Ileana Stigliani, ³Ô¹ÏºÚÁÏ College Business School, with Stefano Magistretti, Politecnico di Milano, Marina Candi, Reykjavik University, and colleagues. Published in Technovation, 2025.
Problem
Most organisations treat testing as something you do as the last step. That approach feels safe, but it is expensive. By the time you discover a problem, you have already spent the budget, missed the window and disappointed the client. The question is whether testing earlier and more often produces better results, or whether it just adds cost and complexity to an already difficult process.
Science
We analysed data from 246 design thinking projects carried out by consulting firms in North America and Europe. Design thinking is a structured approach to innovation that puts the user at the centre and uses prototypes to test assumptions before committing to a solution. We measured two things: how early teams started experimenting, and how often they experimented throughout the project and then tested whether those practices affected both the quality of the final outcome and how efficiently the project was delivered, on time and within budget. We also tested whether using digital tools such as artificial intelligence (AI), big data analytics and three-dimensional (3D) printing changed the picture.
Outcome
Both early and frequent experimentation improved the quality of innovation outcomes. But only early experimentation improved efficiency. Teams that tested ideas early delivered better results faster. Frequent experimentation alone did not make projects more efficient, and in some cases may have generated too many options. When digital tools were used alongside early experimentation, both the quality and efficiency benefits grew stronger amplifying the advantage that early testing had already created.
Change
Our research gives innovation leaders a clear instruction: start testing before you feel ready. A rough prototype in week one is more valuable than a polished one in month three. It also tells organisations where digital tools add the most value: not as a replacement for experimentation, but as an amplifier of it. The research was published in Technovation in 2025 and directly addresses calls from the design thinking community for quantitative evidence on what really works.