Every financial services company wants to be innovative. Innovation brings new lines of business, increased revenue and decreased costs. But how can a bank or credit union encourage innovation amid the daily grind?
By definition, most institutions can’t move their headquarters to Silicon Valley. Changing your dress code may not produce results. Starting innovation labs and appointing innovation czars may help, or may turn into expensive “innovation theater.”
While these ideas may help to attract new talent, innovation truly begins with culture, and traditional banking culture works against innovation. That’s because innovative cultures accept that new ideas inevitably result in some failures — you might call it “the freedom to flop.” Innovative cultures embrace failure and the lessons it can teach, encouraging employees to try new ideas, take risks, and thus learn rapidly. Contrast this with most banking cultures — strongly influenced by regulatory, credit and operational risk management — and you can see the dilemma.
“The solution lies in containment. The tech industry has cultivated tools and processes that allow teams to mitigate the impact of failure while still moving quickly and learning.”
The solution lies in containment. The tech industry has cultivated tools and processes that allow teams to mitigate the impact of failure while still moving quickly and learning. Such processes allow fintech firms like Square and Stripe to move quickly while providing stable services for businesses that rely on them. And they can enable banks and credit unions and their regulators to work together to try new things while limiting potential negative impacts.
Here’s a “sampler platter” of containment tools and processes that can help make innovation work in your shop.
Learning from Failures and Innovating in Small Doses
Research and modeling represent two of financial services’ longstanding tools and they still work. With modern computing power and statistical machine learning techniques, simulation can go far beyond spreadsheets and business plans to reach into the realm of consumer research. Models can predict consumer behavior or analyze how new developments would have performed against past data. Data scientists can offer valuable insights, provided they have access to data in usable formats.
“Experimental results should be shared — regardless of how things turned out — which means broadcasting failure should be encouraged.”
Investing more in research and data analytics early in the development process can reduce the impact of failure. In addition to researching data collected from published sources, teams should have access to the results of any past experiments performed by the company to ensure that everyone leverages the lessons of past failures.
Experimental results should be shared — regardless of how things turned out — which means broadcasting failure should be encouraged.
Innovative companies love to try new things, but they don’t always want everyone to have access to the brightest and shiniest until these have been tested and refined. Again, containment. Tools like A/B testing, feature flags, and beta customers provide mechanisms to limit who has access to what features.
With A/B tests, a financial institution can separate people into groups and present different experiences, measuring their behavior. For example, if you want to understand which value proposition resonates most with consumers, present them with different messaging as part of the onboarding flow. Then measure how each message affects signups.
Similarly, feature flags act like switches, allowing entire pieces of functionality to be turned on or off for groups of users. With feature flags, before launching a new product feature to your entire customer base, you could target only consumers who show strong engagement characteristics and are more likely to be understanding of unpolished or even buggy functionality.
Beta customers choose to opt into new functions, aware the tech might not work as expected. These folks remain eager to try the newest features and give feedback or have their response monitored.
All of these tools act to limit the scope of impact a failure will cause. New features can be exposed to fractional percentages of users and their experience monitored closely. These experiments may take hours, days or even months to fully evaluate, but in the end, impact can be isolated to a small number of customers. With software, it’s fairly easy to implement these tools as long as it’s part of early design.
Of course, all such testing has to meet regulatory requirements. For companies wanting to try innovations that might not fit existing rules, regulatory sandboxes — where available — present a similar approach to financial regulatory challenges. By providing a space for institutions to experiment with new approaches in a small portion of their business and allowing regulators increased access to monitor these experiments, sandboxes provide a mechanism to mitigate risk while encouraging innovation.
- 4 Myths Preventing More Fintech+Banking Partnerships
- 7 Financial Institutions Taking Innovation Labs to the Next Level
Good Brakes and ‘Fast Rollbacks’ Improve Innovation
The best way to be confident about moving forward is to make it easy to move backward. “Fast rollbacks” are a mainstay of software development — they allow teams to rapidly undo changes as soon as they realize those changes aren’t working. This safety net means bold changes can be executed with limited impact to the company. Rollbacks require tools to closely monitor the impact of changes, alarms when core metrics are not performing as expected, and a mechanism to quickly undo a change.
Financial institutions can leverage such tools to encourage safe innovation. By working with regulators to closely monitor new programs and define mechanisms to either shut them down or roll them into existing programs, institutions can try more and bolder ideas with far less fear of failure. Limiting access to these new initiatives, and allocating funds for failure scenarios can further contain and mitigate negative impacts.
Constant Measurement and Evaluation Are Musts
Key to all innovation is close monitoring and ongoing analysis of usage and impact. This requires easy access to metrics and data. Innovative companies employ daily emails and simple-to-use dashboards to provide analytic information to everyone working on a project. This can also help in sharing impact analyses with regulators.
It’s necessary to look backward from time to time to see what’s working well and what isn’t. Making small, measurable changes on products and programs is not enough. Innovation processes themselves should be examined regularly.
To accomplish that, tech companies use “retrospectives,” or “retros” — scheduled meetings focused on a project or period of work. During these meetings, team members share what they think went well and what didn’t go well. Then they brainstorm ways to improve outcomes. Critically, retros focus on the team and the process in addition to the results, meaning changes often address team skill gaps or process deficiencies.
During a retro, participants traditionally share observations on a white board, categorizing matters by “happy face” :-), “meh face” :-|, or “sad face” 🙁 icons. This flags things that went well, things that were neither good nor bad, and things that went poorly. Patterns emerge and problem areas are exposed.
Constant measurement and iteration are core to innovation processes. This allows companies to leverage lessons from failures and to embrace continuous change.
On the security side, the same focus on finding problems drives the practice of offering “bug bounties” — rewards to outsiders for identifying weaknesses and vulnerabilities. This brings in fresh eyes and also leverages the energies of a much larger community than the financial institution’s own development team.
Weaving Agility into the Innovation Process
At a high level, technology companies have nearly all adopted what we in fintech and regtech call “agile development.” This is where stakeholders, designers, and engineers work closely to quickly build, evaluate, and rebuild new software.
Agile development arose in contrast to more traditional “waterfall” processes, in which stakeholders passed on requirements to designers, who in turn delivered completed designs to engineers for construction. Agile processes allow teams to be shifted rapidly, reacting to changes in business circumstances, new ideas, and new discoveries, all as the changes occur.
Agile development is defined by rapid iteration cycles. Cycles usually last a week or two, though they can be as long as a month. At the beginning of each cycle, work is prioritized based on the current state of the business. Throughout the cycle, the team gathers daily for 15-minute meetings called “standups.”
During a standup, attendees literally stand — to encourage a quick meeting — and share what they did yesterday, what they are planning to accomplish today, and what, if any, barriers stand in their way.
Standups keep everyone aware of the ongoing work, allow individuals to raise concerns or issues preventing their work from progressing, and provide an avenue for management to intervene if something’s veering off.
Like the overall agile process, many of the tools and processes discussed above work to tighten the flow of feedback between the development team and the impacts of their changes. By reducing the time between making an adjustment and understanding its impact, a team can reduce the magnitude of any potentially detrimental changes and move faster to deliver value.
In the financial world, closer collaboration with regulatory agencies and more frequent check-ins on new programs could allow institutions to move more quickly, while giving regulators confidence about their visibility into the impacts these new programs present.