Setting up product from scratch at a neobank

I have started product functions before, but never inside a regulated bank. At BUUT, ABN AMRO's neobank, the work was not a clean greenfield. It was a complex family banking product with parent and child journeys, real regulatory weight, and no shared picture of how a new customer actually became an active one. My job was to stand up the onboarding and settings area, then grow it into something larger: an operating system for activation and early retention. This is how I did it, and what I would tell anyone setting up a product function from scratch.

Start where the risk is highest

I did not start with the feature backlog. I started with onboarding, because that is where the money, the regulation and the customer trust all collide at once. Onboarding and settings alone ran to more than 150 screens across parent and child journeys, every one of them touching identity, money or consent. When the highest-risk surface is also the first thing a customer sees, you do not get to treat it as a later problem. You lead from there, because everything downstream inherits its quality.

Starting at the riskiest point also forces clarity early. You cannot bluff your way through a KYC flow. It made me confront the real constraints in week one instead of discovering them after I had shipped something that compliance could never sign off.

Make the invisible funnel visible

The biggest problem was not a broken feature. It was that nobody could see the funnel. People talked about activation as a feeling, not a sequence. So I defined it as one: KYC passed, child added, allowance set, first transaction made. Four concrete steps, each measurable, each a place where a real person was getting stuck.

Once the funnel was visible, the arguments changed. We stopped debating opinions about what users wanted and started looking at where they actually dropped off. That single shift, from anecdote to funnel, is what later let us lift onboarding conversion by more than 70 per cent and cut support demand by roughly half. You cannot improve what you refuse to name.

The arc from onboarding to retention

Onboarding gets a customer through the door. It does not make them stay. The interesting work was extending the scope past the signup moment into activation and early retention, treating them as one continuous arc rather than three separate teams' problems.

A child added but never given an allowance is not activated. A first transaction that never gets a second is not retention. Once I owned the whole arc, I could prioritise the steps that actually moved a customer from registered to genuinely using the product, rather than optimising one stage at the expense of the next.

Structure beats heroics

When a function has no operating system, it runs on heroics. Someone stays late, someone catches the bug, someone remembers the compliance requirement. That does not scale, and it burns people out. So I introduced structure where there was none: clear prioritisation, regular rituals, and a roadmap people could actually see.

Rituals sound boring next to shipping. They are what let you ship. A predictable cadence of planning and review means decisions get made once, not relitigated weekly. Prioritisation that everyone understands means the team stops working on whatever shouted loudest. I would rather have a slightly less brilliant idea executed with structure than a brilliant one lost to chaos.

Leading across Risk and Compliance without losing pace

A neobank onboarding flow touches Product, Design, Engineering, Risk, Compliance and Operations. On this work I was aligning more than 20 stakeholders across those functions, and the easy failure mode is to let that turn into a permanent traffic jam where nothing moves without three sign-offs.

The trick is to bring Risk and Compliance in as early collaborators, not as a gate at the end. When they understand the intent up front, they help you find the compliant path instead of blocking the non-compliant one after you have built it. Pace in a regulated environment does not come from going around the constraints. It comes from designing with them from the first conversation.

A day-one playbook

If I were standing up a product function tomorrow, I would do it in the same order. Find the highest-risk surface and start there. Make the funnel explicit before touching any feature. Take ownership of the full arc from acquisition to retention, not a slice of it. Install the rituals and prioritisation before you need them, not after the chaos arrives. And treat the hard stakeholders, Risk and Compliance, as people who help you ship rather than people who stop you.

None of this is glamorous. Standing up a function is mostly about creating clarity where there was none and holding the line on structure when heroics would be easier. But it is the difference between a team that ships features and a team that compounds. I present that roadmap to the whole company at our all-hands now, and the thing I am proudest of is not any single number. It is that the funnel is no longer invisible.