New research from ESMT Berlin has found that people judge the potential of an idea differently depending on how developed it is. In the early stages, they tend to focus on the reputation of the person behind the idea and how well it’s presented, even though these factors don’t predict success. But as the idea progresses, its popularity and how fast it’s gaining popularity become more important.

The study aimed to understand how open innovation platforms affect idea generation and whether the factors people consider for success change as ideas evolve. It also looked into whether a machine learning model could predict which ideas would succeed. To do this, the researchers used data from the LEGO IDEAS platform, where users propose new LEGO sets, and the crowd votes to pick a winner.

Changing perceptions

The study revealed two key findings. First, people’s criteria for judging ideas change a lot as the idea matures. Early on, it’s all about the idea creator’s reputation and presentation, but as time goes on, it’s more about how popular the idea is and how quickly it’s growing.

The second finding is about the power of machine learning. Can computers do a better job than people at picking the winners? The research showed that, interestingly, in the early stages of an idea, machine learning can provide better predictions than in the later stages. Early indicators are crucial in figuring out which ideas will get rejected by the crowd.

“Our research furnishes innovation managers with a roadmap,” the researchers conclude. “By understanding these critical characteristics of successful ideas, managers can sieve out underwhelming concepts right at the inception. The long-term implications? More informed, strategic decisions in innovation management.”