Validated learning is an approach to demonstrating progress against business goals when traditional key performance indicators (KPIs) are not very useful. In his book, "Lean Startup," author Eric Ries described validated learning as a small unit of progress that can be quickly verified to determine whether a chosen direction is correct.
Validated learning treats product development as a series of experiments that use scientific method to answer questions about market demand. Once the entrepreneur has created a hypothesis about what customers want, the next step is to test the hypothesis by building a prototype -- which in lean-speak is referred to as a minimum viable product (MVP). Potential customers are then asked for feedback about the prototype and the information is used to validate reality and fine-tune the product. The process is purposely iterative and is continually repeated throughout the product's lifecycle.
Validated learning is especially good at helping startups avoid building features that customers don't want or need. The thinking is that by continually validating what matters most to customers, the startup will be more likely to eventually demonstrate progress against traditional KPIs, including revenue.
Although validated learning got its start as part of the build-measure-learn (BML) approach to efficient product development in lean manufacturing, its simple tenents can be applied to the development of a wide variety of products and services, including software development.