement by moving towards higher-level languages coupled with AI integration, the programming experience can become more intuitive and efficient.
David Hsu
As fast which is pretty good but itās not really an order of magnitude jump and i think the reason for that is itās kind of like letās say you had an AI that would go punch cards for you, for Example. Thatās pretty good. You can still program, you know, faster than punching cards yourself. But fundamentally, if you as a human go read the program, youāre like, oh, itās kind of hard to understand. Iām going to try to reason the hole is here and we punched this hole. So what does that mean? Itās kind of hard to reason about. And same with assembly, right? Letās say that, for example, you had Copilot for assembly. Thatās great. You can go write assembly letās say twice as fast now, but youāre still writing assembly. If youāre trying to debug assembly, itās still pretty difficult. And I think thatās kind of where we are with JavaScript, which is JavaScript is still too low level of a language to be using to go write your programs. And Copilot does speed that up, which is great. But I think where weāll see hopefully something like a 10 or 20x jump, maybe you see a 10x jump, for example, by going higher level with Retool, and then you see a 2x jump on top of that, for Example, by incorporating AI into Retool. We think that programming still is too low level, and we want to still make it much higher level. And then you can graph AI on top of that, which will be another 2x on top of that.
Ben Horowitz
Changing gears, I want to ask you, what is an area or aThinking and tying things together is still too low-level if the goal is to bring together thoughts. The fact that humans need structure at all to get going suggests we are not utilizing our biggest cognitive asset: inspiration. The next leap in knowledge tools may be removing structural scaffolding rather than adding more of it.Sustainability of Business Growth Becoming too reliant on venture capital due to excessive spending can jeopardize a companyās independence. Leveraging software in a SaaS company can lead to significant growth with minimal staff, creating a sustainable business model. Maintaining a positive cash flow from the start can enable long-term success and reduce the need for additional funding even during rapid revenue growth.
David Hsu
Think it was mostly a sort of sovereignty thing or something like that, like in the sense that if you start burning a lot of money, you become basically beholden to VCs because itās very Hard to reverse, I think, the sudden burning of a lot of cash. You could do, you know, a 80% layoff, for example. I think your companyās kind of toast if you do that. So thatās, you know, pretty difficult. So that was one of the main reasons. The second was I just didnāt believe that there was that much work. In the sense that with a SaaS company, I think whatās really pretty phenomenal about software nowadays is that itās so leveraged in the sense that you can get to probably, I think, in Almost any industry, probably, I think, one mil or five billion ARR with basically a team of maybe five or ten people, I think. And if youāre able to do that, I think one maybe underrated phenomenon is itās actually very hard then for you to actually burn that much more cash in the future even. So sort of you set yourself up for a lot of success. Because letās say if revenue is growing 3x year on year, for example, itās actually pretty hard, in fact, to grow your company for the 3x year on year. And so what happens is if you start off in a, you know, roughly cash flow positive state, so letās say a few million dollars of revenue, maybe less than 10 people or something like that, Then as long as revenue is growing pretty quickly, you pretty much will just be cash flow positive for the next two, three, four, five years, actually. So I think- Right.Limitation of Language in Communication Communication between humans through natural language can lead to misunderstandings due to the limitations in translating complex thoughts into words. While written instructions are precise for programming computers, verbal communication can trigger different emotions and interpretations due to the low-bandwidth nature of language, resulting in potential misunderstandings and inefficiencies in conveying ideas.
David Hsu
And I think it will lead to great outcomes. But it is definitely not the only tool. And this, I think, itās back to sort of, I was talking to Ryan Lucas, our head of design recently, and his perspective was, when we talk to each other using natural language, we already, Between humans, have so much misunderstanding.
Ben Horowitz
And if that is the only way you have to program a computer, weāre going to be in a world of pain.
David Hsu
Itās funny.
Ben Horowitz
I had a very similar conversation recently where if two engineers send each other a pull request, it is exact what the communication is. Like it is literally saying these are the instructions to a machine. Can you look at them to ensure that they are doing what this contract says that it should do? But if two people are having a conversation, at least this is my understanding, the brainās capability, the spectrum for information in the brain is much higher bandwidth than any Given language. So when I translate a thought into English, it is going through severe compression. And I am communicating that to another person for which they could misunderstand it. Like a word could trigger a different emotion for someone else than I mean when I say it. And then they go through a decompression where they extrapolate that onto their brain. And it could light up a completely different set of neurons than was originally lit in my brain. But we only have this lossy low bandwidth medium to communicate through with language. And
