Building an Industry Before the Market Exists

What does a new technology need before it can become a real market?

China’s civilian drone industry did not emerge from one famous technology hub. Companies appeared across many cities, often before customers, operating rules, or even the industry itself were clearly defined. We wanted to understand what made entry possible, and why the support that mattered at the beginning became less useful as the market developed.

What I did

• Working with Daniel Armanios, I studied China’s civilian drone industry from early experimentation through market takeoff.

• We mapped 2,130 drone companies across 183 cities, classified their positions in the value chain, and analyzed their entry patterns within a panel covering 280 Chinese cities. We connected these patterns with local universities, testing sites, commercialization programs, and regulatory changes, then complemented the data with 30+ interviews at two major drone expos in Shenzhen.

• Unlike an adoption study focused on one product or organization, this project reconstructed the formation of an entire industry across locations, value-chain positions, and stages of market development.

• We followed that evolution through two major turning points: a 2013 licensing system that legitimized civilian drone activity and a 2017 registration rule that introduced tighter oversight.

What we found

• First, prove that the technology works. Early companies depended heavily on aeronautical universities for engineering talent and specialized knowledge. Xi’an, for example, became a center for flight-control and navigation companies because its universities already held the expertise needed to solve those technical problems.

• Then, make experimentation legitimate. The 2013 licensing system established clearer rules and signaled that civilian drones had a legitimate commercial future. Entry increased, especially among companies integrating systems and developing products. As the industry gained recognition, proximity to technical universities became less decisive.

• Next, prove that someone will use it. Once drones could be built, firms needed credible applications. Model farms gave companies places to test agricultural drones under real operating conditions. Science parks connected firms with potential customers, market knowledge, and commercialization support. The bottleneck had moved from engineering feasibility to product-market fit.

• Finally, govern the market at scale. By 2017, the central question was no longer whether drones could become a market. It was whether a rapidly expanding market could be governed safely. The real-name registration rule raised entry barriers across the industry. Earlier advantages in technical knowledge or local market access could not shield firms from a national shift toward safety and accountability.

We call this process dynamic scaffolding: as one uncertainty is resolved, a different bottleneck becomes binding.

So what

For policymakers and innovation leaders, the question is not simply how much support to provide. The more useful question is: what is preventing the market from moving forward now?

The bottleneck may be technical, commercial, or institutional. An intervention that unlocked one stage can become less useful once the market reaches the next. The same logic applies to AI, autonomous vehicles, robotics, and other emerging technologies. Building the technology is only the beginning. Turning it into a functioning market requires different ecosystems and institutions to step in at the right time.

With Daniel Armanios, University of Oxford. Revise and resubmit at Strategic Management Journal.