Selected Work · Reflection 02 · Danske Bank
What Counts as Done
A pilot shipped in five months, and it was real: a working dashboard, live FX pricing, an early cashflow forecast. What it didn’t do was let anyone move money – the one thing “viable” was supposed to guarantee. This is about how a word absorbed that gap instead of the deadline moving to close it.
A launch was promised for June before anyone had agreed exactly what the product needed to do. The commitment preceded the definition. By the time the scope was being defined, the date had already become a constraint.
What we delivered in June was real. Pilot customers had an actual dashboard: account balances, an FX rate widget, an early but genuinely capable version of cashflow forecasting, and embedded access to an existing FX trading platform. Given that almost none of the surrounding infrastructure existed a few months earlier, that was a considerable achievement, delivered on a fixed timeline while the foundations were still being established.
What it did not yet let anyone do was move money. No payments, no way to settle a transaction: the basic act a banking relationship exists for. A viable banking product is one you can bank with. What we had was closer to a rich, working demonstration: proof the idea held together and the underlying technology and integrations could work. That was valuable. It just was not the same as viable, and the gap between the two mattered more than it appeared at the time.
The tension became visible once we started using the term minimum marketable product, intended to distinguish a pilot from something customers could actually depend on. But as the date approached, the boundary proved easier to redraw than to hold. The question quietly shifted from what does a customer need to be able to do for this to be viable to what can we credibly put in front of customers by the date we already committed to. A word intended to create clarity had become something the organisation negotiated around instead.
Holding to the stricter definition would have meant reopening the commitment and slowing a programme that was still proving it could exist. I supported continuing the momentum. At the time, that felt useful. In retrospect, it also made the original commitment harder to evaluate honestly.
Continuing the work did produce something valuable: the programme went on to add real payment capability, arriving eventually at something closer to what viable should have meant from the outset. But the moving definition also obscured how long that actually took: a real cost, never quite named, because the word kept absorbing it a little at a time.
Before committing to a fixed date now, I want the customer outcome settled first: what done actually means, for which customer, completing which task. A date is easy to agree. The real commitment lies in defining what the customer must be able to do when it arrives.