An MVP is not a cheaper version of the product. It is the smallest complete answer to the one question the business cannot answer any other way.
That framing changes what gets cut. Scope reductions that weaken the answer are expensive, even when they look like savings on a plan.
Start from the riskiest assumption
Every product rests on a belief that has not been tested: that people will change a habit, that a workflow can be automated safely, that a market will pay. Build the part that tests it.
Everything else — settings, admin surfaces, edge-case flows — can be manual, partial, or absent in the first release.
Keep quality where the user notices
Speed, clarity, and reliability in the core loop are not polish. They are the experiment. A slow, confusing version of the core flow produces a result you cannot trust.
Reduce breadth aggressively. Reduce depth in the core loop carefully.