Your MVP Should Suck
MVP is such a misused term in product development that it’s now almost useless. In many organisations, it describes the initial version of a product that can be launched within a specific budget or a set timescale. But that really misses the point of what an MVP is all about.
An MVP isn’t about stripping back features and functionality—it’s about maximizing opportunities for learning.
If you’ve got an idea of what the fully-fledged version of your product looks like and you’re trying to wind that back to the minimum set of features you can go to market with, that’s not an MVP—it’s your version one, and you need to be honest with yourself about that.
A true MVP is the minimum viable product you can create to test your assumptions about the value you’re creating for your customers.
Your MVP doesn’t even have to be a working app (Dropbox’s was a video demo of a product that didn’t exist yet), but it needs to be something you can get into your customers’ hands and get real feedback on straight away.
A true MVP is an experiment, not a launchable product.
From Robert Schlaff's excellent article on MVPs:
In reality, though, this approach to MVPs has become the exception rather than the norm.
Working for a product development agency, I've read a ton of tender-style documents that talk about MVPs in relation to features rather than learning, and that misses out all the value of what an MVP ought to be about.
Like Jason Godesky says:
That's not to say figuring out what the v1 of your product should look like and how to get to your ultimate version is inherently wrong, but it's a lot of wasted effort if you're not going to test your assumptions about it before you build.
To paraphrase Mike Tyson, everyone thinks they know what their market wants until they get punched in the face by their customers.
The only way you can really know if a product idea is worth pursuing is when you get it into the hands of the people you think are going to use it. Ninety percent of the time, the idea that you thought was unstoppable will land badly, and you'll need to iterate on it to figure out what the real solution to your customer's problem ought to be.
Putting effort into building a launchable version of a product before finding out if anyone cares about it is hugely wasteful. Even if you keep the feature set super-tight, there's no point in building things that people don't want.
A true MVP can go along way to helping you mitigate that risk.
To go back to the experimental nature of MVPs, this other quote from Jason Godesky sums it up nicely:
An MVP is your first step towards building the product your customers want, not the one you want to get to market. The difference feels subtle, but it's absolutely massive.
Like Ha Phan says:
