A startup working on a product that does not yet make money looks, in traditional financial terms, like a disaster. Revenue is zero. Costs are real. The spreadsheet is terrible. But the startup might be making enormous progress: testing assumptions, identifying what customers actually want, building toward a model that will work. Traditional accounting cannot see this. It only measures what has already happened in financial terms. Ries proposes innovation accounting as an alternative: a three-step process for measuring whether a startup is making genuine progress. First, establish a baseline by measuring current performance with an MVP. Second, tune the engine by making changes and measuring whether the key metrics improve. Third, make the pivot-or-persevere decision based on the evidence. This makes progress visible and measurable even before revenue exists. A startup that can show improving activation rates, better retention, or higher conversion over successive builds is demonstrating real progress. A startup that can only show increasing total users without these supporting metrics is hiding behind vanity numbers. The investors and leaders who understand this difference make far better decisions about where to allocate resources.