THINK: assumptions and hypothesis
Instead of expressing “requirements, ” we consider assumptions and declare hypothesis statements. A hypothesis statement is a way of expressing assumptions in a testable form.
Reality often shows that the creators of a product or service are not sure what a user needs or wants or how it should work concerning the user experience. But quite often, organizations or companies proceed with their idea for functionality or user experience without input what the user needs (or what the user thinks about it) and without checking (testing) with the end-user throughout the development process.
Quite frankly, organizations that have been in business for many years (decades) think (with some level of arrogance) to know what the customer wants and how the experience should be – but this arrogance will bounce back with failure: an unhappy end-user will trash the product and its experience if it deemed to be unfit.
Hypothesis statement
A hypothesis statement will require experiments to determine which assumptions are right or wrong.
The first step is to list any assumptions: an assumption is a declaration of what we assume to be true. But it’s not considered a requirement until we have facts and figures to support it. Assumptions comprehend both business and user assumptions. A possible way to start listing assumptions is to start from a problem statement. Problem statements are full of assumptions. When assumptions are listed, it’s crucial to select the most uncertain ones.
The main principle of Lean Startup is to expose the highest risk first to learn. The whole team should do the exercise of listing assumptions and hypotheses with the whole team: everyone involved so that you can create a shared understanding from the start.
A hypothesis statement has the following format:

Note: This is a possible format, and there’s no obligation to follow this blindly, as long you understand the purpose, and you have the different elements.
Outcomes (objectives) and metrics
It is essential to focus on outcomes first: what are we trying to achieve and what objectives? What kind of user needs? Instead of directly thinking about features (solutions), we focus on outcomes. When creating the product, the progress will be measured against the outcomes. To measure, you need to define a set of metrics. An excellent approach to this is Objective-Key-Results.
Features (functionalities, stories, …)
Features exist to serve the needs that have been identified. You create a particular feature (solution, functionality, story). It’s essential to have the outcome first defined and think about possible features.
User assumptions
It is essential to know something about the customer, the end-user you’re creating the product for. Personas are a well-known technique to make the end-user visible.
The clue is to approach the creation and validation of personas the lean way: personas are not the output of great research upfront but are created and evolve as the product evolves. Personas created with lots of work upfront and not updated afterwards are essentially throw-away personas: they become useless as soon as they are outdated, and concern a source of waste.
We prefer to create “ad-hoc” or “proto”-personas: these represent the best guess of who is using the product and why. A proto-persona contains the information (demographics, behaviours and needs) of a specific type or segment of end-user relevant to the product being designed.
Resources
Lean UX book
Materials in this post are based upon readings on the Lean UX approach.

Lean UX book by Jeff Gothelf.
Read more about the Lean UX cycle.
Read more about the Lean UX book.