Start From Demand

🧠 Free Recall ❓ Quiz 🔍 Cues 🔮 Simulation 🥷 ShadowBox 📐 Encode Cards

Introduction
When thinking about their careers, most people think forward in terms of ‘skills’. By this, I mean that they work out “what are my interests?”, “what are my values?” and “what am I good at?” and then attempt to build a career by extrapolating from their answers.
This isn’t a completely silly approach. It is true that if you do work that you like, that fits your values, and that allows you to use what you’re naturally good at, you’ll likely have a competitive advantage against everyone else in your field.
The problem, however, is that the search space for careers is impossibly large. It is difficult to figure out what lies at the intersection of your values, interests, and skills, and that also happens to pay well enough to support the lifestyle you desire. To make things worse, many of these ‘internal’ factors are discovered through action, not through introspection. That’s just a fancy way of saying you can only truly know your values, interests and skills by trying things out — which makes the problem doubly difficult.
I’m not convinced that you can build a career moat if you only work forwards from skills, values and interests. If you do so, you’ll need a huge dose of luck — you would have to do the equivalent of blind trial and error as you grope your way towards a defensible career advantage. (To be fair, many of the people I profile have found their career moat in this manner, so it certainly is possible! But it’s not as replicable as I might like.)
In this chapter, I’ll argue that it’s easier to work backwards from demand. This is the fundamental philosophy on which career moats are built.
The Entrepreneurship Analogy
We’ll start with an analogy.
Imagine that your friend wants to build a new startup. A few months after he leaves his job at Facebook, he asks you out for coffee and proceeds to tell you about his new idea. You learn that he’s building an app that helps people find friends for lunch. He argues that it fits his background — he last worked at a social networking company, after all — and then he spends 20 minutes explaining that the app is different from its competitors because he has built in all sorts of fancy privacy restrictions. He says he left Facebook because it did not fit his values, and that he’s built these protections because he feels strongly about privacy as a fundamental human right.
How would you react? If you’re like most people, you would probably ask him how he knows that people would want this. You would ask about his target market. And you would ask how he plans to monetise in the near future.
In other words, you’d ask him to work backwards from market demand.
When we think about businesses, we tend to go overboard in the other direction
This is a common approach to startups. In fact, when we think about businesses, we tend to go overboard in the other direction — it’s more common to skip over the founder’s interests and values when evaluating startup ideas. (Don’t get me wrong: those do count for something, but usually over the long term.) With startups, your first questions are more likely to be about market demand: Who would buy this? What problem does it solve? And why is it valuable?
With careers, however, we tend not to ask these questions.
The startup analogy is useful because it maps reasonably well to one’s career: as an employee, you may think of yourself as a business of one: a combination of skills, network, and reputation that make up an offering to a potential employer. Business operate in a market. Individuals operate in a job market. Customers pay businesses to make some pain point go away. Similarly, employers hire you to solve a problem for them.
Seeing yourself as a ‘startup of one’ also does something useful: it helps you realise that starting from market demand is a necessary part of career thinking — in the same way that startup founders reason backwards from market demand into their products, you should use market demand to guide career decisions.
The implication here is that you should start your career moat journey exactly like a startup founder would: with research into the nature of demand in your market. And, like a startup founder who thinks from the perspective of her future customers, you, too, should take the perspective of your future employers.
So far, everything I’ve said sounds reasonable, and perhaps even obvious. But I want to be more concrete: it’s not enough to ‘know’ the nature of demand in your market. You must also know demand at the right level of detail.
Let’s take a look at what this means.
The Right Level of Detail for Demand
The naive view of ‘understand demand in your job market’ is some form of “oh, X is really hot now, so I should go get a job in that industry.” Variations of this thinking have existed for decades — in the 60s, accounting was all the rage. In Asian countries, becoming a lawyer or doctor has long been seen as good, stable vocation, presumably because the pay is high, and the work is respectable. More recently, I’ve lost count of the number of people who say some form of “oh, tech salaries are crazy, so I should get a job as a programmer in the tech industry.”
But this is not the right level of detail necessary to build a career moat.
A career moat is made up of rare and valuable skills. Of the two, the rareness of a skill is more important. So it’s not enough to simply observe that a field is hot; you must also know how the field works — which should then give you clues as to which skills might plausibly be rare.
Understanding SEA’s Startup Ecosystem
In the previous chapter, I told you the story of how I discovered my first career moat. Willis Wee of Tech In Asia told me that startups in South East Asia must expand regionally in order to justify their valuations. I then extrapolated from that comment into a generalised insight: that such startups might need someone to run their overseas expansions, and that the supply of such people might be constrained because such jobs were in uncomfortable, developing countries. I thought I might be able to parlay that into a career moat. I turned out to be right.
The way I tell that story today makes it sounds so simple, so self-evident. But the truth was that I was embedded in Singapore’s startup scene for years before that meeting with Wee. I had interned at a prominent local startup, Viki, a year before they got acquired, and I was running a prominent hackers club at my university, which gave me access to founders, software engineers, and investors in Singapore’s tech ecosystem. When Wee made that offhand comment, he was essentially giving me a set of lenses I could use to look at the ecosystem. And I knew that he was right. I knew this because it was consistent with everything else I’d seen over the past three years.
In other words, I understood the structure of the industry before Wee gave me a useful new way of looking at it. Like a startup founder researching their market, you often need to acquire a deep understanding of your industry in order to identify the skills that are rare and valuable within it.
I tell this story to demonstrate the sort of detail you’ll need to have in order to identify a career moat. Let’s walk through another example. Again, focus on the level of detail that we have to get into to describe the following moat.
Scaling a Startup Sales Team is Rare and Valuable
I mentioned in the previous chapter that getting really good at enterprise sales is one way to make lots of money — after all, the rise of Software-as-a-Service companies and the complexity of the skillset is one good way to build a career moat. But this much is obvious, even to an outsider.
There is, however, a more interesting moat in this domain, one that might not be as obvious to someone who has never worked in a B2B software company. The moat is this: it is really difficult to find someone who can build a sales organisation and scale it, from scratch. The skillset is difficult to hire for because it turns out to be a rather unusual combination of skills. And it is doubly difficult to find people like this if you’re trying to hire in Asia.
I experienced this first-hand. When we were scaling my last company, our biggest and most consistent bottleneck was with the sales team. I ran most of the software engineering in Vietnam, and it was relatively easy to find software engineering leaders — and in Singapore, relatively easy to find product managers. Ditto for customer support. But it was impossibly difficult to find a good startup sales leader. I remember a period of six months where we went through three different VP of Sales; each of them came with good paper credentials, each of them left after a couple of weeks. Eventually, my boss concluded that he had to do it himself.
Why was it so difficult to hire? To understand this, it helps to understand how enterprise sales works in the startup context, and what challenges a B2B software company might experience as it grows. I’ll use the example of Snowflake’s first Chief Revenue Officer, Chris Degnan, in order to explore why this role was so difficult to do, and therefore, so difficult to hire for (source).
Degnan started his career at a company called Covalent Technologies, right as the dot-com bust happened. He survived seven rounds of layoffs. To hear him tell it: “it taught me a lot of grit (…) every day I would have to come in and just cold call and dial for dollars, literally.”
A couple of years in, Degnan moved to EMC. Under EMC, he worked under a ‘very difficult human’ named John McMahon, who taught him the fundamentals over the course of eight years. And Degnan had an objective:
I remember giving him (McMahon), when I sat down with him the very first time, I had on my resume — um, I don’t even know if people do this anymore — but I put an objective. And my objective was to be the VP of Sales of a technology company. I had clarity around where I wanted to go. And so when I sat down with him, I said, that’s what I want to do, help me get there. And I think he was pretty direct with me: “Well, you have to make it uncomfortable for us to not promote you. Just do your job really well.” (…) There were times when people would say, hey, come and cover this big account at this other company. And I certainly got interested in that, and I listened, but what I started to think about is “Does the next move that I make — whatever move that is — does that get me to my objective of being the VP of Sales?” And the answer to that, most of the time, was no.
But enterprise sales was an incredibly competitive field. Near the end of his tenure at EMC, Degnan realised that the product he was selling was starting to become a commodity (“I felt like I was selling a used car”). And he looked ahead, and realised that the career path he was on — as an independent sales rep — wasn’t something he wanted to do.
I have people today that are kind of long time enterprise salespeople. I know plenty of them that like to do that. I particularly didn’t want to be the 50 year old sales guy. Right — I mean, I thought that far in advance, well, the money is better as a rep, generally speaking, than it is as a front line manager. And a front line manager job in tech is one of the more tough jobs.

(But) I wasn’t necessarily (happy) during the days of doing it. There were some pretty dark days. But I knew that, again, my objective was I wanted to be in a position like I am today. And I knew that being at an IC (individual contributor) was not going to do that for me.
Degnan was eventually introduced to Mike Speiser, who helped start Snowflake. And he finally saw an opportunity to get what he wanted: a solid role that gave him a shot at VP of Sales at a tech company.
But the move wasn’t as clear-cut:
Speiser convinced me to take a 50% cash pay cut, to come to Snowflake. (…) what really sold me on him, what I bought in on, was people were moving to the cloud. The data warehousing market was a massive market that was primarily (at the time) on-premise. Amazon had launched a product called Amazon Redshift. Which is still competitive (to Snowflake) to this day. And they had launched in 2013. So I figured, you know what? This is interesting.

Mike told me flat out — he was acting CEO — he said, Chris, you’re going to come in as director of sales. You are never going to be VP of Sales. Let alone the chief revenue officer. We’ll bring someone in with more experience than you to do that. But come help us build this.
Degnan hit the ground running. Snowflake’s core innovation was a cloud data warehouse that could invisibly scale both storage and compute under the hood. Two of Snowflake’s other founders — Benoit Dageville and Theirry Cruanes — were ex-Oracle database veterans who wanted to build a new type of data warehouse, but weren’t allowed to do so while they were at Oracle. So they left and built it with the help of Mike Speiser.
If you don’t know what a data warehouse is: think of it as a database that is optimised for analytics — that is, the type of database that makes short work of queries like “give me the top twenty products by sales for each product category, bucketed by locale, over the past five years, never mind that you have to chew through a hundred million sales orders to do so”. In the recent past, data warehouses demanded that you provision a set amount of compute and storage (this was true for both on-premise solutions and cloud data warehouses like Amazon’s Redshift); Snowflake’s technical innovation was that it would automatically scale the amount of computational power or storage necessary to service your queries, without bugging you about it.
At the time that Degnan joined, Snowflake was about 15 people strong, and still in stealth mode. The engineers had spent two years just building the product. Degnan’s first challenge was to figure out an initial ICP, or ‘ideal customer profile’. This was challenging:
I first started on the large enterprise. And then I pivoted to people that were exclusively on AWS. And man, there was a huge, huge pain. And I found the vein. I found the issue. I tapped that vein. And that’s kind of the beginnings of where I started to see customer momentum. (…)

It took me a little bit to figure out. It had to be a company that was already in the cloud, already on Amazon, and that didn’t care about things like security. Because we didn’t have any security features, really, when we first started selling. And that narrowed it quite significantly, around who I could go sell to. And the people who had the biggest pain were the ad tech industry, and online gaming.

So basically what happens is if you’re an online gaming company, users are in your game, and they’re clicking on things. And as the IT organisation or the CTO in the gaming world, who you’re selling to, they’re trying to figure out what their users are doing. And they’re trying to provide that user information back to the business analysts so that they can send promotions. So that (the player) can buy more swords in your game or whatever it is that they’re trying to sell you. And what would happen in the way the world was working back then was that people would take that clickstream data and dump it in Hadoop. Or (Amazon) EMR, or Cloudera, or whomever.

And then the business analyst would try to access that data. And it turns out that that was super slow. Hadoop was slow, EMR was slow. The business analyst would take three to four hours to figure out what a customer was doing. So their promotions would be hours old. So what we could do is take that log data, that JSON data, and natively ingest it. And take a process that took three and a half hours down to about a minute.

So for us to do that, it wasn’t really data warehousing per se, that I was selling. (What I was selling) was really, okay, you wanted to have faster insight into what your users are doing.
Discovering this initial customer profile was painful. Even with a clear pain-point and an ICP, Degnan spent six months begging ad-tech gaming companies to try their software for free. Eventually, Snowflake signed on two customers, and Degnan slowly started scaling up the sales process. His first step: hiring a sales engineering manager and an intern. The intern’s job was to help him create prospect lists. And the point of the lists was to generate eight customer meetings a week — a number that Degnan forced himself to hit by sending a report at the end of each week to the entire company … and the board.
What was most remarkable about Degnan’s story was how he grew with Snowflake.
Eventually, when he started to travel to hit those customer meeting numbers, he hired two Business Development Reps (BDRs). Then he got the BDRs to replace him completely in the lead generation process. And once Degnan started hitting 15 sales meetings a week, he went to the board and asked for the budget to hire his first three field sales people.
What was most remarkable about Degnan’s story was how he grew with Snowflake. At multiple points in Snowflake’s life, especially when it switched CEOs to Bob Muglia and then to Frank Slootman (the latter of which was notorious as a tough sales taskmaster), Degnan managed to hold on to his role. When asked how he did this, he said:
Luck, I don’t know. I think I owe a lot to two people in particular. John McMahon was on my board. And one of the reasons I almost didn’t take the job was Mike Speiser told me he wasn’t going to be the CEO for long. And I initially worked for him. And I said, Mike, I need someone on the board who cares about me when I join Snowflake.

And he’s like “who’s that?” And I’m like, well, a guy like John McMahon. He’s like, “I know John”. And on the next day he puts John on the board. (…)

And John is such a great sales leader. And he understands — he’s seen so many things. John’s still on our board to this day. And he gave me such great guidance along the way. I learned so much from him. And the other thing (was that) Speiser hired Bob Muglia. And Bob is just this world-class human being. And an incredible product guy. I mean, Snowflake would not be the company it is today without Bob. And he and I developed a very close relationship, and a very trusting relationship. And I think there was pressure for Bob to bring in people at certain times, over me. But we were crushing it. And Bob was like, “Chris is hiring great people.” And he’d go to John and say, “John, what do you think?” And John’s like “well, he’s doing well.” And I think it was kind of like, I had to control what I can control. Going back to what I said before is ‘don’t miss my numbers.’ So I would focus on exceeding every time.

Then, when it was appropriate, I’d say, “Hey, Bob, I think it’s time, people are looking at me as the VP of Sales. I’d like to have that as my title.” And he said, “Well, I’m not going to give you more cash or equity, but you can have that title.” I said, “Fine, I’ll take it.”
I told this story to give you a taste of B2B sales at a startup. If we generalise:
At the earliest stages of a B2B software company, your product isn’t mature enough. Enterprise buyers buy on features — and every enterprise product is in a race to feature parity with every other enterprise product. This means that your initial challenge is to find a subset of business customers who are in such deep pain that they are willing to buy a compromised product. This effectively means that you have to discover a market positioning that allows you to sell despite your product’s shortcomings. Many salespeople aren’t able to do this. If they come from a large company, they are taught to execute proven playbooks, for proven customer personas, in the context of a proven product. There are fewer salespeople who are able to think strategically — that is, who are able to work out the sales approach and ideal customer profile for an incomplete product, in a rapidly changing environment where both the market and the product are changing.
Second, you have to figure out how to hire salespeople. This is difficult on at least two levels: first, you need to design the right incentive structure for these people, and second, you need to figure out how to recruit salespeople, who are traditionally very sensitive to compensation. Then you need to teach them your playbook. Mess up the playbook, or mess up the training, or mess up the incentive design, or mess up your recruiting, and you are likely to hit a brick wall.
Third, and finally, you would need to change all of this as the product matures. As your product approaches feature parity, your set of ideal customers changes, and your positioning has to change along with it. And, as a result of the company growing, you would have to scale the sales organisation, meaning redesigning the incentives and processes and training to fit a larger salesforce. Your job then becomes more org designer and manager than pure sales strategist.
Degnan did all of this. He managed to do so because he had some uniquely skilled people to guide him during that process, one of whom was an ex-Microsoft exec (Muglia), and another who was a sales machine, who sat on Snowflake’s board (McMahon). And so we have the interesting property where it’s possible to pick up this skillset if you’re in Silicon Valley, and you have access to people who have scaled enterprise software companies in the past, but it’s less possible to do so in Asia, where senior salespeople are recruited after the company has hammered out most of the details around its sales strategy.
Eventually, my boss figured that it was simply too difficult to find people who were good enough at this specific skillset. He grit his teeth and got down to executing the playbook himself. And I remember thinking: if you could acquire this skillset and then make it legible to the founders and CEOs in Asia who faced this pain, you would likely have an extremely lucrative career moat for (at least!) the next decade.
Let’s go back up a level: if you had no experience in B2B software, would you have known that this was a career moat that could be built? Would you have known that this was a role that you could hire for? The answer is likely no. It’s no accident that many career moats are opaque to outsiders — you really do need a deep-enough understanding of an industry to figure out where the career moats are. I couldn’t have worked out the implications of Wee’s comment if I wasn’t already embedded in the startup ecosystem in SEA; I don’t think I would have recognised Degnan’s skillset as rare and valuable if I hadn’t experienced the sheer pain of hiring for it alongside my boss.
The Career Moat Strategy, in a Nutshell
The career moat strategy that I’ll cover in the next few chapters will take the form of the following questions:
How does my industry really work? You’ll need to find a perch — a place where you can figure out how your industry works. We’ll cover how to find a perch in the next piece.
What skills does my industry value that is unusual, or rare? (And, more specifically, what types of companies value those skills?)
Is there a plausible case for why these skills might stay rare?
Finally: am I capable of acquiring such skills?
The rest of this guide will follow the structure articulated above. In the next chapter, we’ll talk about how to discover which skills are in demand. Then we’ll talk about how you might uncover reasons for why those skills might stay rare. Finally we’ll talk about the trial and error process required to acquire the skills you need for your chosen career moat.

Unstarred

Dunk (exclude from study)

Start From Demand

Review Questions

Todos

Sensemaker