Ten Years in Startups: Why I'm Building Again

One of our founders sold his cars so we could make payroll. Years later, the company was valued at more than a billion dollars. I was there for both.

After five years at SAP, I joined Beamery as employee number six, one of the founding engineers. I wanted the things that make startups appealing: interesting work, freedom, ownership, and a share in the upside if we built something valuable.

Over the following decade, I watched a handful of people become a company of more than four hundred. I saw early colleagues leave, built an AI team inside the business, and experienced the excitement of becoming a unicorn. I also watched repeated layoffs wear away the confidence that had taken years to build.

By then, I had plenty of reasons to understand why someone would choose a different kind of job. I still wanted to build another startup. What had changed was how I thought about the people, the business and my own place in it.

I kept coming back to a question: if I had the chance to help shape a company from the beginning again, what would I do differently? These are the experiences behind my answer, and behind my decision to become a CTO.

At the beginning, we were all figuring it out

We were cramped into a small office, excited to be doing the startup thing. It was our first time building a company. We had all seen the movies and had some version of that picture in our heads: the tiny team, the chaos, the stories we would tell when it all worked out. We were already talking about how we would remember those days once we were successful.

There was a lot of fun in it. Some people had taken a chance on us, and we were going to see what we could make together. Underneath that excitement were two questions: could we make the idea work, and could we make it work in a way that someone would pay for? We were learning about the product and the business at the same time.

The pressure became much more real as the money started running out. Keeping the company going meant finding enough traction to raise another round. Selling became an obsession. Sell, sell, sell. We needed somebody to take a chance on the product we were building, just as others had taken a chance on us.

Eventually, we landed our first customer. We got feedback, improved what we were doing, and won our first large enterprise account. That helped us raise Series A. It also gave the engineering work a clearer direction: we had customers whose experience could tell us what to build next.

We could start adding people around the original team. That felt like progress, and it brought a question we hadn't had to answer in the small office: who would get to lead the company we were becoming?

After Series A, who gets to grow?

There still wasn't much hierarchy, but we could feel the need for more structure. Leadership started looking for people who had done this before.

The way we had hired early on was part of that story. I remember plenty of relatively junior hires, with keeping salary costs down an important consideration. People received equity too, so they had a stake in what we were building. But we hadn't started with a team full of people who had already led companies through the next few stages of growth.

I understood the attraction of an experienced head of engineering or VP. We were learning while trying to keep the business moving. Someone who knew what the next stage required could help us get there.

For people on the original team, those hires raised a different question. They had taken on work before there were proper roles for it. They had helped create the opportunity. Now that there was a larger job to do, why wasn't there a path for them to grow into it? That tension reached across the company. A sensible decision to bring in experience could still leave an early colleague wondering where they fitted.

Around Series B, some of the core team started leaving. I could see people losing motivation when they weren't getting the responsibility or progression they wanted, and couldn't see enough support to get there. The company was getting bigger while their own future in it felt smaller.

My position was different. I had specialist expertise and a distinct place in the business. I felt trusted and had room to pursue work that interested me. I wasn't being passed over in the way some colleagues felt they were. That difference matters to how I tell this story: I could see the frustration around me while still having opportunities of my own.

Lesson 1: Build the company around the early team

This is something I now tell the people I manage: my job is to help you get where you want to be as quickly as possible. That means understanding what you want, where you need support, and which responsibility you are ready to take on next. Those conversations need to happen while the company is creating new roles, when there is still a chance to shape them.

Looking back, I would have wanted us to be much more deliberate about how experienced hires helped the people already there develop. The new person might be right for the job. The colleague who wanted that responsibility still needs a future they can see and help preparing for it. Equity gives someone a stake in a possible financial outcome; it doesn't answer their question about who they can become inside the company.

As I hire a core team now, my message is clear: I'm building the future of the company around these people, not on top of them. I want them to grow into the responsibilities we create. That makes the first hires especially important. I'm thinking about how they make decisions, respond to unfamiliar work and help other people develop, as well as what they can deliver immediately.

I'm also putting much more experience into that first team. I'm looking for very senior people, and I'm willing to pay more for them from the beginning. AI is part of why that calculation looks different now. I want people who can use it effectively and have the experience to judge the result: what is worth pursuing, what needs checking, and when something is ready for a customer.

For me, that is where AI can multiply what an experienced engineer can do. It can help them explore an approach, build something we can try and work through changes faster. They still have to choose the right problem and recognise when an apparently convincing answer is taking us in the wrong direction. That combination of speed and judgment is what I'm hiring for.

The financial bet is deliberate. We'll spend more on salaries each month, with the aim of taking fewer months to put something valuable in a customer's hands. I care about the total time and money it takes us to get there. If we can demonstrate value sooner, I expect us to be in a stronger position for the next funding round. The measure is simple: how many months pass between hiring the team and the first customer paying for what it built. That is a bet on what this team can deliver, rather than an assumption that using AI guarantees a quicker raise.

Hiring senior people doesn't finish the job. Their work will change as the company develops, and at some point we will bring in additional experience. When we do, I want it to strengthen their ability to lead rather than cap it.

As we scaled, the culture changed

As we raised more and hired more, the immediate feeling of running out of money eased. Keeping people motivated became harder. The company was succeeding at growing, but some of what had made it enjoyable to work there was becoming difficult to recognise.

Lesson 2: Pay attention to who carries the culture

The early culture lived in the people I worked with every day. We had some written values from the beginning, but the ones that felt most real were unwritten. They showed up in our expectations of each other, the room we had to try things, and the sense of being in it together. We had shared the uncertainty of getting the company started. That experience influenced how we worked without needing much explanation.

When early colleagues left, some of that went with them. Replacing their responsibilities didn't automatically replace their influence. They were the people who could show a new joiner how we approached the work, often without thinking of it as teaching anyone about culture. Some of the people who felt they had run out of room were also carrying the habits we wanted to keep.

Meanwhile, the managers we brought in were shaping the daily experience of the growing teams. Some came from large enterprise companies and brought the processes they knew. We needed more coordination; I understood that. What frustrated me was how much of their previous way of working seemed to arrive with them, without enough questioning of whether it fitted the company we were building.

For someone joining at that stage, their manager's approach was the culture. They hadn't been in the early office with us. They were learning what was expected from the people making decisions around them now. We could keep telling the founding story, but it had less influence on an ordinary working day than the habits of those managers.

There wasn't one process I could point to as the moment everything changed. It was the accumulation: more structure, more steps, more effort to get the same things done. We were adding people and experience while losing some of the speed that had made the work exciting.

As we scaled, we tried to put more of our values on paper. That made sense: new people hadn't shared the early experiences, and we wanted to describe what mattered to us. But I remember an expensive two-day offsite with the whole leadership team, with workshops to come up with a new set of values. We were putting substantial time and money into choosing the words.

At one point, we ended up with more than twenty values. “Be the change you seek.” “Act with kindness.” I remember the phrases, and I remember the posters in meeting rooms. I didn't see those posters shaping our everyday work in the way the early, unwritten values had. The experience of working there still came from the behaviour of the people around us.

That gap stayed with me. A company can be very good at describing how it wants people to behave while the people inside it experience something else. I want a new colleague to be able to recognise our values in the team they join. If a value only makes sense after a long explanation of what we aspire to be, I need to ask how much we are actually living it.

The lesson I took from those workshops is to keep the list short. Write the values down, but keep them few enough to remember and close enough to daily life to mean something. Start with the behaviour already worth keeping, so that codifying it helps people recognise and continue it.

The other half of the lesson is about who we hire to manage. Their habits become other people's experience of the company, far more than any poster does. Preserving a culture means weighing that influence as carefully as the experience on their CV. A workshop can't do that work for us.

A startup inside the startup

Around the time of Series C, I got the green light to build the AI and research team, focused on applied AI. We had found product-market fit, with annual recurring revenue in the tens of millions. The business needed to stay ahead and make the product meaningfully different. Leadership trusted me to take that work on.

For me, it was a chance to build a startup inside the startup. There was interesting technical work to do, a business reason to invest in it, and room to create our own working culture. I could do something about the loss of freedom that had been frustrating me.

I remember the company being on Confluence and Jira while we used Notion. People questioned why our team was using something different, and I pushed for it. It was a small, visible decision about how we wanted to work. I wanted the team to know we could make those choices and stand behind them.

That helped give us an identity within the larger company. People had a say in the environment they were working in, and we did work I was proud of. It also gave me a clearer understanding of what I enjoyed about leading a team, beyond choosing the technical direction.

Lesson 3: Give trusted people room to make decisions

I've found that smart people love autonomy. They want to use their judgment, explore an unfamiliar problem and have some freedom in how they approach it. They also want to feel that they are growing through the work. I recognise that in myself. Being able to work something out, rather than just carry out an answer someone has already decided on, was part of the appeal of joining a startup.

Learning is a big part of that growth, and learning takes time. Sometimes an idea needs to be followed far enough to discover whether it is useful. Sometimes somebody needs room to investigate an approach before they can tell you what it will lead to. I want curiosity to have a place in the team's work. Expecting people to develop their judgment means giving them opportunities to exercise it while the answer is still uncertain.

That mattered particularly in an applied AI team. We were exploring what could make the product different in the future. The team needed room to develop ideas about that, alongside the work of making something useful. If every investigation had to begin with a known outcome, we would have limited the work before it started.

The difficult part for a manager is being comfortable with that freedom. For me, it comes down to trust. I need confidence in how someone approaches the work even when I don't know every detail of it. Without that, I can keep checking, stepping in or reopening decisions I had supposedly handed over. My need for reassurance can become the thing that removes their autonomy.

I was trying to pass on the freedom leadership had given me. That means accepting that someone's approach may differ from the one I would have chosen, and offering support without taking the work back. The difference is part of what I am trusting them to decide.

Most of my team at Mav9 came from my previous team at Beamery. I already knew their work and trusted them to run with things without being micromanaged. We had built relationships over time, and I wanted to build with those people again. Those relationships are one of the most valuable things I carried out of that decade.

There is a very practical side to this at an early-stage company. I'm juggling enough things that I can't supervise every piece of work or be involved in every decision. I need to give someone responsibility and turn my attention elsewhere, knowing the work can keep moving. Starting with people I trust gives us that ability much earlier.

Later, growth stopped being the whole story

By our Series D in December 2022, Beamery had reached a valuation of more than a billion dollars. Going from employee six to working at a unicorn was extraordinary. We had customers, a much larger team and several funding rounds behind us. It felt like recognition of years of work.

There was a lot to be proud of, and I found it easy to let that pride become confidence about the future. We had come through so many earlier challenges. Reaching this stage made it tempting to think the business was on much firmer ground.

The pandemic had already arrived before Series C and D. We were reaching those milestones while the conditions around the business were changing. Growth and uncertainty overlapped; the funding announcements were only one part of what was happening.

Later, the downturn and pressure on sales brought a very different experience of the company. There were layoffs, then more layoffs, then more again. I watched morale fall as people went through that repeatedly. After seeing the company expand for so long, it was difficult to take in how quickly we could find ourselves shrinking.

Lesson 4: A funding round doesn't make the company secure

The repetition is what stayed with me. After one difficult round of cuts, there is a desire to believe the company can settle and move forward. Another round makes that harder, and the possibility of it happening again becomes part of how everyone looks at the future. A financial decision happens quickly. Recovering the team's confidence takes much longer, and with repeated cuts people have to adjust again before they have settled from the last time.

During the growth years, looking ahead had been exciting. There was another customer to win, another stage to reach. After repeated layoffs, thinking about what came next carried a different weight. I could still be proud of what we had built while feeling much less certain about where it was going. Becoming a unicorn hadn't resolved that uncertainty, even if it had briefly made it easier to overlook.

The people who remained still had work to do. There was a product to develop and customers to serve, while the group doing that work kept changing. They had to absorb what had happened, adjust to who was left and keep investing energy in the next plan. Watching that made it clear how much a company asks of the people who stay, and I started to see their recovery as part of the work ahead of the business.

It changed what I took a funding round to mean. The money gave us time and choices. We still had to sell enough, make good decisions about spending, and cope with conditions we didn't control. I had allowed the progress we had already made to carry too much weight in how I imagined the future. The company needed to keep earning the next stage, however impressive the previous one looked.

I also looked back differently at some of our bigger decisions. We had turned down acquisition offers and made acquisitions of our own. Some of those acquisitions, from my perspective, didn't deliver what we had hoped for. As the circumstances changed, I kept thinking about the choices we had made when the future looked more generous. Their consequences stayed with the business as its options became narrower.

Those experiences make me ask more about the alternatives before committing to a plan. What would we do if it took longer, or if the conditions changed? The optimistic outcome is usually easy to get excited about. I want to spend enough time understanding what the company would have to do in the other outcomes too.

And when a company asks people to commit to a new direction, they bring the experience of the previous one with them. A new plan doesn't reset that history. I now take much more seriously the work needed to make the next plan believable to the people being asked to deliver it.

AI did not shorten the sales cycle

One of the opportunities we were working through was the surge of interest in AI. Our team had been doing applied AI work for years by then, and the attention raised the question of where we could take the product next.

I saw customers want AI and be nervous about buying an AI-driven product at the same time. They could see possibilities in the technology while being unsure how to bring it into their organisation. Their enthusiasm was real. So was the work they needed to do before they could approve a purchase.

I remember a lot of questionnaires. How could they use AI safely? How should they assess bias? What would an audit need to show? We were answering questions about the product while customers were still working out what their own evaluation and approval process should look like. Sometimes the process around buying was as unfamiliar as the technology itself.

Their caution made sense. We were building for talent teams, where fairness mattered to how the product would be used, and the people assessing it had responsibilities of their own.

There was substantial work around bias audits, including independent audits. An outside assessment gave customers something concrete to examine alongside our explanations: what had been tested, what the results showed and what they could reasonably conclude from them. It contributed evidence to a decision the customer still had to make.

That work made explaining the technology part of delivering it. We needed to make the product understandable to the people responsible for assessing and approving it. A technically useful capability could still be difficult to buy if those people didn't have what they needed to evaluate it.

There is more shared structure to draw on now. ISO/IEC 42001, published in 2023, sets out requirements for an AI management system. I see frameworks like that as a useful starting point for a more consistent conversation between customers and suppliers about how AI is managed. The questions about a particular product and its use still need answers.

Lesson 5: Allow time for customers to get comfortable buying

As an engineer, I was drawn to what we could make possible. I could see the progress in the product and the opportunities ahead of it. This experience made me pay more attention to the customer's ability to act on that progress. Their evaluation process was part of the route to adoption, and it had a timetable of its own.

Enterprise sales were already lengthy. AI added questions that customers themselves were still learning how to answer. We could move the product forward while a buying decision remained some distance away. The urgency inside our company didn't make their process move at the speed we needed. That was especially difficult when we were looking to new technical work to help the business find its next direction.

This is also the limit on the bet I'm making with a senior team and AI-assisted development now. We can try to shorten the time it takes us to deliver something valuable. We still have to understand what happens between delivery and a customer's decision to buy. I want to know early who needs to approve the product, what evidence they will need, and how we can help them assess it.

The lesson for me was to allow for the period when interest in a technology runs ahead of an established way to buy it. Customers become more familiar with it and assessment processes become more consistent, but the company has to survive until that happens while still paying for the work that made customers interested in the first place. A customer can have good reasons for taking more time, and we still need the work to become revenue. Understanding their position doesn't remove the pressure on ours.

The financial upside, and the wait for cash

There was a personal financial story running alongside all of this. The possibility of sharing in the value we created had mattered when I joined. As employee six, I was early enough to help shape the company itself. I wanted that commitment to lead to a meaningful financial outcome too.

What I understood much less clearly was how long it could take before ownership became money someone could use. The company could raise another round and continue hiring while the question of when an employee could sell remained open. Funding the business and giving shareholders a way to receive cash were separate events.

To understand that part of the story, go back to Series C. That was when I saw the first secondary opportunity at Beamery, already years after I had joined. A secondary allowed some existing shares to be sold while the company was still private. It was an opportunity to turn part of that ownership into cash without waiting for an acquisition or a public listing.

Access was limited. From memory, only a handful of people across the company were able to take part. Saying that the company had done a secondary could make it sound like a milestone for the whole team. In practice, its significance depended on whether you were included. Even being there almost from the beginning had meant years of work before seeing that first opportunity.

The unicorn announcement came later, at Series D. I remember the excitement of thinking about what our shares might be worth at that valuation. We had helped build something investors valued at more than a billion dollars. It was natural to start doing the arithmetic and imagine what that could mean for our lives.

Lesson 6: Paper wealth isn't money in the bank

The difficult part was separating that excitement from what was available. The funding announcement didn't mean employees could sell at the values they were imagining. It didn't establish when another secondary would happen or who could participate. I could understand the distinction in principle and still get caught up in the possibilities.

Later setbacks changed how I thought about those numbers. A valuation from a funding round could remain the number people remembered while sales and the outlook for the business became less certain. I had to stop treating an earlier headline as a fixed expectation of what my equity would eventually deliver.

It also changed how I compared joining at the beginning with joining after several rounds. They are different financial decisions. Later employees can have valuable equity, but I wouldn't assume that their opportunity resembles the one offered to the first few hires. I would want to understand the actual ownership on offer and what would need to happen for me to receive money from it. A company's size or reputation can't answer that for me.

For a long time, I put Beamery first. I treated the company's success as if it would automatically become my own. That was a mistake. The company could reach its milestones while my financial outcome still depended on a separate set of conditions. Caring about the business and doing everything I could to help it succeed didn't settle those questions.

I also became more aware of how incentives can diverge after some people have cashed out. Someone who has already received a meaningful payout can be comfortable with the business running on terms that suit them. Someone who received very little may still be chasing the outcome that makes their commitment financially worthwhile. The same company can represent a comfortable lifestyle for one person and a promise of future reward for another. Being in it together doesn't mean everyone still needs the same thing from it.

That made me more willing to be selfish about my own financial outcome. I want the upside to be clear, and I want to understand the conditions attached to it. Loyalty to a company shouldn't mean leaving my own interests unexamined. The years I put into building it are years of my life too.

My rule now is simple: equity should come on top of a base salary I'm happy with. I won't compromise that base pay in exchange for a possible payout later. I still want meaningful ownership in what I'm building, but the cash compensation has to stand on its own. Joining should make sense before any future sale of shares.

Why I wanted to do it again, as CTO

After all of that, I still wanted the experience of getting something off the ground with a small team. I knew how much I enjoyed building, and I wanted a bigger part in the decisions that shaped the company around the product. The next opportunity had to make sense with everything I had learned along the way.

When I joined Mav9, funding was already in place to cover at least three years under our plan, including hiring a full team. That gave me a useful horizon. I could think about the people and the product with the resources to carry out the plan we were making. After experiencing the pressure of a company running short of money, having that starting point mattered.

The work was greenfield again. I could get back to building, make early technical choices and shape the product as we learned. Building the AI team had given me a chance to create one part of a larger company. This time I wanted to be involved while those choices were still being made across the business.

We also had design partners and sales channels already in place: people we could work with on real problems, and routes through which we could reach customers. That gave the technical work a commercial starting point. I wanted to understand who we were building for and how we would get the product into their hands before getting too far into building it.

Choosing the industry was part of that decision. At Beamery, our buyer was HR, and I saw limited budgets affect sales. A team could want what we were offering without having the money to buy it. That made me much more interested in where the budget sits, and whether the people who see the value have the means to act on it.

The fintech opportunity at Mav9 looked stronger to me on that front. I saw more scope to work with buyers who had money available for the problems we were trying to solve. I wanted to think about those commercial conditions before committing years to building within them. The quality of the technology was only part of the opportunity I was choosing.

And I really liked the founding team. The trust was mutual. I was choosing people I would work closely with through difficult decisions as well as exciting ones, and I wanted to spend my time building with them. The funding, the opportunity and the work all mattered; so did that feeling about the people.

I was joining early again, this time as CTO, with the technical direction and the team as my responsibility. The previous decade had made me much more deliberate about accepting it.

What I'm taking into the next chapter

Looking back, the biggest change is what I pay attention to. I arrived with a picture of startup life and a lot of enthusiasm for the work. Now I look more closely at what makes that work sustainable: whether people can grow, whether customers can buy, and whether the opportunity makes sense for the people putting their years into it.

The good parts of those ten years are still why I want to build. I remember the excitement of the early team, the freedom of creating the AI group, and the people I get to work with again. The harder parts changed the questions I ask before committing, and the choices I want to make once I'm there. Both belong in the story.

Becoming a CTO means taking responsibility for putting those lessons into practice. I can have clear ideas about trust, culture and hiring, but the job is deciding what to do when the needs of the team, the product and the business pull in different directions. Seeing those decisions from inside a growing company gave me a useful education. Now more of them are mine to make.

That's where Becoming a CTO picks up: what happened when these convictions met the day-to-day job. The series follows my first year through delegation, team building, AI, compliance, mistakes I was confident about, and the work I kept taking on myself. It starts with Nobody Handed Me a Manual.

A few useful reads

• Ben Horowitz, The Hard Thing About Hard Things. The book I'd put alongside the downturn in this story, when the business still needs decisions and the team has already been through a lot.
• Nasdaq Private Market, Who can sell in a tender, and how much?. A useful explanation of why an opportunity to sell shares can be open to some employees and limited for others. It puts practical questions behind the word “liquidity”.
• Holloway, The Holloway Guide to Equity Compensation. The reference I wish I had read before signing my first option grant: vesting, exercise windows, preferences and what each one does to the number on the headline.