At Revolut, innovation is a rigorous, rhythmic, and scalable process. At the antipodes of long and centralized development cycles, Nik Storonsky claims an approach based on repetition, volume and selection by figures. In three years, the Fintech has launched 27 “Bets” products, five of which have established themselves as major growth relays. Autopsy of a high frequency innovation model.
A product philosophy based on structured experimentation
The Revolut Product team does not advance to intuition. Each launch is based on a logic of bet-a BET-led by an autonomous micro-team of a dozen people. These teams work as independent units, with an experimental budget evaluated between 2 and 3 million pounds sterling per year.
“” We can lead 20 Paris in parallel without friction, because each team is independent “Explains Nik Storonsky. The challenge is to test at controlled cost, without diluting attention or overloading existing structures.
Clear success criteria from the first months
The system is based on a simple rule, namely that products must demonstrate measurable traction in the first three months. If the curves do not go up quickly, the teams are not reinforced. Conversely, when a bet takes off, the company unlocks additional resources to speed up.
“” We compare the income generated by each BET on the first four months. The top 30-40 % obtains the right to scaler. A team becomes two, then three. And it becomes a department. »»
Among recent winning bets, the ESIM offer, praised for its ease of use, or Revoints, the integrated loyalty system.
A granular piloting tool
Revolut has developed an internal interface to control its Paris portfolio. Each BET is followed in Jira, aggregated in dashboards which centralize performance (gross income, growth rate, adoption). The progress state (development, production, scaling) is also visualized there.
This dashboard allows management teams to make quick decisions without depending on heavy committees. The logic is Data-Driven, but located: the arbitrations are taken produced by product, without monolithic vision.
An assumed success rate, an optimized return
Of the 27 Paris launched, around 5 were clear successes, 5 to 6 failed, and the rest is between the two. This return is consistent with a “Venture Studio” type approach: few winning projects, but with a strong growth lever. “” If we look at the generation of global cash, the wallet is exponential He says.
The initial filter is not based solely on the perceived potential. Revolut consults a committee of former product managers. When the consensus is positive, the project is launched. If the opinions are shared, it is launched anyway.
An ability to kill quickly what does not work
The case of the “Salary Advance” product illustrates this culture of dry iteration. Designed as an alternative to consumer credit, the product made it possible to unlock part of the salary before the end of the month. The concept did not take, the adoption funnel being too complex: sale to the company, then to its employees, integration with payroll software.
“” For many efforts, we only had a few hundred users. We arrested », Storonsky slice.
A model apart in industry
This “Multi-Bets” system makes an exception in the banking universe. Where the neobancs were at low noise on a unique product, Revolut built a complete portfolio of services – credits, ESIM, loyalty points, trading, etc. – With a logic of modular expansion.
The key lies in team independence, test speed, and selection discipline. Each product is judged on its intrinsic performance, and only the best access the means of massive deployment.
📌 In summary:
| Element | Key data |
|---|---|
| Paris launched in 3 years | 27 |
| Paris led in parallel | 20+ |
| Average budget by bet | 2 to 3 million £ |
| Clear success rate | 5 of 27 |
| Scaling mode | Team progression → Department |
| Selection criterion | Traction from month 1-3, gross margin, adoption |
