Builder --> Architect CEO

The shifting role of CEO as a startup begins to grow

Though I’ll send it in 2021, I’m actually writing this on New Year’s Eve 2020, which is surely as good a day as any to reflect. This will be a bit introspective, and to make matters worse, I’m going to mix my metaphors. Wish me luck, and Happy New Year wherever you are!

Being a startup founder feels a bit like playing a platform game, where you have to defeat bosses to get to the next level. The first two bosses are product-market fit and early-stage fundraising (and can be tackled in either order, or repeatedly attempted over a few cycles). Unlike typical games that start easy and progressively ratchet up, both are very difficult to beat.

I write this from the middle of Level 3, and the boss is a different beast entirely. Along the way, it involves more technical challenges like revenue attribution tracking and HR policies and consolidated global accounts and data warehousing. Meanwhile, the weapons required to defeat the next VC boss in the battle of growth-stage fundraising are repeatable growth levers and quantifiable moats and strong cohort net revenue retention and bottom-up TAM projections and so on. They’re totally different, and in the old days of venture capital (…I’m told!), the founder would very rarely have remained the CEO during the transition.

To make it psychologically harder, every level feels like you’ve accidentally got to the last boss in the game, and all these amazing gamers around you somehow got a cheat code you didn’t (fellow founders raising $3mn pre-seed rounds, or $30mn Series A, or whatever).

My biggest learning from the last few months is that the transition from Level 2 to Level 3 involves a role change from the CEO. In Levels 1 & 2, the CEO can work as a player in the game, with a rag-tag team of other strong, passionate players on the field with them. Level 3 requires the CEO to work at a different level.

Builder vs Architect

Time for a second metaphor. I’ve heard this referred to as the move from ‘product builder’ to ‘company builder’, or ‘founder to CEO’, but let me give this a spin: it’s the move from builder to architect. With a few bricks and a passionate team, anyone can be a builder. You can build a simple hut fairly fast, without worrying about the foundations, and it’ll stand up and stop most of the rain coming in. But with tons of bricks, a big team of builders, foundations to lay, and a plan to turn the hut into a skyscraper (and then a city of them), the leader has to act as an architect, and become a great one. The challenge is that learning to be an architect is really difficult, and stopping oneself from leaping in and trying to lay the bricks is surprisingly difficult, too.

The inescapable changes brought about by a larger team and fast-growing business in the last 12 months forced me to learn a new set of things very quickly. For Lingumi, a lot of this learning was packed in simultaneously in the last three months, as the little hut we’d built began to reveal its weaknesses under the duress of very fast company and team growth. We’d grown the business from a low six-figure annual run rate to 10x that in 10 months, and the team has grown fast to keep up with it. These are what people at VC dinners call “good problems to have” and make for good founder-brags or fundraising pitches or ego boosts. However, at ground-level, for the majority of the team, any problem feels like a problem. They’re all painful. We needed to lay foundations for our hut as it grew upwards, while putting out a few fires and patching the roof, very quickly. (Proper) budgeting, exec hiring, quarterly planning across several teams, fundraise planning, tricky HR issues, and increasing legal, people & financial complexity force a lot of learning very fast, and pulled me away from any semblance of being a builder on the site, laying bricks each day.

Reflecting on my leadership in Q4 2020

I learn quickly, and enjoy nitty-gritty, complex stuff, so found the end of 2020 an enjoyable if hectic quarter. A member of my team, when describing this state of chaotic, high-uncertainty multi-variate planning told me that “deep learning is never comfortable”, which I love and will steal! At the end of it, I needed a proper Christmas holiday, and have been using this holiday to read the novels I’ve wanted to read for months, and to reflect on the quarter and year.

My main reflections have been on my own ‘fails’, and the weaknesses they reveal. When problems arise, my team sometimes hear me say “the fish rots from the head”, which is an expression I learned in Russia, though many languages and countries have their variants (鱼从头烂 !). I firmly believe that the CEO is accountable for the problems that the team or company is struggling with, even if the problem only emerges lower in the org chart. Indeed, unless the CEO designs the feedback culture in the team and the board to be very free-flowing, it can at times be too easy a ride due to low accountability - after all, who manages the CEO? Who gives the frank feedback? I push for feedback from my team, but it’s easy to get lazy about this. This newsletter is part of my public accountability system, a way I hold myself to a high standard, and encourage my team and wider group of advisors and shareholders to do the same. Many of them read this.

A selection of things I’ve done badly

At Lingumi, Q4 planning was made unnecessarily stressful for my team because we started budgeting and strategy-setting for 2021 too late. Why? Because I didn’t kick it off early enough. The wider team were reporting a lack of collaboration on company OKRs. Why? Because we’d set too many goals, pushing teams into their own channels to achieve team OKRs, at the expense of collaborating on a few, critical goals. I was ultimately accountable for that goal-setting. Every article I’ve ever read has said “hire senior leaders six months earlier than you think you need to”, yet we began the hiring process for a Chief Commercial Officer, a critical senior role, too late. Why? Because I was too deep into the day to day to think about the quarter-to-quarter at a really practical, what-needs-to-have-happened-by-when-and-who-do-we-need-in-the-team-to-achieve-it level. It’s easy to argue that these things are only clear with hindsight, but I intend to improve my foresight.

All of these share the same ‘why’: I spent the beginning of Level 3 playing the game with my Level 2 avatar and weapons. I bet this is a classic situation as companies scale. Unlike in gaming, you don’t see the end of one level up ahead and plan accordingly, nor can you watch yourself spawn into the new level, or pick new weapons on a menu. It’s much more of a gradient…but the signs are always there. The speed at which we grew in 2020 is a reason, but not an excuse, for not focusing on those signs; I was laying too many bricks to draw architectural plans. Reading back my previous post on mistakes I made at seed stage, lessons 8 and 10 have hung around. If I could reflect the advice about senior team hiring onto myself, it would be “the CEO needs to begin the transition to the next level six months earlier than feels natural”. Creating some external accountability here (having a great coach, or close non-exec) might help trigger this, but only the CEO can actually make the necessary changes. That’s on me.

The idea that “companies ship the org chart” is very relevant to this situation, even though I’m talking about shipping things internally. The annual strategy needs to be tested with, and then ‘shipped’ to the team, the budget to the board, etc. The quality and clarity of what we were shipping - or not shipping - at an internal planning level was reflective of an org chart that has a first-time founding CEO at the top, making first-timer mistakes. QED.

I’m not beating myself up here - nothing went wrong, it was just less smooth than I’d have liked, and it’s my fault if I or the team felt more tired going into Christmas than we’d hope for (with the added strains of a very intense year for us, and a difficult year for the world in general). Regardless, there’s not much point in beating myself up. Any judgements I make on my performance in the last months reflects the success of my work in the window between 6 and 12 months ago, because its outcomes were only visible in the last quarter. 6-12 months (and longer) is the gestation period for a CEO’s decisions and output as the company scales up. This is also my first growth cycle. If I were scoring an A, it’d be first time lucky. The hiccups and chaos are essential for achieving that “deep learning” state. Making the mistakes, rather than reading books or blogs about others who have made them (sorry, readers), is the way we learn this practical stuff. If a first-time CEO can navigate from Level 2 to Level 3 with perfect timing, poise, and with no internal hiccups, they either got lucky, or are some form of management deity.

What I can do is use these reflections to do better in 2021, and to push myself to fully be the CEO that our company - and mission - needs for it to succeed.

Holding myself accountable for success in Level 3

Enough navel gazing! Time to look forward, as the New Year looms. If everything I’ve written above sinks in, the evidence will be my solid, well-prepared high-level plan for June-December 2021. It’ll contain something like:

  1. Quantified strategic “trigger points” when our strategy for 2021 will measurably be showing its success or failure, and we need to iterate or push harder on it

  2. A skeleton plan for my team offsites and board meetings, and how the outcomes of trigger points above will play into those plans

  3. Finding, hiring and onboarding (onboarding really well) the senior leaders we’ll need to successfully navigate Level 3 and move Lingumi to ‘Level 4’, whether internal or external

  4. The group of other CEOs and senior leaders across different functions at Series B startups I’m going to spend time with to understand what excellent looks like across different functions at the next stage

  5. How I’ll need to change my own behaviours, my team, and my routine to continue navigating Level 3 (more team growth, fundraising, greater business complexity), and when I’ll need to have made those changes

  6. What level of performance and change I expect from my team members in the next 12 months. Over-communicating that to them, with regular, frank feedback on progress

Though inspired by the NFX piece from which the graphic above is taken, this list comes from my head. It’s probably missing a lot, but feels very specific to me, rather than plucked from a blog. These are all requirements that feel very different from what I became good at as a ‘builder’. I’ll have the same title, but a new job.

I’ve worked on some of these already, and others only in my head, or not at all. They are and should be my top priority. But in the hurly-burly of January, I need to hold myself accountable to not losing sight of what excellent looks like internally at this and the next stage, and pulling myself and the team into that shape. Looking back, the urgent can often supersede the important too consistently to hold open the space for this type of long-term work. Knowing myself, I will have to sternly resist, and time-block accordingly, to avoid this happening.

How might this timeline change over time if Lingumi continues to succeed in our mission? At Level 25 (or whichever level of my metaphorical game you get to when your market cap is nearly a trillion dollars), Jeff Bezos’ version of this timeline is a three-year one:

"Friends congratulate me after a quarterly-earnings announcement and say, 'Good job, great quarter,'" Bezos said to Forbes. "And I'll say, 'Thank you, but that quarter was baked three years ago.' I'm working on a quarter that'll happen in 2021 right now."

- Bezos, quoted in Business Insider in 2018

Regular site inspections are table stakes

This isn’t to say I’m ignoring the short-term, Q1 and Q2. It’s just that the next six months’ successes and fails will be the outcome of the plan we made in Q4 2020. The subtle change I need to recognise is this: the execution of Q1-Q2 needs to be run by my team, whom I need to hold accountable for that (something I’m working on too). It sounds so simple, but is a huge shift in how I think about my work. To bang the same drum again, my job is no longer to lay the bricks! Indeed, beginning to do so would hinder the team’s space to grow and make mistakes…and if we’ve hired the right people, I’d do a much worse job anyway.

As such, I think my active input in Q1 in particular should only be:

  1. Monitoring progress against our North Star Metric and company OKRs

  2. Holding my team accountable for a very high standard of execution and leadership, and being extremely frank if this is not being met

  3. Spotting early if the strategy we set was the wrong one, and mobilising the company extremely fast to re-plan

Those feel very cookie-cutter and simple, but I suppose they are - they’re my table stakes. To return to the building metaphor, maybe it is the CEO’s job to make regular ‘site visits’, but not to become the site manager, making the daily decisions. Or maybe I’m over-stretching my threadbare metaphor.

De-glamorising startup leadership

What’s never mentioned in the glamorisation of startups is that the senior team, if they are capable and decide to stay on through different levels, have to step away from designing stuff and standing at the whiteboard and writing the code each day, and learn new skills like “how to be an effective manager” and “how to run a board” and “how to prepare a budget” and “how to spot problems in a balance sheet”, and and “how to handle really complex people problems with total calm”. They all require a load of work, and avoiding becoming a corporate shill while navigating the transition is probably going to be harder still. For loads of startup leaders, those are probably just not as fun or as fulfilling as the early stuff. As a result, and maybe this is controversial, I think staying on at different levels should be a pro-active decision, made carefully and with board and team input, rather than a presumed continuation. Reed Hastings explicitly tells his board that they should fire him if they find a better CEO. This, along with many layers of how Netflix operate (which I recently read about in his new book), is admirable, though at points rather doctrinal. I will ask Lingumi’s board to hold me to that same standard.

Writing this, I’m already feeling excited for next year. I enjoy being perpetually uncomfortable and pushed to the edge of what I know or can do. I embrace the paranoia and uncertainty of startups with some glee. This is probably not natural - maybe I licked the floor too much as a baby or something. Every month, and quarter, and year, the job changes, or the next level arrives, and the person in that hard, bumpy CEO seat has to adapt very fast, and very effectively, to succeed. It’s definitely fine to not enjoy that, but I’m loving the discomfort of the “deep learning” so far.

Happy New Year! Here’s to a 2021 less full of Covid, and just as full of learning. Thanks for reading and sharing this as ever!