Investing in developing digital tools is a critical strategy to attract one of the most coveted customer demographics: millennials. In March, research from bank tech provider Morphis found that nearly 92% of millennials said they would make a banking choice based on the bank's software and online services.

The good news is that a number of large banks, like Capital One and JPMorgan Chase, have dramatically increased their attention on creating easy-to-use digital experiences. The bad news is there are scores of institutions that have yet to evolve their services and/or don't know where to begin. More troubling, banks not only must compete with each other for this coveted demographic, but they increasingly compete with other nontraditional financial services providers that focus on mobile experiences.

According to research from a unit of Viacom, a third of millennials say they won't need a bank in the future. If big banks want to continue their early success with millennials, they will need to compete with fintech upstarts like Betterment and Acorns — both of which have come out of the starting gate with mobile-first service models.

To get ahead of incoming competition, banks must embrace these three startup principles.

Adopt startup speed
Traditional banks often leverage expensive and lengthy processes when they deploy new services. Fintech startups that attract millennials, meanwhile, are unafraid to debut new services even if they are not entirely polished at the outset. Banks need to adopt this fearlessness and stand up new offerings quickly, but at a smaller scale.

By building one feature of one potential service for, say, one branch, banks can use consumer responses to refine and eventually scale the project across other branches. In taking this approach to developing new products and services, banks may also help internal stakeholders feel more comfortable with the perceived risky innovation.

Gather feedback in real time
When identifying new services to deploy, today's banks typically use a blend of quantitative feedback, surveys and focus groups. These methods are often poorly timed, impersonal and, quite frankly, outdated. Focus groups are held in isolated environments that are divorced from the situations in which customers would ever use a banking service. Further, synthesizing feedback also happens after the fact, and therefore, doesn't provide the opportunity for a real back and forth with customers.

Instead, banks should run small, live experiments with prototypes that allow them to observe customers using the service. Co-creating along with a customer yields constructive feedback in real time.

Brainstorm legacy tech integration
Many banks today develop new services in labs that are completely isolated from other teams. This "innovation in a vacuum" approach is in total opposition to startup methodology. In order to test prototypes effectively, banks need to take into account how their existing infrastructure will be able to support the launch and growth of new services and integrate into the bank's existing service offerings. For example, many banks currently use legacy software systems. If a new app ties back to this system, the bank needs to test how they would work together.

Gordon Hui is vice president of strategy for Smart Design, a design and innovation consultancy.