Why Your Startup Fails to Scale: $12B Company Founder Explains

Dheeraj Pandey built Nutanix into the largest tech IPO of 2016 before stepping down as public company CEO to address structural silos inside enterprise software. With DevRev, Pandey connects internal corporate data with natural language models, creating unified knowledge graphs across engineering and customer support. His executive philosophy emphasizes taking market risks over simple technology risks, practicing unreasonable hospitality with B2B clients, and leading with total operational authenticity.

My name is Dheeraj Pandey, and I'm the co-founder and CEO of a company called DevRev. We are in the business of blurring the lines between the public wisdom of the internet, natural language wisdom, and the wisdom of the enterprise. That is basically what DevRev is all about.

Before this, I had started a company, Nutanix. We were the fastest to a billion dollars in revenue, and we took it public in a record seven years on NASDAQ. It was the largest tech IPO of 2016, and I ran it as a public company CEO for almost five years. Then I realized there was a big opportunity opening up in AI, which is how DevRev started almost four years ago. I also sit on Adobe's board, where I'm on the audit committee, and last but not least, I am a big student of both AI and design.

I had a fairly dysfunctional childhood. I grew up in the poorest state of India, where the Himalayan and seasonal rivers would flood half the state for half the year. It was the most lawless state as well as the poorest one. That state's GDP per capita is only as much as sub-Saharan Africa, so you can imagine the level of poverty that I actually saw. That is what catapulted me to get out of the state and do something.

Back in the day, you could only be a good engineer or a good doctor. There was no third profession, so I picked engineering. Luckily, when I was growing up and had to pick my field of engineering, the internet had just happened. This was 1993 and 1994, and computer science was going to be the next hot thing, just like AI, which is the new computer science, is hot right now. I built my entire career around the internet.

I can stay in India right now, but back in the day, the only option was to leave the country if you were an engineer. I had to be in that bubble to know what working in the industry really means.

A big part of this is having seen dysfunction. As a child, the worst is not bad enough. If you have seen sad and disappointing days as a child, that actually carries through with you. But it also means you can take care of tough situations, and entrepreneurship is a lot of that. It is a lot of highs and lows, and it is important not to get too sad with the lows, which also means that you can't be too happy with the highs. Staying on that average line and modulating your emotion is the way you will deal with a lot of failures.

The only direction when you start a company is forward. That is the best you can do. Nutanix was my first startup, and the core of Nutanix was to really do something around data management. We said that we were good at data-intensive applications and that we reliably manage data at fast speeds like no one else. We could ship software to you so that your data centers would look like Google's and Amazon's and Facebook's.

We were also trying to bring commodity hardware, which was Taiwanese equipment, to the enterprise. Before that, people were buying hardware from Dell, HP and IBM, along with other branded hardware companies like EMC and NetApp. They were all into shiny hardware, and we were saying no, you can take the same approach that Amazon, Google and Facebook take. You can source hardware from Taiwan, bring it to the US and put software on top, so you have commodity hardware and pure software running on top of it.

The first four or five years were extremely problematic because the hardware was flaky and unreliable, and the software needed to overcompensate for that. So we took our time to really make the software compensate for flaky, fragile hardware. In fact, our company almost shut down three times.

At the core, the big thing in my head was that you are walking a tightrope and you have to cross a valley. When you are halfway through, you can't turn around. You will fall if you turn around.

Any new product that you launch, any new initiative, you have to think about following through. Sometimes moving forward might mean shutting it down completely, but to say that I want to go back in this direction is the worst thing that entrepreneurs actually do. So we did have a couple of big pivots. The only constant in building a company is change, so if you can embrace change and make it seamless, you can be a little bit ahead of the market.

You have to be really good at looking around the corner, and that means listening. When you hear something from the market the first time as a stubborn entrepreneur, you will say that it is an aberration. When you hear it the second time, now it is a coincidence. When you hear it a third time, it is a pattern. You need to understand how to recognize patterns and look around the corner, be at least six months ahead of what will eventually be a hard pivot, and see if you can make soft pivots in that direction. There is no point in making 90 degree turns or 180 degree turns if you can make soft turns to the same degree that you would have achieved anyway.

If you look at the idea of Nutanix, it was where the puck was as opposed to where the puck was headed. With the first idea of my first startup, I had started to really take technology risk but not market risk, because the market by 2012 had already seen cloud. Consumer cloud companies were saying that you don't need to own infrastructure if you can stream the infrastructure. It is like what happened with music eventually, where you don't need to own songs. The market for Spotify was already happening, and the public cloud was that version for infrastructure.

So the puck was headed to streaming, and we were still at iTunes. On-prem software was basically headed into that challenged market where you would constantly have to think about whether you needed to own or whether you could stream. Do I need to own, or can I stream? So the idea of Nutanix was where the puck was, it was a technology risk, and we got paid for it with a $10 to $20 billion valuation for Nutanix.

But to really build a $100 billion business, you had to take a market risk. This is something that all entrepreneurs will actually face, as an investor but also as an operator. When you do your own company, you are doing both in one. At some point you might have to decouple those, where you become an investor in your current idea and then you become an operator of a new idea. Around 2020, when COVID happened, is when we asked what the next chemistry was that we could bring together, and it really happened to be around business software.

When GPT came about, we already had our own vector database. We had built our own vector database to do search, because we knew search was the first killer app for anything intelligent. What GPT has given us is the ability to ask it questions and bring in the natural language wisdom that we would otherwise have struggled to build. But at the core of it, the enterprise data piece was the biggest challenge, because I can't do anything in an enterprise software environment with Wikipedia's knowledge or the internet's knowledge. I need to bring the knowledge of the enterprise together in a coherent whole, which is what a knowledge graph actually brings, and taking that knowledge graph and mixing it up with the wisdom of the internet is what we really do. So the biggest bang for AI is building a Perplexity-like experience for all enterprise assets, and that to me was the light bulb moment.

In the absence of the public cloud, we could have been a $100 billion company, but in the presence of the public cloud we would be hovering around the $20 billion range, because the market has moved. If technology doesn't integrate things and bring the chemistry, as opposed to just new features and new ideas, it has not done its job. Big ideas are about alchemy. It is about mixing things together and making it look like a more robust solution, and even a chemical solution is about bringing chemistry together. Bringing a lot of these things together is where a lot of value gets created.

Product-market fit is actually a journey, not a destination. You will need to struggle with that at 1 million, at 10 million, at 50 million, at 100 million, at a quarter billion, at half a billion and at a billion. At every such milestone you have to keep thinking again and think ahead of the curve, because if you waited too long you might start to sputter at 100 million. You might have needed a new product for going from 100 million to 250 million, and you might have needed to see a new region of the world to do sales and marketing. This idea of PMF is relative, and it is not a destination.

Every time you think you have product market fit, you will start to sputter at the next threshold. So you are at 10 million and you think you have PMF. That means you have not thought through that at 50 million you will stall, because you don't have a pipeline that was large enough for more features, more capabilities, bigger deals and bigger customers. Or you didn't think hard enough about the regions of the world, which will now create a portfolio effect.

What you need to keep doing is create a portfolio of products, customers and geographies around the world that will give you that balance every quarter, because not every existing customer will pay you every quarter. So you need to find a portfolio of large customers so that they can all balance each other every quarter. That to me is a journey more than a destination.

We take a step back and think about the funnel of ideas and the life cycle of an idea. There are a lot of ideas up there at the top of the funnel, and then they become concepts. From concepts, when it becomes a little bit real, it is still a feature. You talk about all these companies that are basically building agents, but at best they are features, and you can get some early wins by showing a feature to your customer.

But how do you take a feature to a product? A feature alone is not going to get me $100,000 from a customer. How you organize all this into a robust product, and a product into a company, and a company into a business, is the real test of entrepreneurship. Going from a company to a business requires building relationships with customers and making sure that you don't leave customers in the lurch when they need the most amount of help. One of the reasons for our success is underpromising and overdelivering, and this is the core of repeat business.

If your existing customers are not paying you more, then you are basically just selling and running, which means hunt, hunt, hunt. But the balance between hunting and farming, which is new customers versus getting more from existing customers, is at the core of building large businesses. Jeff Bezos calls it the flywheel effect of companies, where you need to add new logos but you also need to go make money from existing logos.

I think that tests the real mettle of a company. Do you have a product that is authentic? Do you have customer support that is really authentic? And finally, do you have salespeople who really know how to underpromise and overdeliver?

I'm a big fan of Will Guidara. He is a restaurateur who had two of the most famous restaurants in the world, and he talks about this concept of unreasonable hospitality, which means giving people way more than they expect. That is where great entrepreneurs differentiate from the mediocre ones.

A lot of people think that entrepreneurship is about selling your dreams and ideas to prospective candidates who become employees of your company, but also to prospective customers and partners. Very few people, and I would say less than 1% of people, are really good at following through. They will sell and move on, and they will sell some more and move on, as opposed to selling and staying. That idea of unreasonable hospitality is what really creates the flywheel. When you go to new customers, they hear from existing customers, and that is how you reduce the friction.

At the end of the day, there are two emotions going on in a new relationship, the fear of missing out but also the fear of messing up, FOMO and FOMU as we call it. A lot of people think that they just need to create FOMO with prospects, but they don't realize that most customers, especially in B2B, and I'm talking about B2B software, work differently. When you are doing larger deals, like $500,000 or $200,000 deals, a lot of it is actually based on FOMU, which is fear of messing up.

So building that critical relationship, providing that unreasonable hospitality and letting your existing customers speak of that unreasonable hospitality is the only way to blunt that emotion of fear of messing up. To me, entrepreneurship is about building that authenticity at every level of the way, because the market is actually smarter than you think.

What is disruptive in technology? The biggest disruption in technology is miniaturization, turning big things into small things. Lighter weight, sitting in a pocket, not on a desk. Even the idea of infrastructure was like that. We were buying things for five years in big metal monolithic purchases, and now you can just come and use the compute for two hours if that is what it takes for you. When you make things smaller and lighter weight, that is the biggest disruptor to incumbents.

SaaS has become extremely heavyweight and extremely complex in terms of implementation, so the disruption that will happen is consumption. We will go from subscription to consumption, and consumption models are what will disrupt subscription models. Remember that subscription models disrupted the licensing models of the 20 years before that, where people were buying five-year enterprise license agreements.

We moved to subscription and said that you can pay by the user. When Slack comes in, you don't have to pay by the user, because you can pay by the monthly active user, making it even more lightweight and a little bit more towards consumption. Snowflake came in saying everything is consumption-based. I think all SaaS will probably become consumption-based, and that will disrupt legacy SaaS.

The worst decision is to sit on things for too long, to sit for too long on people who are not a good fit, on bad products, unhappy customers, unhappy employees and mediocre leaders. To me, sitting on the suboptimal thing for too long is the worst thing that can happen. It is the tyranny of mediocrity, not changing and doing nothing.

That is exactly what you have to go and fight when you go and sell your new idea to customers, because a customer's most comfortable choice is to do nothing. Your biggest competition is not that other idea or the other competitor. Your biggest competitor is inertia, and that is also the biggest enemy within a company. That is the worst thing an entrepreneur can do, and it is the worst thing a company can do.