Solving scale's organisational problem: six companies, one HQ

Scale makes you slow. Our answer: category-focused vertical companies with full authority, one HQ engine for G&A, and a fluid structure.

When you scale fast, a problem occurs: you get slow. A growing organization needs structural rules to hold everything together. And each rule means one more thing to pass through before anyone can act. You can go on believing you're still that lean startup from the first days. But reality hits hard. Past a certain headcount threshold, speed stops being something you can decide to have.

We are a case in point. Codeway is six years old and bootstrapped: $0 of outside investment, each new product funded by the ones before it. By now the numbers look like this:

worldwide users

600M+

products

60+

team of builders

400+

It's a good problem to have. But you still need to overcome the challenge to keep growing.

Big results stopped needing big teams

The great news is that the maths changed. In the age of AI, you don't need a big team to achieve a big result. A handful of people can now do what would have needed a department a few years ago. Our belief is that one person here has the multiplier effect of a couple of people at a company that isn't built around AI, and even more in some cases. It's an estimate, but also the guiding assumption we build on, and it shapes our hiring strategy.

There's no point pretending you're still that agile startup that moves fast, believing every visible hint of your well-defined culture and every promise your team's average age might give. It's been a long time since we passed that phase. What we need is a combination that might seem near-impossible to crack at first: be fast and act like an indie, but also keep achieving big numbers.

The pods were already most of the answer

Most of the structure for this already existed at Codeway. We ran agile, compact teams we called pods, bringing small groups together across the bigger Marketing, Product and Engineering teams. Apps went from idea to launch, and from launch to a million users in a week. Speed like that only happens when a small team can act on its own. The hard part is keeping it inside a company of this size.

So the question became: why stop there? The pods had the speed. What they didn't have was full authority and a name. The answer was to give them full independence and authority over the job they were already doing and the category they operate in. Not more coordination. More independence. That meant handing each one the whole game: product, growth, creative, AI and engineering.

We gave them the authority, and we gave them a name.

Vertical is a weird word, but it explains the logic

We called them verticals. They are companies, but we are not a venture fund. They are teams, but not exactly. They are actually more than teams. We thought vertical was the right word, for now, to explain it. There are six, each owning a category:

• Wishlabs, AI creativity. Home of Retake, the #1 AI face editing app in the US.
• Learna, language learning. One AI tutor that 80M+ people have used to build confidence in a new language.
• Sparked, education. 10+ learning apps spanning maths, study tools, practical skills and brain training, with 68M+ learners.
• Wellture, health, fitness and sleep, rebuilt around AI. Our youngest vertical, with 10+ products in its first year.
• Catalyst Apps, utility and productivity. Where Codeway began. Home of Cleanup, the #1 storage cleaner on iPhone, still #1 in its category six years on.
• Cosmic Jam, media. Our drama studio, with 70+ original series.

They are still part of one Codeway. What changed is the level of autonomy.

The engine none of the six has to build

Behind the verticals sits a whole working engine. It runs everything they shouldn't have to run themselves: HR, Finance, Workplace Ops and the rest of the general and administrative work (G&A), plus the shared platforms. Each vertical runs without friction. We call that engine Codeway HQ, one shared backbone with two halves.

• HQ Tech: Data, AI Research, Infrastructure, Security and IT. Shared platforms and portfolio-wide data, so a vertical never builds its own from zero.
• HQ Operations: Finance, Legal, HR, Talent, Mobility, Brand and Workplace. The G&A work that no team should have to run six times over.

All of that buys one thing: speed. That's why we can hold ourselves to a rule: a new app ships within a couple of weeks of spotting a gap. It's also how know-how moves horizontally. Voice started as Learna's problem, because a tutor that teaches by talking has to sound right in more than eighty languages, and it became a capability every product can reach for.

The verticals also keep one right: in time, any of them can set up its own G&A operations. HQ exists to power them, not to hold them.

Reality has already disagreed with the plan

On paper, this draws cleaner lines without losing the glue: one brand, one company's strength, under the Codeway roof. However, big organizational changes bring their own rigidity, and if you insist on one big change exactly as you drew it, you're screwed.

Ours hasn't held entirely either. Since we verticalized, people on the teams have changed, the question of how HQ should power the verticals is still an open-ended discussion, and we've even changed the category of one vertical. None of that means it isn't working.

The idea isn't a finished org chart. We expected to redraw it. The plan is only useful if it can keep changing with our needs.

The lines between small teams keep moving

AI keeps shifting the ratio towards small teams and individual builders, and that changes what a vertical's edges look like. One builder can power a couple of verticals from inside one of them. Two verticals might work far closer to each other than the chart suggests, while the others stay completely separate. We're fine with that. The point of the lines was never to keep people apart. It was so everyone knows what they own.

Scale makes you slow because it asks for rules. What we need is actually one rule: keep lean teams with authority, so they can have a bigger impact, not only on their work but on all the other small teams around them. Today that rule has six names, one HQ behind them, and a plan we'll keep rewriting. We'd rather be six fast companies than one slow one.