The Two Inconvenient Truths: Most Ideas Fail and Iteration Can’t Wait

Most people believe that if they plan carefully or gather enough opinions, their next big idea is sure to be a hit. But decades of innovation research and the lived experience of top companies show the opposite: at least half of all new product features, shifts, or solutions flop. Not because people are careless, but because it’s almost impossible to predict true value, usability, or technical feasibility in advance. Even among the best teams, surprise failures are common.

Once that’s understood, the smartest teams design their whole approach around rapid, low-cost experiments. Instead of waiting until after months of building to find out if users care, they invest a few days creating a quick, non-production prototype or test—just enough to gauge reactions. Instead of getting attached to one path, they adapt and pivot until real evidence appears. It can feel humbling, or even counter-intuitive, since it means letting go of the perfect plan.

Picture a team rolling out a new feature, only to get lukewarm feedback or see usage stall. Rather than stick with sunk costs, great teams reconstruct, retest, and measure success by how quickly they learn what works—not how many features they check off. The technical term for this is operating in a 'continuous discovery and delivery' mode, where learning drives development, and progress means improving results through cycles of evidence, not just output. It's a hard shift to make, but those who do build products and careers that last.

Start today by letting go of the comforting idea that you or your team can plan your way to guaranteed success. Name, right now, one thing you’re working on that could be riskier than it seems, and frame it as an experiment. Instead of waiting for perfection, put together a basic prototype, draft, or test—something raw but functional—that you can share with a real user, customer, or supporter in the next few days. Watch and listen for what surprises you, and note which assumptions break down. Use that feedback as your blueprint for what to tweak, and don’t hesitate to try again with a fresh angle. If you make iteration and visible learning the norm, not the exception, you’ll be much closer to finding real breakthroughs—fast.