Making sense of minimum viable products
johnnyholland.org
johnnyholland.org
If you were trying to launch a Twitter competitor now, you'd be hard-pressed to get away with the same degree of minimalness.
However, if you had a unique take on the approach, say, all Tweets get posted backwards and upside down, then you can probably get away with a page that does that, as minimally as possible, to test and see if people actually like that. If they do, then you start fleshing it out. If everybody hates it, move on to the next idea, or pivot it into something better.
The point is that the MVP allows you to qualify whether or not people will use it. If you're solving somebody's pain, they probably don't give a flying hoot if it has fancy mouseovers, or a WebGL rotating background, or even good typography. Do those things help? Sure. But if the idea at its core is rotten, then all those things are is a waste of time. They are things that slow you down from figuring out whether or not the core of your idea has validity.
If the idea is proven, by somebody else entering the market first, then the bar is raised on what gets counted as 'minimum', unless you're approaching the problem in a uniquely different way. Google Drive is, as near as I can tell, a feature-for-feature knockoff of Dropbox. That said, it couldn't be crappy. The Google Docs integration isn't enough of a differentiator to allow for the bad initial first impression of a half-assed client. If it worked in some fundamentally different way though, perhaps it could have.
The point though, is exactly what you said... MVPs are contextual, and the context is the market. How unique is the idea. Is it radical thinking? Does the model need proven? If nobody's done anything like it before, why not? Is it a great idea that you're the first to stumble onto, or is it a bad idea that everybody else has realized stunk, and haven't done it for that reason. Get your product out there to find out, as quickly as possible.