How Small Should Prototypes and MVPs Be?
Part of the problem people have with prototypes and MVPs is not appreciating how small you can make something while still learning from it.
If you envision the full product, a first SLC version (Simple, Loveable, Complete) should be an order of magnitude smaller. And prototypes should be several orders of magnitude smaller still. For two reasons:
No one should be able to mistake the prototype for the real thing. That's how you end up with broken products and abusing users as bug finders because stakeholders decided to release the prototype.
You must be able to make the prototype in a matter of hours so that you'll take the learning and won't feel regret about sunk cost or get emotionally attached to the solution.
Now, building the actual product should still happen in small releases, but those small releases carry a lot of overhead, like all the extra work around security, localization, internationalization, accessibility and all that stuff that leads to quips like "90% done, 90% more to go". Plus, integrating a new feature into a larger codebase, even if it's small, requires extra care to make sure it fits neatly into existing features.
With all that in mind, then, a prototype should be ridiculously small. Just large enough that you can front-load at least some of that value risk.
