Taking a lean startup approach

An interview with Richard Puckett, Development Manager from Concur Labs, on developing new concepts for Travel and Expense:


Q: Concur Labs launched recently to build prototypes. What does that mean?

A: What we’re doing is looking at what’s ahead at the newest technologies and building concepts to gather early feedback. Prototypes are meant to be built fast and allow room for experimentation. They can break and they can evolve; in fact this is necessary. Eventually a prototype will show itself to be viable, have potential, or simply serve as a test. In all cases, we’ll learn from the process and apply those learnings to the next problem.


Q: How do you formulate or capture new ideas?

A: Hackathons, meetups, lunch and learns. Several prototypes originated at hackathons. The 24-hour events are a marathon and an opportunity for uninterrupted development time. Teams try out new SDKs, TDKs, APIs and hardware to see what makes sense for users. We attend and sponsor internal hackathons to connect ideas from the inside, too.

For example, the prototype “Trace” enables you to expense miles right from your car dashboard. We first built this at a hackathon using Ford’s Technical Development Kit (TDK), a physical machine that allows us to develop as if we’re using a connected car. Through a series of trials and updates, we matured “Trace” into a Concur service that Ford Motor Company will offer through its SYNC 3 family of apps.


Q: Are there any preferred methods you use to develop new prototypes?

A: We rely on lean startup approach: We put something out and see if there’s a good fit. We talk with users, collect feedback, iterate and release another version. On a regular basis we’re putting out a Minimum Viable Product. We don't have a monthly release cycle; we don't have a bi-weekly cycle; we have a daily release cycle. What’s important for the learning process is to release new functionality with high frequency rather than scheduled predictability.


We use Amazon to deploy at scale continuously. This makes it possible to try new things, get user feedback and analytics, and evolve a prototype toward a better solution.

Q: What happens when a prototype doesn’t go as planned?

A: That’s part of the process. And it’s vital because we often don’t know the complexities around we’re trying to solve and these only become clear as we’re on the journey. One particularly relevant story is that after Edison invented the light bulb, a reporter asked him, "How does it feel to fail a thousand times?"

Edison replied: "I didn’t fail 1,000 times, the light bulb was an invention with 1,000 steps.”

Everything we do is a step. Whether it becomes a prototype or code sample, it's a step that helps us explore the space and get closer to creating an “expense report that writes itself” and an effortless “perfect trip.” We’re always seeking new possibilities.


Q: Can you tell us what’s coming next from Concur Labs?

A: We’re exploring connected cars, connected homes, machine learning and messaging. With “Trace,” there’s a connected car experience. The Concur IFTTT (If This Then That) beta, ifttt.com/concur provides a connected home experience by letting you combine your favorite apps and services. This means that while you’re on a business trip, you can turn your house lights off or on, for example; or manage your Nest thermostat from your smartphone.


Also in development are new integrations with Amazon Echo and chat bots for customer service. Innovation is often a remix of existing solutions applied in new ways so we constantly challenge ourselves to look at problems (new and old) with fresh eyes.


Do you have ideas? Reach us at: concurlabs@concur.com and learn more at concurlabs.com.


Loading next article