Formula One Racing Team Speeds to Agile Development

You hear a lot about organizations that must make split-second decisions or risk losing to competitors. Split-seconds? That’s usually an exaggeration, unless you’re talking about Formula One racing.

During the 90 to 120 seconds it takes to run a lap in Formula One, engineers in the pit crew use telemetry and business analytics tools to find adjustments to make on a car racing at speeds of more than 185 mph. Every fraction of a second does count, says Graeme Hackland, IT director for Lotus F1, a three-driver team co-sponsored by Lotus Cars and Renault. At a recent race in Bahrain, a Lotus F1 driver lost first place by just three seconds.

No one says IT was to blame—the driver said afterward that he used his brakes at the wrong moment. But Hackland says he wants to make sure that IT runs friction-free and that application development, in particular, goes better than it has. “If software is causing problems to end users or distracting them from their main job, that’s the worst possible scenario for us.”

Racing to Agile Development

Lotus F1 recently ditched traditional application development methods for agile development, including scrum and visualization. The goal is to produce better tools faster for both pit crew and car designers. Hackland says he also wanted to erase the perception that sometimes IT doesn’t understand everything about how the pit crew and car designers like to work. “They wanted us to deliver more right the first time,” he says.

The continuous communication between developers and users under agile development can clarify how the end product should look, feel and work much earlier in the process, says Margo Visitacion, an analyst at Forrester Research. But organizations underestimate the culture shock of going agile. “Collaboration is a lot different from taking requirements, throwing them over a wall and expecting perfection many months later,” she says.

The Lotus F1 team has finished fifth on the Formula One racing circuit for the past two years. Now they’re driving toward a championship within the next couple of years, Hackland says. There’s “an outside chance” the team could take it this year, Lotus Ambassador Emerson Fittipaldi has said.

In Formula One racing, 12 teams compete in 20 races every two weeks in cities around the world between March and November. Each team designs and builds its own car. Some specifications and rules change every year. In 2014, for example, turbocharged engines and new fuel-flow restrictions will be mandated. Lotus F1 is working on those long-term designs. But the team is also focused on the minute-by-minute rush of racing today. For Hackland, that means pumping out faster and more sophisticated analytics tools to give the team an edge.

Half of his 14-member IT group is dedicated to software development. Even so, the traditional waterfall development method the team’s developers had been using could get slow, he says. Gather requirements, create a functional specification, create a technical specification, build and test. The ultimate users of the system were involved mainly at the beginning and end, but not during the whole process.

Yet it’s all that time in between when small corrections and additions can occur that make the end product much better, says Visitacion at Forrester. “When you see something tangible faster, it creates a stronger sense of reality of what’s going to be valuable to you and what’s going to be extraneous. It allows the team to focus on the essence of the project.”

Under Lotus F1’s new agile approach, users attend daily meetings with developers and look at progress every three weeks. Testers are involved all the way through. “By the time we deliver, there’s been a huge amount of interaction, which we never used to have,” Hackland says.

Visualization in the Pit

His team recently used visualization software from iRise to improve a key software tool: a race strategy system that helps the crew determine when to pit cars. The tool considers factors such as lap number and speed, position in the field, wear on the tires, what competitors are doing and data from the 200 sensors on Lotus F1’s cars. It can calculate 10,000 race simulations in the time it takes to run a lap. If crew engineers have to fiddle with a clunky user interface or wait for an analytics result that takes a few minutes to compute, they waste time–and opportunity, Hackland says.

The visualization his developers produced with iRise let pit crew members interact with new features in the user interface, giving feedback to incorporate into the finished product. Testers could try the unfinished code to get a better idea of whether it would work for them. In the old way, users may just have seen static screen shots, or nothing at all, during development. “It was difficult for users to say whether they liked it or found it easy to use,” he says.

Hackland says he expects his staff to refine its approach to agile as work continues. “If [IT changes] reduce the amount of manual work engineers have to do, they can focus on the car and driver,” he says. The team, meanwhile, prepares for the next two races in Spain and Monaco this month. Maybe Lotus F1 will find that three seconds.

Kim Nash is a senior editor for CIO Magazine. Follow her on Twitter @knash99

Follow everything from CIO.com on Twitter @CIOonline, on Facebook, and on Google +.