Scaling the world's largest platforms by making them simpler.

Early in my career I assumed leadership was mostly about being right: the sharpest analysis, the best plan, the clearest answer. Over time I learned that being right is necessary but nowhere near sufficient. People don't commit to a plan because it's logically correct. They commit because they trust the person asking, and because they believe their effort will be seen and will matter. Empathy is what builds that belief.

People choose to follow leaders they trust and respect, and that has hard operational consequences. When a team genuinely shares a mission, you spend far less energy on coordination, politics and rework, and far more on the actual problem. That shared commitment compounds: faster decisions, higher retention of your best people, and a willingness to raise the difficult issue early rather than hide it until it's expensive. It is, quite literally, a driver of business performance, not a tax on it.

Empathy in practice is not about being soft on standards. The opposite, in fact. The best leaders hold a very high bar and a very high level of care at the same time. They tell people the truth, including the uncomfortable parts, because they respect them enough to be honest. They give real ownership rather than hovering. And they pay attention to the human context behind the work, because someone who feels understood will run through walls, and someone who feels like a resource will do the minimum and look for the exit.

This is why I care so much about mentoring the next generation of leaders. Most people are promoted for individual excellence and then discover the job has quietly changed underneath them: it is no longer about their own output, but about creating the conditions for others to do their best work. That shift doesn't happen by accident. It can be taught, modelled and practised. Leadership, in the end, is less about personal achievement and more about building environments where people and performance thrive together.

Most growth plans are addition plans. A new feature, a new segment, a new market, a new partnership. Each one looks like upside on a slide. In reality, every addition is a liability you carry forever: more surface area to maintain, more decisions to coordinate, more ways for a customer to get lost. Complexity compounds quietly until one day the organisation is spending most of its energy managing itself.

Simplification flips that maths. When you remove a step from a checkout flow, conversion rises across every market at once. When you collapse three half-built systems into one, every team downstream gets faster. When you say no to a feature, you free the most expensive resource you have - your best people's attention - to spend on the thing that actually matters.

This isn't minimalism for its own sake. It's a recognition that clarity is leverage. A simple product is easier to explain, which makes it easier to sell. It's easier to operate, which makes it cheaper to run. It's easier to reason about, which makes it safer to change. Those advantages stack, and they stack in your favour every single day the product is live.

The hard part is that subtraction has no champion. Nobody gets promoted for the feature they killed. So simplification has to be led from the top, as an explicit strategy with its own goals, not as a clean-up exercise squeezed between launches. The leaders who treat "what can we remove?" as seriously as "what can we ship?" are the ones whose platforms keep growing long after their competitors have collapsed under their own weight.

There's a comfortable version of product leadership that lives entirely inside the product: backlogs, discovery, design reviews, ship dates. It's necessary work, but on its own it produces leaders who can describe what they built and not what it earned. At scale, that gap becomes career-limiting and company-limiting at the same time.

The operators I respect start somewhere else. They know the unit economics cold. They can tell you which markets carry the business and which ones cost more than they return. They understand how the product connects to revenue, to cost, to risk, and to the commitments leadership has made to the board. When they prioritise, they're not just ranking features - they're allocating capital.

This doesn't mean becoming a spreadsheet. It means holding two truths at once: that you are responsible for the customer's experience, and that you are responsible for the business that experience exists to sustain. The trap is treating those as opposing forces. Done well, they're the same force - a genuinely better product almost always shows up in the numbers eventually, and a healthy P&L is what buys you the room to keep improving the product.

For anyone aiming at a CPO seat or a board, this is the threshold. The boardroom doesn't ask what you shipped; it asks what changed because you shipped it. Learn to answer in the language of the business, without losing the instinct for the user, and you become the rare leader who can be trusted with the whole thing.

Few companies choose to be fragmented. They get there honestly. A market needs a special case, so you fork the code. An acquisition comes with its own platform, so you keep it running "for now." A big customer wants something bespoke, so you build a one-off. Each decision is reasonable in isolation. Together, they produce an estate no single person fully understands.

The cost shows up later, and it shows up everywhere. A change that should take a week takes a quarter, because it has to be made five times in five subtly different systems. Talent burns out maintaining duplicates instead of building. Data can't be trusted because every system defines a customer slightly differently. Worst of all, the organisation slowly loses the ability to move quickly - which, for a platform, is the only durable advantage there is.

The instinct is to fix this with a grand replatforming programme. Sometimes that's right, but more often it just trades one risk for a bigger one. What works better is treating consolidation as continuous discipline: a clear architectural spine that every market and acquisition is expected to converge onto, a high bar for creating new exceptions, and the patience to retire old systems deliberately rather than letting them linger.

The deeper point is cultural. Fragmentation is what happens when speed today is always allowed to borrow against speed tomorrow. The leaders who keep platforms coherent are the ones willing to be slightly slower now to stay fast for years. That trade is invisible on any single roadmap - and it's the difference between a platform that scales and one that seizes up.

The most common mistake in going global is to treat each new country as a fresh build. The second most common is the opposite - to assume one market's product will work everywhere if you just translate the labels. Both fail for the same reason: they skip the actual work, which is deciding what is universal and what is local, and being ruthless about the difference.

Across dozens of country sites and many languages, I learned to draw that line explicitly. The core - the data model, the commerce engine, the design system, the way the platform thinks about a customer - stays the same everywhere. It's the asset. The edges - payment methods, regulatory and tax rules, language, cultural expectations around content and trust - are where you let the product flex. Everything else is overhead masquerading as localisation.

This needs a playbook, not heroics. A repeatable way to enter a market, a shared platform that absorbs local requirements as configuration rather than code, and a team structure that doesn't reinvent the wheel in each region. When that's in place, the marginal cost of the next country falls instead of rising - which is the entire economic point of being a platform in the first place.

And it needs respect. The teams closest to a market usually read it better than headquarters ever will - which is exactly why they should be the ones judging where the global product fits and where it has to bend. The job of a global product leader is to give them a foundation strong enough to build on and the freedom to adapt at the edges - then to resist the urge to make everything a special case. Get that balance right and scale stops being a tax and starts being a moat.

When you serve customers across dozens of countries, the limits of a homogeneous team show up fast. A flow that feels obvious in one culture confuses users in another. A payment assumption that holds in one market breaks in the next. A tone that reads as friendly to one audience reads as flippant to another. You can discover these things slowly and expensively through failed launches - or quickly and cheaply by having people in the room who already see them.

That's the practical case for diversity, and it's the one I make most often, because it's the one that's hardest to argue with. Teams built from different backgrounds, geographies, disciplines and lived experiences simply have a wider field of view. They catch more of reality before it catches you. For a platform whose addressable market is the planet, that breadth isn't a nice-to-have - it's a direct input to product quality.

There's a decision-making case too. Homogeneous groups converge quickly and feel confident doing it, which is exactly when blind spots become expensive. A bit of well-managed friction - people willing to ask why, from genuinely different starting points - produces slower meetings and better outcomes. The leader's job is to make that friction safe and productive, not to engineer it away in the name of harmony.

None of this happens by accident or by statement. It comes from how you hire, who you promote, whose objections carry weight, and whether the quiet voice in the corner gets heard before the launch instead of after it. Build the team that looks like the world you're building for, and the products tend to follow.

Most commerce is forgiving. Demand arrives spread across the day, and a slow moment here or there costs little. High-demand commerce is the opposite. An on-sale concentrates a month of traffic into a few minutes; everyone arrives at once, motivated, impatient, and watching. In that window, the experience the platform delivers is the experience the brand becomes - and there is no second chance to make it.

Engineering for that reality changes how you think about everything. You design for the worst minute, not the average one. You assume queues will form and make the wait honest and fair rather than hiding it. You decide in advance which parts of the system can gracefully degrade and which must never fail. You rehearse the spike before it happens, because the moment itself is no time to be learning how your platform behaves under load.

The lessons generalise far beyond ticketing. Any platform built on moments of intense, simultaneous demand - a drop, a sale, a launch, a market open - lives or dies on resilience. And resilience is not something you bolt on at the end. It's a set of product decisions about fairness, transparency and what the customer sees when things are at their hardest, made long before the traffic arrives.

What I keep returning to is that under peak load, your values become visible. Whether the system is fair, whether the communication is honest, whether the experience respects the customer when it would be easier not to - that's all on display in the one minute that matters most. Get it right and you earn trust that no marketing can buy. Get it wrong and it's the screenshot everyone remembers.

Leading a small product team and leading a global one are different jobs that happen to share a title. At a certain size, you stop being the person who makes the decisions and become the person who makes sure good decisions get made - by people you'll see twice a year, in time zones that barely overlap, working on problems you'll never touch directly. The tools that got you there stop working, and most of the failures I've seen at scale come from leaders who didn't notice in time.

What replaces direct control is clarity of intent. If the team understands what matters and why - the strategy, the few things that are non-negotiable, the bar for quality - they can make thousands of good local calls without you. If they don't, no amount of process will save you; you'll just be the bottleneck on everything, slowing the whole org to your own throughput. Alignment, not oversight, is what scales.

Distance also changes the texture of trust. When you can't read the room, you have to be deliberate about candour, about writing things down, about making space for the objection from the region that headquarters didn't see coming. The best ideas rarely originate at the centre, and a global leader's real value is often in moving insight from the edge to the middle faster than the org chart would allow.

And it changes what you optimise for in people. At scale you're not hiring for output, you're hiring for judgment - people you can hand ambiguity to and trust to resolve it well. Set the bar high, give real ownership, then get out of the way. The leaders who can do that, across cultures and continents, are the ones whose teams keep performing whether or not they're in the room. That, in the end, is the whole job.