How Harness runs 16 “startups within a startup” at scale | Jyoti Bansal (Co-founder and CEO)

Jyoti: Early on when you think of go-to-market, people talk about product-market fit. Sometimes they say you should think of product-market sales fit. But if you have not figured out how to sell it, you may have to redesign the product.

Brett: For today's episode, I'm sitting down with Jyoti Bansal, founder and CEO of Harness. In one of the rare examples of a second-time founder who's built not one, but two enduring enterprise software companies.

Jyoti: If someone asked me, as a second-time founder what you learn, I said, the main thing is to not go in the assumptions at what worked the first time will always work the second time.

Brett: After selling AppDynamics to Cisco for 3.7 billion, Jyoti could have done anything. Instead, he dove back into the startup trenches this time building a platform to automate software delivery for engineering teams. What stands out most about Jyoti is how systematically he approaches company building, from building a go-to-market motion all the way through scaling up the sales team.

Jyoti: When I was doing AppDynamics, until the first 10 million of revenue I didn't care about go-to-market. As we built the best product go-to-market will happen. And then I realized to scale, really, we need to be equally excellent and go-to-market as we are excellent in product.

Brett: In our conversation, we talk about why starting with enterprise customers has been an advantage, and how to diagnose whether your challenges stem from product or go-to-market issues.

Jyoti: If someone comes to me and say, "This product is not selling and go-to-market is not doing a good job," the first question I ask them, "How many happy successful customers do you have?" We don't really have any happy successful customers.

Brett: He also shares how he thinks about founder-led sales, what it really means to take go-to-market seriously from day zero, and what separates an exceptional sales leader from an average one. Let's dive in. When you look broadly at the software ecosystem, there's a lot of companies that start mid-market and struggle to crack enterprise. Do you think it's normally that the problem is just different in enterprise, that it's surprisingly hard to get go-to-market right in terms of transitioning from mid-market to enterprise. Do you have a theory as to why a lot of companies tend to get stuck in the mid-market?

Jyoti: It's almost like over time you build certain muscles at a company. The go-to-market muscle in mid-market is around... PLG is very important to it. You have to have the high velocity. It's mostly inside sales. The heavy-duty enterprise sales folks will be too expensive for it. Now you've built your culture and muscles around it, for now you want to change to something else or build something in parallel that's just organizationally hard. That's the one part that I see, you build all the machine around, it's mostly marketing-heavy, mostly PLG, you have cheaper, younger inside salespeople. Enterprise will be more sales-led. You have partners and channels and expensive enterprise salespeople, longer sales cycles. There's a shift in how the financial planning will happen in your company. And with the CFO will be onboard with the salespeople, what you celebrated before, how the product managers work with go-to-market. Those are just a company's DNA get created over time. The second is the product. The products that are designed for mid-market and products that are designed for enterprise have certain differences. In enterprise, you need to hide... You want lots of capabilities and lots of flexibility. Lots of capabilities and lots of flexibility could become a complex user experience in mid-market. What's your strength in enterprise could become your weakness in there. Versus, also applies like in mid-market you designed a simple product with less capabilities, but now it doesn't work and do the things in enterprise. It's possible to do both. It's possible to do both, but you have to design the products in the right kind of way where you remove the things that are not needed in market, you price in a certain way. But it's just hard. For enterprise-focused companies to go mid-market is also hard.

Brett: You very rarely see a company that begins by dominating enterprise be able to go down market.

Jyoti: Yes, yes. It's very, very hard. But that's why at Harness one of the lessons I learned was I wanted to do both. I wanted to set a charter for let's design for both. There's no reason why we can't design our product where it works in enterprise, it has all the capabilities, it has all the flexibility, but it also has the ability to simplify and work in the mid-market. For two reasons. One reason is I didn't want the mid-market companies to come in enterprise at some point in future and compete with us there. Second is, I do think that designing the simplicity of the experience helps an enterprise as well. Because if you want do your... Even enterprises you would still do some PLG. You still want people to get going. Even someone who's already bought your product, for them to adopt, you still want the simplicity and the virality that comes from it. There is an advantage of, as an enterprise-focused company also to not completely let go of, okay, we are so heavyweight and so complicated. But that becomes our issue. You want to get almost the accountability that comes from by selling into commercial. We call it commercial space with the mid-market space. We are at 80/20 right now. I would actually like it to become 70/30. We don't need it, but we just think it actually creates the right discipline on how the product should be designed.

Brett: How did it translate to you think about the first year of the company and building Harness, with this idea of primary/secondary you wanted to be able to play in both? What did it actually look like tangibly in terms of how you approach building the business?

Jyoti: Well, the large enterprise you won't nail for some time. The product won't be ready. So you really start with mid-market anyways. At Harness, I knew that I wanted to sell in enterprise quite a lot, and you start having those dialogues, you start having dialogues with the large banks, your City Banks and Barclays and all kind of companies. But you're still mostly selling into the mid-market who are more fast-moving, they don't need so many capabilities, they can move fast, all of that. But by working with them, you're getting the product to be ready with it has to work an enterprise. And some point you start nailing your first enterprise customer, and then you nail your second. And that's what happened in practice with us.

Brett: What was your first enterprise customer?

Jyoti: I think one of the banks, Charles Schwab.

Brett: And how long into the company's life did you land them?

Jyoti: They were early, that were two years in the company life. But that was, they were one and then you get more, then it takes more. And now it's like we nail enterprise customers every week. But early in the company life, I won't say we only sell into enterprise only. We would starve otherwise, so we build the product. We don't need to build the... Again, it's about this is the path we are on. But on the path we don't need to wait for the full path. You find customers wherever you find, really, and most of them will create the pipeline in enterprise will also create a pipeline in mid-market. The mid-market pipeline will close faster. The enterprise pipeline will close much, much, much longer. But you start closing one and two, and then you start learning from them and the product starts get to mature and then you build all of it.

Brett: Were you particularly opinionated about your first three or five or 10 customers when building Harness? Or broadly speaking, anyone in mid-market that needed a solution that you had was fine to get going?

Jyoti: I don't believe I'm being opinionated too much on it. It's like, you want the right use cases though. You have to be opinionated on you are building the product that would be broadly needed, not by what this one customer wants or this two customer wants. Something like that, right? That's what for sure, but not who the customer is. Our first customer was online retail company. Second was a speech translation. And some of those companies become bigger over time as well. They grow from there. In AppDynamics, our customer number three was Netflix, and Netflix was just transitioning. And it sounds like an old time ago, very long time ago now. They were just starting to transition from DVD company to a streaming company. They were, in terms of the tech footprint, was not big. And they were our customer number three. And then they suddenly like, okay, we are going to, boom, completely in the cloud and become the online streaming company. And when we started working with them, they were like 300 servers, that's a total thing, and they became 3,000 servers running in the cloud in a couple of years.

Brett: Did that help you mature the product rapidly-

Jyoti: Definitely.

Brett: ... because you didn't want to lose them?

Jyoti: Definitely, yes. It helped us mature the product very rapidly. Some point of time, and that was hardest decision and I probably ever made, we decided to lose them. They were so much bigger than our customer number. They were customer number one in terms of their size, and our customer number two to 20, the mind were not their size. And we are producing a lot of revenue and the business was growing. And so much our effort was going into scaling and making it work for them, that at some point we had the conversation and dialogue with them that we have to let you go. It was probably the only time I had to let go a customer like that, but that was the right business decision for the company.

Brett: How did you figure that out?

Jyoti: Just the engineering time. I couldn't get anything else done in the engineering roadmap. At that time we had 200 customers, and they all have the requirements and things, and feature requests and things that they want. And then we had Netflix's number one customer. They have a lot of requirements and things that we have to do. At some point it became it was only for them in our case at that time. And that's where it was clear I cannot serve the other customers, it's just very hard choice to make that we just cannot serve other customers supporting that one customer. Which was such a great customer to have, and to lose them after three years was a painful choice to make.

Brett: What does it look like to take go-to-market seriously from day zero? Oftentimes, before you really have a product built and all that type of stuff.

Jyoti: In the very early on when you think of go-to-market, you think of people talk about product market fit. Sometimes I say you should think of product market sales fit. Which is, if you design a product which is what problem you're solving, what is your solution? That's a product market fit. But if you have not figured out how to sell it, it will eventually you may have to redesign the product. If you're going to sell it to enterprise or mid-market, you're going to sell it through PLG or you're going to sell it through sales led. How you design the product is very dependent on how you're going to sell it. If you don't think about it like the product is solving the problems, your go-to-market is going to PLG, and you didn't really design for that experience, that's not going to work. You're going to sell into large enterprise and you didn't design it properly. To me, figuring out what the go-to-market would be is very, very key. You have to trade through that also. The product market fit iteration also need to figure out the go-to-market fit iteration. You just have to deliberate about it. At Harness we were more deliberate about it, like, okay, this is what we want to build out. We knew what we need to build out, otherwise it won't be sellable. We can build a product that works, but it won't be sellable because it's not ready for our sales force to scale through that. Talk about, how do you become a more scalable go-to-market organization? When you convert that into your product and you have some sales that can scale, you can go from 10 units of sellers to 50 units of sellers to 500 units of sellers, there's a scalable machine that you have to build out. The more you put in, the more will come out because you have a repeatable, predictable, and go-to-market sales machine.

Brett: Say more about how you go about doing that.

Jyoti: Well, so much goes into that. A lot of it is the fine-tuning the kind of sellers you will have. Fine-tuning, what does the demand generation funnel look like? What comes on the top? What converts at every stage? So you know if I get 5,000 leads of this kind, they will convert into this many demos for our sellers, and these demos will convert to this many opportunities of this kind, and this many will go to this kind of close rate. Now, at some point you can predict based on how many leads what you're going to close in six months because this is your sales cycle, you can start planning your business in terms of how many new sellers you'll hire, how long it takes them to ramp, how many you iterate, what's the sales capacity that comes from it, what's the demand generation capacity that comes from it. Then you know, okay, we want to go from 5,200, you know what you need. Otherwise, if you are... We are going to go to 5,200, you have this many sellers today, and we are going to close business in these accounts and deals. That doesn't work. To me it's like, a scale sales organization is when you're not talking about deals, you're talking about capacity. The deals will happen when you have the right capacity. When you say, okay, I have 20 million a quarter of ramped sales capacity, so I know I'm going to close 90% business of that, roughly 18 million. Without even thinking of what deals you have for 18 million. Then you have your predictable system. Of course, at the end of the day you have some deals you have to close, but you can get to a point where you don't need to think about what deal whatnot.

Brett: And call it the one to five or five to 25 scale revenue chapter of the company. How do you think about diagnosing if there's a fundamental product problem, or if there's a go-to market kind of distribution problem in the business if it's not growing in the way that you want?

Jyoti: I think a lot of it comes down to the product market fit definition again. My definition of product market fit is always not that we have sold the product, but did we deliver the value to the customer after we sold it. For most enterprise software products, over time the rule of thumb that I've learned is that if you have delivered value to 25 customers, you have a very strong product market fit. Not just that you've sold 25 customers. When you sold, you said this is the problem we'll solve and this is the ROI you're going to get, this is the benefit you're going to get. And three months later, six months later they bought your product and they have got that ROI, and you've successfully implemented and ensure it. If you don't have successful, happy customers, it's most likely a product problem. It's not a go-to-market problem. If someone comes to me and say, "This product is not selling and go-to-market is not doing a good job," the first question I ask them, "Okay, how many happy successful customers do you have?" And it's like, "We don't really have any happy successful customers," or we have three or four or five, product is not ready yet for go-to market to really really scale on it. That's where I look at many times that is the founder-led sales job is to deliver not just the first 25 customers but also the first 25 success stories, or the successful outcomes that come out of it. A lot of the product maturity happens in that process, not in the selling processes. The delivering the value into those. Once you have those, most products should be able to scale. If you have those happy successful customers and go-to market, it's not scaling after that, then you have to feel like, okay, did something change since you found that product market fit, the competitive dynamic to everything? Part of the most likely it's a go-to-market problem otherwise.

Brett: The definition of a happy successful customer, is that generally easy? It's visceral, you talk to the customer, "This is exactly what I needed."? Or there's some, with more precision it looks like X or Y?

Jyoti: To me, just comes down to one thing, is the value delivered? And the value delivered perception in there... It's like if you ask a customer, "Are you happy?" Sometimes they might be happy but they haven't gotten any value out of it. It all really comes down to, have you gotten the value you thought you will get when you bought us? In Harness, we run a very structured sales process where we will create a assessment before we sell. Which is, this is the value you're going to get. If you buy Harness and using Harness for continuous delivery, your deployments. We do an assessment of right now you're deploying once in two weeks, you can become once a day. Your engineers spend this much time per deployment, like six hours per deployment, that can come down to 30 minutes per deployment. Your failure rate is this much today, a 20% deployment failure rate, we can bring it down to 4%. You define outcomes in the sales process. And then after once you sell it we work with the customer on, have you achieved those outcomes? And have you not achieved those outcomes, you haven't delivered the value that we promised. Even if you're not that structured around it, I really think the happy successful customer definition is very simple as, have they achieved the value outcome they thought when they bought it? And if they have, then most customers are happy if that happens.

Brett: In the case of Harness, when you think about building out the early go-to-market team, is your philosophy you will hire a senior go-to-market leader, they build a team? You hire more junior people, you begin to scale and then you put a senior leader on top of them? What's the order of operations? Do you have a point of view on it?

Jyoti: I tell every early founder, if you don't have experience managing salespeople, try not to learn as on the job for the first time there. If you can, you have the money and ability to recruit a senior person, do that. Because you're de-risking, you're learning and a lot of things that will come in. Normally, I look at how the most senior you can attract and you can afford. If you have constraint on cash early on, which a lot of startups are and you don't want to, then it's a different thing. But if you have the cash and you can attract someone, I would look for a director, VP sales, someone who knows how to hire the right people, the reps, how to run a sales process, all of that. If you have experience managing salespeople, it's okay to hire a few reps and then bring the manager. I'll tell from my own experience. AppDynamics, when I was looking to hire my first sales rep, I still remember I was someone who was like, okay, you should hire some sales reps, you don't need a VP of sales. And I started interviewing those reps, I couldn't even figure out how to interview them. It's like, what do I ask to really... Because they-

Brett: And you're an engineer.

Jyoti: I'm an engineer. I couldn't really figure out, okay, is this person the right person or not? What to ask in the interview. And that's when, okay, I'm going to probably... I should hire a VP of sales who knows what to do, and I can learn from the VP of sales to do. And so that's what I did. I was able to attract a VP of sales at the time, and then the VP of sales brought in now, started to build it out and all that. At Harness, actually I did decide that I know now at AppDynamics I worked with salespeople all the time. I know enterprise software sales and all that. I had enough skills that... And I could attract head of sales where he's in Harness, but I actually wanted to have a couple of reps just work with me initially. I felt like I need to be close, very close to the customers in that. I was still in the transition from founder-led sales very early on. At Harness, to me it was very clear that I need to be... The second time founder, I was a bit paranoid about that. I don't want to be too far removed. At AppDynamics, right before I was leading a org of a thousand people, it was my last job there as a CEO. And now you come in and just completely on the ground at five people, and that's a shift. And then you may get into, okay, I need these layers of people to do things. And I was very paranoid about not doing that, that I need to go on the ground and not bring layers. Actually, Harness, I initially didn't hire. Until the first million I didn't hire head of sales. And I said, okay, I'll just hire a few reps. I know how to manage them, I know how to interview them, I know what to do there.

Brett: One of the big points I think you're trying to make is that the way the product is delivered and the way the product is built is one and the same and they have to go together. And so, when you're selling it and you're close to the customer, and you're building the product, you have a very tight feedback loop between those-

Jyoti: And you can compare it very fast and kind of... Yeah.

Brett: When you do think about hiring your first head of sales, do you think there's some special things you're looking for in your first head of sales that's different than an excellent at scale sales leader? Or a great head of sales can go from one to 50 million or 50 to 500, or that type of thing?

Jyoti: Most of the times it would be a little bit different skills. Someone who would go from one to 50 and someone who go from 50 to 500 might have most likely different skills. Sometimes you may find someone who can do all of it, and that happens too. If you're hiring your first head of sales, I would look for someone who hasn't been too far removed from a first line or a second line manager. If they come in from a place where it's like a 300 million sales org, there are six layers between them and the sales rep, and they have been like the last time they had it was eight years ago before they were close to the customers. It's an adjustment, and you're taking a risk with it. I normally would recommend people don't hire that senior of a person trying to bring... It's hard for them to adjust down as well. Normally, a first time head of sales I normally look at what are we forecasting as a business side. Let's say half a million in revenue and you're looking at your first head of sales. Or a million in revenue looking for head of sales. And you're projecting that for a million in revenue in the next three years I want to get to 25 million. At 25 million revenue you normally would probably need a team of maybe 15 sellers. I would normally look for someone who has experience within that managing teams of 15, 20 sellers, and delivering outcomes of it is a good person. Because at least you have the next three years. And now you're taking a bet on, can they scale from that? And if the scale is great. If they can't scale from it, then it's, okay, you bring someone maybe on top of them or something.

Brett: Do you have any reflections on when you think about the top three sales leaders that have ever worked in your org, what's different about them than all the average sales leaders? Of which I think there's by definition a lot.

Jyoti: No, I had the fortune of working with really, really, really world-class sales leaders, and including now in Harness. I'll tell you that the number one thing that they do extremely well is recruiting. And it sounds like you're talking about sales, why is recruiting the number one thing? And all I've found is eventually the sales scaling comes down to ability to attract the right people. And people who are very good in recruiting or want to put the time and energy in recruiting the right people, they scale very well. The second part after that, which is tied to recruiting, is the enablement. If you bring people in, can you understand what it takes to enable people well? The value of that, the ramping the people. If you don't enable them well, good people are not going to make money. And if they're not going to make money, they will leave. And now you don't have the sales capacity that you need. Then the third is a disciplined sales process. The sales process is not very loose and how you generate pipeline. How do you evaluate every part of where you are in the deal? How do you look for what is the right pain? There are many frameworks around a structured sales process, but you want someone who believes in it and who can run that in a disciplined way. Otherwise, the predictability of the business goes away. A sales rep comes to you and we are doing this POC and the things are going very well. Suddenly, oh, we lost a deal. Why did we lost a deal? Because there was another person who was pitching for competitor we didn't know about. That is a bad sales process. In a structured sales process you will try to find that early on. You'll try to build the right champions in the account so you don't have surprises. A great sales leader will do that well. The fourth is really people who have the mindset of finding a way to achieve success. In sales, the world of sales, the excuses can build up very fast. Many times those are valid reasons. The best sales leaders in my mind always take extreme ownership. Extreme ownership and pride in their ability to achieve the numbers. At any level, if I interview a salesperson and they're not very clear... If I ask them normally, okay, how did you perform on your targets in the last 10 years? And they don't know it easily. Best people will have in their resumes even. 2020, I achieved 110% of my quota, '21 I achieved 109% of my quota, '23 I did this, whatever. Right? The reason I like that is not just because they achieved more, but because they are anchored as a measure of success around that. Do they take pride in it and do they consider that you want to achieve your goals and targets is the thing you take pride in? People who don't, you ask them, okay, how did you do, and they all start speaking in abstract terms. I know I want a salesperson and say, "I achieved 100% of my target, or 70% of my target for these reasons." But someone who doesn't even know their target or talk about that is normally a bad sign. What I've seen is you recruit very well and you are almost obsessed with the talent that you can recruit. You enable them well. Enable could be at the rep level, enable could be at the manager level, at the leader level, all those. You know run a tight discipline sales process. And you take extreme ownership of the number, and you drive that ownership of the number across the org, those are the best ones.

Brett: You mentioned this in the last thing that you just said. When you're interviewing a sales leader, if I were to sit down and watch you spend time with the person, what are you doing? What are you asking? What are you probing on?

Jyoti: It depends on different stages of the company, but I really interview along those lines. Okay, we have to hire five reps or 10 reps, or 20 reps, whatever it is. And managers and leaders and all. How would you do that? And what's your prior track record of hiring and attracting the right people? There's leaders who make the hiring recruiting easy and proven that they have done it well, and they are this magnet for good sales talent, that's [inaudible 00:23:51]-

Brett: Is that a big part of it is that they're the kind of person that people just follow?

Jyoti: That's a very big part of it. Follow for multiple reasons, like they're a good leader, they are fair in the how to do it, good people want to learn from them. All kind of reasons. But, yeah, that's one big part. The second is, I like to see do they understand a disciplined sales process or not? Have they done it in the past? Many times they've worked in organizations where there is a disciplined sales process, so you know if they're spent five years in this company, they likely know the disciplined sales process well. But then, I really would interview around that. Which a lot of it comes down to, tell me about the deals you lost, tell me about the deals you won.

Brett: And you want to hear the sales process come through.

Jyoti: You want to hear the sales process and how they even talk about it. When they talk about this is a deal we lost, what is the terminology they even used to talk about that. Or this is the deal we won, how they talk about it. And the third thing I talk about, tell me about when you didn't meet your numbers, and why? Or when you beat your number significantly, why? And do they understand it? And people who don't even care about their numbers, to me that's a problem. But when they care about their numbers then how they architect towards that number. I was in this job last year and we did 70% of our number. Do they understand the mechanics of why did this did 70% number? It's okay that you did 70% of the number, but at least what happened and you understand what the full dynamic of it. You did 130% of your number, and what happened and why not? And do they think about architecting the sales machine towards the number? Those are the things I would normally interview for.

Brett: Do you care if they've sold similar things to similar customers, or no?

Jyoti: To some extent, but not... It depends on the definition of similar. In most enterprise software, this is my learning, is you can put into three buckets in terms of sales experience. One is business applications, CRM and all that. And there are very good sales people who are very well-trained to do that. They worked at Salesforce or someone for a long time. The second will be the enterprise infrastructure folks. Enterprise infrastructure, you're selling to IT and technical people. And the third would be hardware oriented things, like firewalls and all that. And I do care about that you have... When I'm selling an enterprise infrastructure I've seen, yes, people coming from business applications can succeed, but there's a risk there. People coming from hardware selling can succeed, but there's a risk there. These three buckets are broad enough. I don't need to be an infrastructure software that someone has sold a CI/CD product, doesn't matter to me. If they have sold anything infrastructure software, that's good enough. They know how to sell to IT and how to sell to infrastructure people, and technical people and developers and all that. That's normally I would look at. I do look at SMB versus enterprise for sure. The people who are only experienced in enterprise selling, and you want to sell to SMB, they will struggle. And same with the people who only sell SMB, they will struggle. That you definitely want. If you want your business enterprise selling, you need someone who has experience in enterprise selling.

Brett: Have you noticed any patterns in hiring go-to market people from the third-best company in the category versus the first? Do you like the people that worked at the premium company in the category? Do you like people who had success selling the third-best product? Do you care about that in any way?

Jyoti: It's a good question. I remember when I was hiring one of our sales leader in AppDynamics, became our CRO, and very successful many companies after. In the interview process was telling me we had product number five in the market and we're doing well. I normally don't look at completely like that, but yes, there is something to it. Normally, what happens is the number one company is number one not just because they have great product, but because they also have a very strong sales culture and DNA. Most of the time you do want someone who has spent... So they know what good sales culture and good sales discipline means. I'm not biased against number one company because of that. If someone is on number three company and they've done well because the product was not very good and they've sold it, there is value to it. I would rather take someone who's coming from a highly disciplined sales culture, because then we don't have to teach them that. Because they already know, they've spent three years, five years, seven years in that highly disciplined sales culture.

Brett: One of the thing in talking to you that's very clear is you either always had or you've developed a lot of commercial taste. And as someone who's a classically trained software engineer, I think it's an attribute that tends to be under expressed. Do you think you honed it and developed it? Both the nose for value, like spending time with a customer and understanding is there an opportunity here, all the sales and go to market stuff, do you think you just developed that sensibility or you always sort of had that as something that was a part of you?

Jyoti: Before I became software engineer, or before I went to school to study computer science, I grew up in a small town in India and my dad had what you'll call a mom and pop shop of selling irrigation machinery to farmers. And in that small town we didn't have too much proper job, so everyone almost had some small business that they will do around my family. I was going on my dad's shop since I was six or seven years old on weekend and after school to help him on the shop. Anytime I sell... I learned business before I learned engineering, but that was just some exposure to the basics of the business. I studied computer science, became software engineer. Then when you start as an entrepreneur you have to hone it, you have to put your energy into it. Many times software engineers come to me, it's like, "Hey, you became from engineer to entrepreneur. What skills you had to learn?" Tell them is you really need three skills. There's a technical skills, don't underestimate and forget that, because that is your strength. There is the people skills, because you cannot build anything of size without being alone. And third is the business skills. Technical skills are your core, so you have to make sure you don't lose your core power, but you have to put deliberate energy in the other two skills, which is the people and the business. Yes, I had to put energy into refining and learning and refining and learning. And you do that over time, because that is really the only way. Many times I tell early founders is learn sales. They ask me, "What do I learn?" It's like, learn sales. And you learn sales by selling it. Sell yourself, sell the product. Don't be like, oh, I'm not good in front of customers, and I need to hire a salesperson to do it. There's no way you'll be able to manage and make that sales org work if you don't learn how to sell yourself. Yes, you need sellers to grow, but you got to learn how to sell yourself too. And understand the get the empathy towards what it takes, what do customers really want? Otherwise, you'll be building things that no one really cares.

Brett: When you started to develop the sales muscle, did you enjoy learning sales, or did you like doing product and you were forced to do sales and you did it because you had to do it?

Jyoti: In some ways, when I started AppDynamics, this was April of 2008. And I raised my first... That time the 5 million round used to be called a Series A round, right? For sure.

Brett: Series A is 5 million.

Jyoti: Pre-seed. Is the pre-seed, right? I did my series a 5 million in April, and in September the Lehman Brothers crash happened and everything was falling apart. There was no more funding for the next couple of years really. That almost forces you to succeed.

Brett: You got to have revenue.

Jyoti: It creates this very tight focus. For me, it's like we were a small team, like 10 people. And we were like, we got to build a product. But everything, do we survive this next two years is about do we get revenue? You become much more revenue oriented. And that was very, very key, just because that was the only way to come out of it. And it was no easy path. Other than revenue, there was no next round I could raise without it. I think maybe that was it, you force yourself to do it. You find revenue, find your way to learn how to get to revenue. It's very simple. I tell everyone is the only lifeblood of any company is revenue. If revenue is there, most things will kind of you'll figure out.

Brett: And if not? And if not, they won't.

Jyoti: You can build the best product, everything have the best, everything is... The revenue is not there... Ultimately, you need the revenues. I always liked that. I actually had one exec early on in AppDynamics who somehow thought that was a bad thing. He will complain like, "You are a very sales oriented founder. You are not being able to spend enough time on engineering. You should do more." Yes, I know how to build products and I want to, but the sales orient is a good thing, it's not a bad thing. I got to sell and make sure we get revenue, otherwise I'll be building things that doesn't really sell. And if you don't have revenue, nothing is going to work.

Brett: What was the founding story of AppDynamics? How did it actually get started?

Jyoti: I was working as a senior engineer architect. And another startup which had a first generation of application monitoring product, a company called Wily Technology, which was acquired by Computer Associates. I had exposure to the application monitoring-

Brett: You were there before they were acquired, and then after-

Jyoti: Before they were acquired. Before they were acquired. And I was there for a year after they were acquired. That was the first generation of application monitoring. And the reason, actually, I joined a startup because I always... As a developer, I'm struggling with troubleshooting these complex issues that are happening, and this company is doing something interesting there, so let me join them. And I like the problem space. This is when the cloud was just starting to come out, and people are building more and more distributed systems, microservices. That how you troubleshoot what goes wrong and fix things in these requires a very different way of doing things. And it requires a notion of a distributed application tracing and ability to trace between a lot of different things. I knew the problem is emerging, and a little bit of the industry knowledge because I was working in a company that was in the previous generation of that kind of solution. I was passionate about that problem. It was clear to me we need to next set of products and companies have to build to solve this. As a founder, what happens is many times you start getting into the strong conviction and you can't sleep. You are like, someone has to solve this problem. And then you start, okay, why don't I solve it? If someone has to solve this problem, you completely are convinced and you start looking at, how do I solve it? And that's when I, okay, let me start building this thing and start pitching to VCs and investors to get some capital. And I still remember I was still in the job when I was pitching, because it was a large company. And this one investor was... He asked me, "Do you really believe in this?" I said, "Yeah, of course I do." He's like, "Why are you still in your job then?" And I came home and I was like, oh, that's a good question. Next day I resign. It's like, okay, let's go and do this. And if I can't get started, I can always go back in an engineering job. Why not? I started pitching to investors, at that time it was just... For me, it's like we didn't have many investors in San Francisco at that time. I live here, I had to go to Sand Hill Road to do the track too, to pitch to them. Mostly, during the days I would go and trying to find investment, and during the night I will go and code.

Brett: How long did it take you to raise the first round?

Jyoti: About three months. And I got a lot of rejections. I got about 30 rejection. Especially, at that time it was a bit harder on you're a first time founder, you're just an engineer, you're technical, you don't have any business background. Some people will say you find a business co-founder, otherwise we can't really fund it. Actually, my first term sheet I got, it came with a condition that a business co-founder has to join as CEO. And they had a good guy who they wanted to come in, but I was not sure, I don't want to do a shotgun marriage like that. Even with a great guy, I don't want to. This market is too small or all kind of things. Now, observability people will think, oh, you have companies worth lots... Datadog and New Relic and Dynatrace and AppDynamics and Splunk, and a lot of successful companies, but back then we were [inaudible 00:34:57] observability. This was before with things we called monitoring. This is such a small market, it's a niche market. No big company could be built. All of this.

Brett: And why did you have such conviction? It just seemed obvious to you?

Jyoti: It seemed obvious to me. To me, the world running on software, everything you do is software. And I knew as a software engineer it's very hard to troubleshoot when something goes wrong. And the impact of something going wrong is very, very high. Things slow down on an application web app or a mobile app, people lose hundreds of millions of revenue. It has to be fixed, and there were no good products. And I knew there were no good products because I worked in one of the products who were the previous generation product. The conviction to me is the conviction on the problem. How can you operate a world on software if you can't fix or troubleshoot when things go wrong? Which, things go wrong all the time. I knew the current solution set, so that was there, but I also had conviction on my approach of solving it. I knew that, or at least in my mind, I was pretty convinced-

Brett: Technical insight of the-

Jyoti: Technical insight that I can do a distributed brace would be the solution for it. And maybe my insight would've been turned wrong, then maybe we didn't succeed. But that was also the conviction. And the insight was right, and it all worked out.

Brett: Earlier in your life did you say, I want to go start my own business at some point? Or that wasn't even-

Jyoti: No, no, that was my goal. A lot of times... I graduated from as an engineering school, on the top engineering schools in India. A lot of people would come in and do the US after to do a master's or PhD. And I was like, I don't want to do that because I want to actually start a company and build businesses. That's always something I wanted. That's why I was like, let's me go and work in startups so I would learn it. I had to also wait for my green card before I could start a company. I was like, let me work in startups while I do that, which was kind of what I wanted.

Brett: One of the many interesting stories about the company is you're about to go public, and then you ended up selling the company. What's your reflection on that whole set of decisions? And was that easy at the time, was it agonizing?

Jyoti: It was agonizing, for sure. We sold the company for $3.7 billion. It seems like a big number. Even these days, 3.7 billion doesn't seem like a big number.

Brett: It's still a big number. It's a big number.

Jyoti: But back then it was bigger-

Brett: Bigger than now.

Jyoti: Bigger than now. Many people think, oh, that would be such an easy decision to make. But it was not. We were on the IPO path, we were on the road show, we were about to ring the bell on Nasdaq on Thursday. We had our first 20 employees or so who are in New York for the ringing the bell and everything. And this conversation started three days before the IPO Cisco came in and said, "We'll pay you more than you the price you will list." And we said, "No, we'll go IPO." We said no. They came in with another price, which is like, "We'll pay you more than one and a half times of what you will trade at."We said no, and then say, "We'll pay you two and a half times of what you will trade at." At that point now it's like, okay, do we take it or not? Do we take the risk of what the execution that will come out of it? What's the right thing for the shareholders? Yeah, it was like a two or three days of nonstop debates and board meetings. And of what is the right thing to do? And we said yes. When we rolled it out on Wednesday to our employees who are in the New York already, it's like, yeah, we come back, "We are not going IPO." Everyone was so sad. We made a deal with Cisco that, okay, if we close this deal, can we still go and ring the bell? And they made it happen. The Cisco stock is a big ticker on Nasdaq. So with all the same people, we all went back a few months later and did it. In hindsight, when you look at it, we had one of the... I would say the strongest visibility product at that time. We also had one of the best sales organizations also. We're growing very, very fast. We're growing 65% or so at that time. And of course the revenue sizes those days at IPO were smaller, we are at 150, 160 million or so revenue at the time. Would we have built much more bigger platform? I look at the other companies in that space, they have done better. Yes, so we there. But the other part of the variable was also I would say not the best investor board dynamic, and that creates execution uncertainty on what will happen in the future, what we'll do. That was a factor in there as well. When I also learned the lessons from the first company to second company, to me that's very important as well. You got to make the right aligned set of investors and board. At Harness, you have this vision for building it for a platform for the long-term. But if I have investors who don't want to do that for the long-term, who want to do in something different vision for what the company should be, then you're misaligned and that creates a big execution, uncertainty and risk. Unfortunately, we had those issues in AppDynamics, and that was a factor in deciding to sell.

Brett: That note about alignment with your investors, is that much easier for you to understand the second and third time because you know all the investors? Or do you think you could actually start the whole thing over and build your first company, and there's conversations you could have before you accepted a term sheet where you could have figured out are we aligned or not?

Jyoti: Yeah, that's a great question. I don't think so, but I didn't know how to ask the questions at the time. I really had no idea what to ask when I was doing my earlier rounds of financing, so it was a little bit harder. But a lot of it comes down to what you want to build out in the... And it's not that I had full clarity. You get your investment and get the business going, you don't have too much choices and control. Sometimes you want to get the business working. And later on if it turns out it creates misalignment or sometimes the board dynamic is not the best, that does create issues. I'm very fortunate that... And I probably spend one to 2% of my time managing the board. And Harness, on maybe not even that. In AppDynamics, I had a broken board dynamic that I had to spend 25% of my time managing the board and shielding our team from the challenges that come with it. Because otherwise what happens is if someone wants to do certain thing and someone wants to do certain thing, and someone want to do a thing, and the business is flip-flopping from one to another, from a strategy that can break the company. I had to spend a lot of time to make sure that doesn't happen. At Harness, I deliberately have picked the people and build the board and the dynamic of it that I don't have to spend time. And I'm very fortunate that I was able to pull that off.

Brett: If a founder friend of yours is raising their first round, they've never done it before, is there any advice you would give them in terms of how to figure out if this investor or set of investors is going to be aligned with you?

Jyoti: Yes. One is, are they investing in because they believe in the vision and they believe in the founder, or are they investing for FOMO? Because half of the investors are just, oh, this is a hot deal, I need to chase it. And they don't really fundamentally believe in the problem, or you. That's one thing I say.

Brett: How do you figure that out?

Jyoti: You know in the process how much... Someone who's want to invest, how much conviction they have, do they have a point of view of why they want to invest in? Not just because you're suddenly a hot deal and that's why they want to invest. That definitely is one. But I also look at, do the reference checks and all on how they would behave in hard situations. What's their belief in the role of the founder and a board member? It's where the boundaries are. That's what I ask if a founder friend comes to me. Look at that and look at the track records. There is the talk to other former entrepreneurs they have invested in. It's important, because you're going to be working with them for a long time.

Brett: It's easier to get a divorce than it is to get someone off your board. What, in your mind, does a healthy, productive, useful board look like?

Jyoti: I think a healthy productive board would be the board that will challenge you, ask the right questions in the decisions that the executive team, the management team is looking to make, but then let you run with it. Not that they are trying to think that they need to make the decisions. And I always look at if the board believes that the executive team is not capable of making the right decisions, they should change the executive team instead of trying to make the decisions. That's the one healthy dynamic of it. Second is the board should provide the right degree of governance and oversight from people are doing the right things. Just that's to me is the board's job. Nothing wrong ever happens in the company. Nothing wrong is happening and you provide the right degree of governance and create some degree of accountability around that. The third is to bring in perspectives from a broader lens. Sometimes when you are executing, you are executing in the lens of what you are doing. That's what you believe in day-to-day, everything you see. But there's a broader lens of what might be happening in the industry that something a board can bring in the perspectives. But to me, the healthy board dynamic always is the board members will challenge you. You have a debate, dialogue, discussion. But there is a trust that they know, okay, I asked my questions, I challenged, and I see the arguments on why do something, why not do something. But I trust the judgment of the executive team on why they... Because they have the expertise, because they live in it every day to day. Most of the time board members don't have the domain expertise or the industry expertise or the expertise. They think they might, that's a [inaudible 00:43:34] problem.

Brett: They really have nothing.

Jyoti: They think they might. And these are very smart folks, but they're not living and breathing that particular problem every day. The expertise level that the management team will have will be much more than the board. But the board's job is to challenge them and ask them, have they thought through things or not, but you have to trust them to do that, do the job. That's very important. The other thing is, of course, for a lot of early companies you need help on things like recruiting and sometimes sales intros and all. People are willing to-

Brett: Do the work.

Jyoti: ... to do the work and help you. That's important because that makes a difference.

Brett: How has your thinking on competition changed in the last 15 plus years at building companies? Is it something you obsess over? Is it something you focus on? If you do pay close attention to competitors, how does it practically express itself in the way that you want to run a business?

Jyoti: I am more in the camp of, I don't really pay as much attention to competitors, but I don't obsess about competitors. You don't want to be blind about competitors. You want to know what competitors are doing and what things are happening, et cetera. But chasing competitors and being too obsessed by chasing them, you can just go into one direction to another, do this and that and not really... The lens of revenue and customer value delivery works much simpler because you could be very focused with that on we need to drive this much revenue. And how do we drive the revenue is by delivering value to customers. And how do we prove that? Can we constantly increase the value that we are creating by more use cases, by some use case, by do doing it better than we used to do before? That's what driving... And we are listening to customers very closely. We have a very tight feedback loop from our customers. If you have that, you know what you're building because you know that will create more value for your customers, and new customers and existing customers, and you are building for that. That, to me, is the primary north star. You don't want to be blind about what competitors are doing. You want to watch that, but not obsess by that. That's how we normally operate at Harness. We look at competitors, but ultimately, are we building the right things that our customers want? Will pay for it? Are they getting value from it? And would that allow us to sell more over time?

Brett: Building on this slightly, when you think about delivering new products to the market, if there are seven other companies that are doing something similar, does that inform anything? Or it really is, let's just solve the problem in the best way for the customer and do a world-class job selling the product and...

Jyoti: Well, you have to think about if there are seven other companies that are solving the problem. You just say, let's solve the problem in the best way possible. And if your best way possible is no different than other seven companies, or maybe even inferior than the seven companies, then it's not good enough. If there's a market that I'm building a product with seven other companies, the approach I take is we have to look at, okay, what are those seven companies doing? That's when you do want to have a competitive analysis very strongly. Out of those, what are the core capabilities that are table stakes to compete with those seven? And those are the core capabilities you have to have compete with them. And then, what are the defen... I call them binary differentiator capabilities on top of that. The rule that I like to follow is we are not launching a new product until we have, these are the table stake parity capabilities, and these are the binary differentiator. That we have at least one strong binary differentiator. Maybe it's two, but at least one.

Brett: What's a good example of it?

Jyoti: Let's say CIE. A lot of competitors in space, continuous integration, you have all kind of companies in the space for a long time. Product number two that we built out at Harness, so if you're building CIE, what would that mean? You start with, okay, what are the things that all the other CIE companies would have? A modern CIE company will have a declarative pipeline, it will have this, all the kind of things. This is the list of table stake parity features. But what is our unique binary differentiator that we built? In our case, we looked at what is the biggest pain with all these companies? Which just really comes down to the developers think the builds are too slow. And that they submit a code and it takes too long, it takes them 30 minutes to wait or 40 minutes to wait for the build to... And then we look at, okay, why are the builds so slow? And the builds are so slow because you're running a lot of tests. We created a technology about, how do you reduce the number of tests you need to run based on what you changed? Let's say you changed 50 lines of code, and you sum it up a PR for a CI system. And you most likely these days you'll have a suite of 3,000 tests that will all be executed. We created a AI model. This was, by the way, before LLM. When we launched our CI in 2020, called test intelligence, which where we'll learn based on what you changed, what tests really need to be run. Out of the 3,000 tests, we probably can cut it down by 80% to just 300 tests or 400 tests, 500 tests based on what you changed. We can significantly increase your productivity around it, you don't need to wait for builds. When we launched it, we had a very binary differentiated test intelligence that will do everything that your current CI does, but our builds will likely be 4X faster. And we will prove it. We have 4X faster because we cut down all this unnecessary testing with our test intelligence. But that was the binary differentiator. When we launched our product on our cloud cost management, which is another module in Harness we built. Cloud cost management was about FinOps. Like in any products in the FinOps space, for quite some time like Cloud Health, VMware Require and Cloudability, and all those kind of companies. But the problem in FinOps still a lot of cloud-based. When we build the product, okay, what are all these FinOps products do? The FinOps product are reporting, you're collecting data, and you're slicing and dicing the data and reporting and dashboarding on the data. That's mostly say the current state of the FinOps products. And we looked at our binary differentiator that we wanted to bring is, what do you do with the data by automating actions? Can you just bring automation actions to the developers so that normally otherwise in FinOps a report... Someone will look at the bill and a report, a FinOps person and then chase the engineers and developers say, "Hey, go fix this." Our approach was, instead of them chasing the developers, can we just give it an automated response the developers can do so it doesn't even happen? We launched with that market. It's the current status dashboards and visibility, which you have to do to replace as a parity thing, but you have to build the binary differentiators which are about the, how do you automate that? Almost everything that we do, that is the rule we follow on that. You need to parity as well. Because you don't have parity, you're not going to be able to replace something. Maybe it's not 100% parity, because maybe people have the existing seven incumbents, they have 50 features that only 20 matter, 30 don't really matter right away. You just focus on the 20 that really matter to get the right parity. But you have to have a binary differentiator, one or two binary differentiators.

Brett: What else may on this theme of launching multiple products? When you're getting a team together or if you were to explain your philosophy around, what are the things that we need to do to make the next product successful, similar to what you were talking about around a binary differentiator, are there other things that are part of your philosophy on when we are doing the next thing, this set of conditions has to be met or other similar things?

Jyoti: One thing I didn't talk about, which I'm surprised that I didn't so far yet, is this how we operate at Harness is what I call the startups within a startup concept. And this is a concept I actually started to implement at AppDynamics. We were doing it at AppD, and AppD was growing so fast because of that. We are building more and more products in there. At Harness, we have made it to a V2 of that startups within startup concept. What that means is when you build a new product, it operates as its own independent, semi-independent startup. There's some owner who's almost a startup CEO, a product manager who's leading that. And with the same kind of hunger and if you start a company outside, a seed-stage startup, you have 5, 6, 7 people and you have to go and build something and win. We almost like to create the similar kind of structure, the environment, how they do it. With a good degree of freedom around it also. You have a shared infrastructure, shared side of our platform. A lot of the enterprise code and all you will get for free because it's all part of the platform. After that, if you are the startup CEO, you do your founder selling, you find the product market fit, you do founder selling, find the first 20, 25 successful things.

Brett: Yourself, you don't get to use your sales team.

Jyoti: You work with sales, but you are responsible on sales. You lean on sales. You say you have a salesperson in an account, like a Citibank. You will say, "Hey, Citibank will be a great potential customer for this thing. Can you introduce me to the customer and then you will do the selling?" Not, our salesperson doesn't know how to sell this new thing there, but the product manager does. It's really the founder selling in the beginning to get that going. And we only start teaching the salespeople to actively sell something once we reach that bar. Normally, it's like in the 20, 25 customer kind of bar, that's when now the sales can run and take over the founder.

Brett: And it's the same thing, it's 25 customers and you've met the product [inaudible 00:52:00] .

Jyoti: Yes. Some things could be slightly lesser, some could be more, but the bar definition depends on. But it's roughly in the range that we have sold the product, people are getting value. Very large accounts that maybe you don't need 25, you need smaller numbers. We operate in that startup within startup concept. When we start a new startup, the definition that I set for everyone is we have to be best of breed in that area. We don't have to be best of breed on day one, but that is what we keep building. What does best of breed mean? To me it's very simple. Someone says, "What are the top three products in this thing? We should be one of the obvious names in there." Then we are [inaudible 00:52:32]-

Brett: For this job to be done.

Jyoti: Whatever that is. Like a CD, if someone says what are the top three products in CD? And people don't think of us. In our segment, say we're selling an enterprise segment in there, we probably are not there yet. Or, what are the top three products for feature flags? But that's we keep building for that we are just perceived by our target segments as among the best products in that. And we have those kind of binary differentiators. We have the right kind of capabilities, the right bar on the value that we are delivering to the customers after. We are focused on not unnecessary features, but the features that will deliver, then deliver an outcome. And that's the goal we set. Let's keep building towards if we don't stop until we get there. The thing about the seed model, at Harness we even built a internal funding model that's aligns to it. A startup starts at a seed stage where you normally have like five, six people, a product manager and a five, six people who are the product manager is the startup CEO, and you are five six people. It's not too different than a seed company outside. Once you hit a million in revenue, you hit a Aeries A stage, then you have a bigger team, you have more customers. We fund it like that. Once you hit 5 million in revenue, 5 million in AR, then you have the next Series B startup. You hit 20 million in revenue, that's a Series C startup. You have Series D startup is 50 million in revenue. And we have 16 startups inside Harness now, and they're all of these different stages. The advantage is the seed stage startups can fail fast. Our investment as a company is not that high. We can pivot very similar to what happens in the startup ecosystem. We can start with five people and start experimenting on something. And if we can't find a million in revenue or a path to a million in revenue, we know either we need to pivot to something else, which we do pivot, or we need to stop it. Let's stop doing it. But we can do that at a lower cost to the company.

Brett: Will they stay in this startup concept basically forever, or at some point it gets reorganized into the rest of the company?

Jyoti: They stay in the startup concept forever. It's all one shared platform. For our customers, if you buy five modules from five of our startups, you're still buying one. You won't even know if five things, is one product you're buying, right? But we create a model that we don't bundle things. Each of these modules are sold on their own to create accountability. They will never be best to breed in them if we don't have accountability to the customer. The minute we bundle it, then you can have three great modules for best of breed and two very inferior ones. And you'll start getting in this false sense of these things are selling and these things are... We look at, they're not even selling because they're just bundled part of something. By not bundling, it creates accountability to the customer. Our customer, we says we have these 16 modules, you can start with 1, 2, 3, 4, 5, 6, whatever you want. And we'll earn our business by proving that we are best of breed on this module number three or four or five. And you don't think we are, don't buy it, by the things that you think we are. But that creates this accountability to the customer, now they are only buying what they think is good because they are likely replacing some existing thing or something. And we're competing with those existing things in there. And same applies on renewal also, if they don't use that because they didn't get value, they won't renew. It creates a lot of internal accountability now. Now the startup team, they need to know and earn their business by being... Because let's say someone is using our CI/CD and they are very happy with it. Now they want to look at our feature flags. We don't bundle it part of CI/CD. We are like, we have to compete against the best of breed products on feature flags out there. And if we can't win, we know we don't have the best of breed product. Now it creates accountability to the customer and create accountability internally to us, to our teams that what they're building towards it. Of course there's a lot of synergy that comes with the shared platform. The shared platform, you get a lot of advantages.

Brett: And does the customer get advantages?

Jyoti: Customer gets massive advantage because it's one integrated experience. But we'll still look at that should not be the only thing. If someone is buying customer is just buying because you're inferior product for something, but it's part of one integrated platform, in the end I feel we are not going to win.

Brett: That's very interesting. [inaudible 00:56:18] It's the opposite of the Microsoft strategy.

Jyoti: I feel like in developer tool space you want to build good products for each of those, otherwise people won't use them. People will just build their homegrown things and all that.

Brett: When they buy the fifth product, you don't discount, you don't do... It's all has to stand on its own.

Jyoti: There are so much discounts. If the more you buy from us, there'll be some discounts. Yes, there are some part of it, but we still have to prove that we are, compared to the alternatives that are out there that are best to breed could be in that, that we can beat them. There could be some price thing, but that's not the primary reason people pick something, right?

Brett: Ultimately, you see customers land with all sorts of different products, and start here and then go to this one and-

Jyoti: Yes. There are certain, we call them the most common landing zones. Where people will start with, there are three or four that most people land with, and then the people will grow from there. But the advantage of that is now we have these startups running and they're scaling, and all the compounding effect of startups within startup is very, very powerful. Because now it's like I have these startups who are kind of go from a million revenue in first year to three, 4 million in second year, to eight to 10 in third year. To maybe 20, 25 in fourth year, or something like that, right? And every year we are launching three. You start layering the waterfall of that. We are growing well now. Actually, our business growth has accelerated, which at a certain size the growth starts coming down. Part of is that compounding that comes from that multiple startups that are happening.

Brett: When you think about the PM that you are putting in one of these new startups, is there something different about them than a good PM at Harness that makes them good to do that specific thing?

Jyoti: Extreme ownership, entrepreneur mindset. it's very important, I call it the entrepreneur mindset. It's because you are a startup founder inside it. Yes, we create the platform, the sales platform, the shared platform, all kind of things. But the entrepreneur mindset, which is like you have a startup founder, you find your way. And you are a startup founder, you can't just say, "I do my product, I couldn't figure out sales or I couldn't figure out marketing." You couldn't figure it out, your company startup dies. It's similar mindset that we try today. About actually half of our startup founders are former founders. Which actually helps, because now they have seen... But you don't have to be former founder, but a former founder always helps because they have gone through it and they've seen it and they've worked on it. We also acquire small companies. Because many times what happens, actually all the time, everything's said with AI to build a world-class best of breed product in each of these, it takes time. It takes about two years. By the time you say, "Okay, let's build this thing," and you matured it enough that you can sell to our enterprise customer base, it's normally a two year process. Maybe sometimes 18 months if we're building completely on our own, which is what we would normally do. But we also look at, we bring a small team? You hired a tech in acquisition or something, to cut down the two years down to a year or to six months because we have expertise and iteration. But whenever we acquire a small company, we rewrite the code completely in six months too on our platform. All the old code goes away on our platform, that's one shared experience and everything, but it's the expertise that you get. That you get expertise and you get that sometimes those one or two people who could be the startup founder for that particular area for you. That's how we accelerate and cut down that two year cycle, or many times to six months, a year. And we have done a few of those.

Brett: Do you have a unique compensation model for the people that are doing that, or they're just normally compensated?

Jyoti: I don't believe in overly complicated compensation models. Because the problem is overly complicated compensation models, they could drive bad behavior easily. If you don't like to do that. Then you start getting bad behavior, internal competition in the wrong ways, and all that. It's like, everyone's interest has to be aligned towards the Harness interest in the end. And-

Brett: Creating shareholder value and then everybody-

Jyoti: Creating shareholder value. But at the same time, they are heavily motivated to make their startups and their products successful.

Brett: I guess as we start to wrap up, if I were to watch you over the course of a month or two months running the company and talking to your team, are there things you're constantly explaining to them that they would roll their eyes that you're saying it again, but is a big part of your philosophy on building companies that we haven't talked about or explored yet? The most important ideas, this is how we build products, this is how we sell products. Is there anything we didn't explore?

Jyoti: Internally, people hear from me a lot about the startup within startup. We have this scaled autonomy, but you run it on your own, but you have still in a scale aligned framework. Which is important for... We don't want to slow down, you want to be... I always talk about the concept of continuous improvement as a company. Because when startups start to grow, people are like, this not perfect, this not perfect, this not perfect. And I always look at... I repeat one thing all the time, it's not about, are we perfect on something? It's all about, are we improving on something? And if you're constantly improving, that's the mindset. Because we don't need to... If the day we are perfect on everything, we'll be too slow. We cannot be stagnant. Whatever the problem is, we have to keep improving on that. And we find next set of problems and keep improving on it. That's something I talk a lot about, the value of customer value delivery. We talked about it. Everyone at my team will hear the concept of value delivery a lot. That's a core part of what we do in the company.

Brett: What did you mean by scaled autonomy?

Jyoti: By scale autonomy, I mean people run on their own without being micromanaged and too many gates and all that. It's almost like it's a balance. When you're a startup, you don't need any structure. You will grow faster by not having structure, by having a lot of people can just run on their own. But a certain time, if you don't have structure, that will slow you down. What is the balance? What structure you need and what autonomy you need? I started to call this concept, how do we internally create a blueprint for this? Startup within startup is an example of how we do product innovation and scale. How we are building a broad platform, and we are doing product innovation at scale with a lot of autonomy that people have. It's a similar thing we apply on the sales side as well. There's a framework in our CRO, Carlos, would run where there's a very structured sales framework, sales discipline, sales process. But how do you get every sales manager, first line, second line, third line, all of them to operate as it's your business and that you're responsible for your business in maybe New York or London, or whatever you are running? And how do you take it, that pride in you are the business owner and you have the autonomy to the right degree to manage your business? It's a concept of how do you give people more, but you still have the framework of structure so that they can still keep operating as startups?

Brett: And the important thing is that the general physics of companies is as they scale, you reduce autonomy.

Jyoti: Everything starts falling apart. People are like, "Oh, you don't have a...," and then you become a bureaucracy.

Brett: This has to be approved by this person, this person.

Jyoti: Yeah, yeah. Then it's like, how do you create the high degree of autonomy, but you have the right structures and checks and balances?

Brett: On the point around continual improvement, other than you talking a lot about it, how do you indoctrinate that? Or are there rituals or are there things that you get people to behave in that way?

Jyoti: Talking is very important. Sometimes people underestimate how important it's just to talk about something. But then it starts becoming part of the mindset of how people would think about it. But a lot of it is like if something goes wrong, no blame games. How do we bring the spread of continuous improvement? Of what do we learn from it and improve for future? And that's a big part of the culture that we built out. Things go wrong all the time. We're very intellectually honest post-mortem, learn from it and find the continuous improvement way of how we improve it. And we must still have something, again, go wrong on something, but we still get better. And people who are not doing that, they're stuck on something, maybe they are not the right people or they are not the right cultural fit for us. That's the part of it. But also, transparency around things. Having the notion of continuous improvement, people are a bit more transparent. They don't need to feel that they need to hide things that are not working. Because they can come in and say, "These are things working well, these are things not, and these are things we need to improve on." And they'll put this in the bucket of continuous improvement that we need. And this is what you show the improvement a quarter later or a month later, or something like this, is how we got there. You just have to get it on different layers that part of the company culture.

Brett: Wanted to wrap up where we always do, which is the question of, who has had an outsized influence on the way that you think about these topics of building companies and scaling companies? Is there someone that comes to mind that has imparted something particularly useful? What did they teach you or what's a part of your life philosophy in building companies?

Jyoti: I would say the company I admire a lot and the person would be Amazon and how Bezos build Amazon. A lot of the stuff that I'm doing in the enterprise software world, the startups within startup and how we are innovating and how we are scaling, and how we are going towards newer markets, and have the sense of ownership responsibility by a different autonomous level. I take a lot of inspirations from how Amazon was built. How Amazon went from selling books to selling everything. That probably has the most outsized influence in when I look at it as a role model of a company that could be built and how I would like to build it.

Brett: Cool. Great place to end. Thank you so much for spending all this time.

Jyoti: Okay. Yeah.

Brett: I really enjoyed it.

Jyoti: Yeah, great conversation. Really enjoyed it as well.