What A Drone Industry Rivalry Teaches Us About Organizational Styles

Who would dominate the drone industry? It was an open question back in 2012, as two fierce rivals battled for market share: Berkeley, California–based 3D Robotics (3DR), and Dà-Jiāng Innovation Science and Technology Co. (DJI), of Shenzhen, China.

Culturally and organizationally, the two ventures couldn’t have been more different. 3DR was flat and relied on the bottom-up ingenuity of a large open-source community of drone enthusiasts. DJI, meanwhile, was hierarchical and dominated by the vision of its founder, Frank Wang.

Those different organizing strategies grabbed PhD candidate Robert Bremner’s attention as he worked with his advisor Kathleen Eisenhardt, Stanford W. Ascherman M.D. Professor in Stanford’s Department of Management Science & Engineering. He began to examine how organizational design and experimentation can help or hinder ventures. “Depending on how you structure your company, that can play a big role in its success,” he says. In the face-off between 3DR and DJI, he had an ideal case study.

Bremner’s analysis, drawn from interviews he conducted, news articles and other public records, revealed that, although DJI proved to be the ultimate victor in the drone industry, each company’s approach had strengths and weaknesses. Both flat and hierarchical structures can succeed, Bremner found. It just depends on what kind of problem you’re trying to solve.

For instance, 3DR’s bottom-up approach is useful for finding a path forward when it isn’t clear what is blocking your company’s or product’s success. “An example in the drone industry is the drone architecture,” Bremner explains. “DJI was focused on helicopters, but safety was a huge issue that wasn’t fully recognized at the time.”

For 3DR, the answer to the drone architecture problem arrived serendipitously as members of the open-source community started using quadrotors instead of helicopters or planes. The quadrotor design had existed since the early 1900s, but 3DR implemented it in a way that was more affordable than ever before. That kickstarted the company’s growth, and the quadrotor spread through the industry like wildfire.

Sometimes, though, it’s obvious what problems need to be solved for a new technology to take off in the market. “The challenge might be finding the right solution or just implementing it,” Bremner explains. In cases like these, hierarchy wins.

In the drone industry, the obvious but unsolved problem was figuring out who wanted and needed drones. The quadrotor made drones more accessible, but it was obvious to industry insiders that the technology was still missing a “killer app.” In solving that problem, Bremner noticed that DJI’s approach excelled. “Hierarchy is good because it enables organizations to quickly and systematically test different alternatives,” Bremner says.

DJI experimented with ideas including public safety, utilities inspection, and agriculture, attending trade shows and working closely with experts in each field. Ultimately, the team settled on marketing drones to filmmakers and video producers. “What we saw was, at that point—when the bottleneck to growth in the industry was clear but there was still a lot of uncertainty as to how to solve it—DJI’s approach worked really well,” Bremner explains.

DJI’s hierarchical structure helped the company in other ways as well. “An issue that the team at 3DR ran into was the challenge of building a complex hardware product, like a drone, when you have a distributed team of part-time engineers working for you,” Bremner explains. “They were all brilliant—3DR had a ton of engineering talent—but when employees have free reign over what to do and how to do it, it becomes a struggle to rally them to work through really complex issues.”

Inspired by events in the drone case study, Bremner has turned his attention to experimentation in the video game industry. “One thing that was common in the drone industry, and certainly elsewhere, was rushing to get a prototype on the market to see if it would stick,” Bremner explains. “For DJI, this really worked. But sometimes it really backfired.”

Traditionally, video game developers have relied solely on internal user testing to determine whether a new game was ready for prime time. In the last decade, however, many companies have turned to a new development model called early access, in which they launch an unfinished game to test it on the market. The model is similar to the lean startup method, which emphasizes “being as lean as possible, saving resources and just getting your product out there in the market,” Bremner explains.

For developers with limited budgets, early access is an appealing option. “It’s easier to launch the game, get market feedback … and then improve it or move on to something else from there,” Bremner says. “The question is, what are the risks for taking this type of approach?”

To answer that question, Bremner has begun a quantitative study, which looks at how many people purchase and play each game each day, as well as how long they spend playing. “There is a pretty clear pattern where, over time, games launched in early access start to engage a higher proportion of users than regular games,” he explained.

The catch is, for certain types of developers, these improvements don’t translate into sales. Small independent developers, Bremner’s observed, might take a reputational hit by releasing early versions of not-quite-ready games. “Gamers might simply stay away because of the negative impression formed by a few early players,” he says.

Though the research is still in progress, Bremner says it suggests one reason why entrepreneurs should be cautious when applying concepts from the lean startup method. “I think it’s easy to get carried away in the hype of something like lean startup or early access, so that we forget that there might be situations where the process might not work as intended,” Bremner explains. And there may well be other scenarios where different methodologies are better suited to startup success: “I’ve made it my mission to figure out what those edge cases are,” he says.