I think the reason the FDE model works only at Palantir and nowhere else is that 99.99% of startups are founded with the goal of achieving "product-market fit," and the entire VC model and conventional org chart revolve around this idea. The law of the land has been to build a product based on an insight, iterate with feedback, and eventually reach escape velocity to scale to as many customers as possible.
While this approach has undoubtedly resulted in many incredible products, especially in B2C, it fundamentally contradicts what enterprises often need—where the ideal model would be "outcome-market fit." This is why, despite the limitless options of BI tools available to enterprises, few actually help them act upon those beautiful insights.
I know a jewelry manufacturer who spent $1 million purchasing SAP, only to realize it was too rigid for their specific process. The consultants kept making up excuses and plans—five years later, and it’s still not implemented.
Palantir is unique because its goal was always "outcome-market fit" from day one, never tying itself to a single identity. Unfortunately, once a company has started on the "product-market fit" route, it becomes extremely painful to change course because it loses the ability to start from first principles.
I saw this play out at DataRobot. The company had an AutoML product that made for an incredible demo and promise: “Buy our product and democratize data science for everyone.” The problem was that once we sold it, we could only rely on "hope" to get customers to actually use it and bring models into production.
We had a solutions team that tried to help customers solve their problems end-to-end, but in an effort to avoid becoming a “dev shop,” we were always constrained by trying to fit our existing products into the solution.
Not only that, but incentives across the company were misaligned. Salespeople preferred selling something "understandable" and off-the-shelf. Engineers preferred building something “durable” into the core product. We were able to deliver some great projects and results, but it was an uphill battle.
What’s interesting to me about the route Palantir has taken is that it differs in two key ways: intent and patience. In other words, you have been highly intentional in providing as much value as possible for each customer, understanding that there will be many custom solutions that may not generalize immediately. However, with trust in the process and patience, patterns will emerge, revealing which components are truly composable and can be turned into repeatable products.
Certainly, this could not be done without: 1) Having "outcome-market fit" as a north star from day one. 2) Founders who control voting power and can resist board pressures to "build repeatability" for an extremely painful and long time.
Well said! The role of an FDE was a groundbreaking one. It was the first opportunity I had to solve problems that I believed needed more attention for the growth of the company. You just aren’t able to take that passionate viewpoints in other environments. I appreciate having the opportunity to experience that within a growing company. I would still like to explore how a training I put together 10+ years ago could be solved by Palantir Gotham today. As a teacher, I am always looking to elevate a person’s comprehension and how they are empowered to ask questions. My time at Palantir definitely allowed me to do that.
Does this mean, every Customer who buys Palantir software must have a FDE to help them solve the problem, then why focus on building a Developer eco system when only a FDE can deliver value to a enterprise using Palantir software ?
Excellent description of the essence of the role of the FDE at Palantir.
I've also always considered it the mark of sustainability of a model by the fact you openly disclose what you are doing, and the only way to replicate it would involve very hard and difficult work, which most people do not want to do.
I think the reason the FDE model works only at Palantir and nowhere else is that 99.99% of startups are founded with the goal of achieving "product-market fit," and the entire VC model and conventional org chart revolve around this idea. The law of the land has been to build a product based on an insight, iterate with feedback, and eventually reach escape velocity to scale to as many customers as possible.
While this approach has undoubtedly resulted in many incredible products, especially in B2C, it fundamentally contradicts what enterprises often need—where the ideal model would be "outcome-market fit." This is why, despite the limitless options of BI tools available to enterprises, few actually help them act upon those beautiful insights.
I know a jewelry manufacturer who spent $1 million purchasing SAP, only to realize it was too rigid for their specific process. The consultants kept making up excuses and plans—five years later, and it’s still not implemented.
Palantir is unique because its goal was always "outcome-market fit" from day one, never tying itself to a single identity. Unfortunately, once a company has started on the "product-market fit" route, it becomes extremely painful to change course because it loses the ability to start from first principles.
I saw this play out at DataRobot. The company had an AutoML product that made for an incredible demo and promise: “Buy our product and democratize data science for everyone.” The problem was that once we sold it, we could only rely on "hope" to get customers to actually use it and bring models into production.
We had a solutions team that tried to help customers solve their problems end-to-end, but in an effort to avoid becoming a “dev shop,” we were always constrained by trying to fit our existing products into the solution.
Not only that, but incentives across the company were misaligned. Salespeople preferred selling something "understandable" and off-the-shelf. Engineers preferred building something “durable” into the core product. We were able to deliver some great projects and results, but it was an uphill battle.
What’s interesting to me about the route Palantir has taken is that it differs in two key ways: intent and patience. In other words, you have been highly intentional in providing as much value as possible for each customer, understanding that there will be many custom solutions that may not generalize immediately. However, with trust in the process and patience, patterns will emerge, revealing which components are truly composable and can be turned into repeatable products.
Certainly, this could not be done without: 1) Having "outcome-market fit" as a north star from day one. 2) Founders who control voting power and can resist board pressures to "build repeatability" for an extremely painful and long time.
Well said! The role of an FDE was a groundbreaking one. It was the first opportunity I had to solve problems that I believed needed more attention for the growth of the company. You just aren’t able to take that passionate viewpoints in other environments. I appreciate having the opportunity to experience that within a growing company. I would still like to explore how a training I put together 10+ years ago could be solved by Palantir Gotham today. As a teacher, I am always looking to elevate a person’s comprehension and how they are empowered to ask questions. My time at Palantir definitely allowed me to do that.
Does this mean, every Customer who buys Palantir software must have a FDE to help them solve the problem, then why focus on building a Developer eco system when only a FDE can deliver value to a enterprise using Palantir software ?
Interesting to note that in 1992 MSFT had similar revenue to PLTR today but almost 12,000 employees
is this inflation adjusted?
Excellent description of the essence of the role of the FDE at Palantir.
I've also always considered it the mark of sustainability of a model by the fact you openly disclose what you are doing, and the only way to replicate it would involve very hard and difficult work, which most people do not want to do.