To me the SDLC starts when the development project starts — when we have some-what fixed requirements or begin development sprints from an established backlog. Depending on context obviously. I have a bit more about it here: https://jlottosen.wordpress.com/2017/05/24/testing-across-the-it-landscape/
Scott mentions how we can add value in establishing the business case and business ideas. These can be tested/challenged. Some business schools teach classes in simulations and experimentation with business ideas. Something along these lines: https://www.coursera.org/specializations/wharton-business-financial-modeling but also in rapid prototyping there is always an implicit scientific model (hypothesis/trial).
In essence: The PO’s how do the decide, how do they choose one item over another. What assumptions and probes do they take when deciding?