Your MVP isn’t supposed to be a product 一 it’s a process
An MVP isn’t just a product you work on for three months and then launch ー it’s an iterative process you need to repeat until you’ve found product market fit.
“You know that old saw about a plane flying from California to Hawaii being off course 99% of the time — but constantly correcting? The same is true of successful startups — except they may start out heading toward Alaska.” — Ev Williams
You need to identify your riskiest assumption, find the smallest possible experiment to test that assumption, and use the results of the experiment to course correct. You can identify and validate these assumptions by putting your product in the hands of real people as quickly as possible.
Just Keep Iterating
Don’t be discouraged if you have to go back to square one, because most startups do ー multiple times.
In fact, most startups that get accepted to the prestigious accelerator, Y Combinator, leave the program working on a different project than they applied with.
The founders who fail fast win. Figure out which of your assumptions are false by getting real user feedback as quickly as possible.
As long as you learn from your failures, you will get better and better. And if you’re in the idea stage, you can keep iterating your idea until you have something people actually want.
“The fastest way to succeed is to double your failure rate.” ー Thomas J. Watson Sr., Founder of IBM
The only way you can fail is by giving up.
Start with a RAT
The term ‘minimum viable product’ was popularized by Eric Ries’ The Lean Startup. Instead of spending years developing a software solution, Ries advised that development should only take a few weeks, or a month ー tops.
However, as lean methodologies are evolving, entrepreneurs are realizing that you can test your assumptions much faster by conducting riskiest assumptions tests, or RATs.
Consider IBM’s speech-to-text experiment. They hypothesized that humans would be able to interact faster with computers by using speech instead of a mouse and keyboard. Instead of spending years developing the technology, they designed an experiment in a few days.
They set up a room with a microphone, a computer monitor — and no keyboard. They told the test subjects that they had a speech-to-text machine for them to try, and all they had to do is speak into the microphone, and their words would “magically” appear on the monitor.
They pulled it off by hiding a fast typist (with a keyboard) in another room. The microphone output was fed to a speaker, and the hidden typist translated the speech into keystrokes which appeared as text on the monitor with amazing speed and accuracy.
From this experiment, IBM found that loud working environments would make speech-to-text unappealing and lack of privacy would be an issue for many potential users.
Needless to say, they didn’t continue with the idea. But had they choose to develop the technology 一 it would have failed.
Read more: “How to Use Smoke Tests to Validate Your Product or Feature Ideas”
Minimum Remarkable Product
When you’ve tested and iterated on your riskiest assumptions, you can start building the first version of your product. Build it as quickly as you can, and launch. The goal is to get real world feedback on your idea as soon as possible.
But don’t move so fast and cut so many corners, or you may end up missing the core value of your product. Your product doesn’t have to be scalable or capable of living on its own. It doesn’t even have to be well designed or polished.
All it needs to do is capture your core value.
“I think Jeff Bezos’s “minimum remarkable product” terminology is a better phrasing than minimum viable product. A viable product brings to mind this whole ecosystem of support like a viable organism, which has all of these things it needs to live. You don’t actually want that.
If you’re launching something that is self-sufficient and can go off and live on its own, you definitely waited too long.”
Once you’ve gotten to the point where one person in the world thinks your product is remarkable, you’ll have something worth pursuing. And if nobody finds remarkable value in your product, you need to figure out why, iterate your product, and try again.
Skateboard Your Way to Product Market Fit
After you’ve created that “skateboard,” or v1, it’s time to start working up to a “car.”
However, don’t hide and spend a year developing your product. Keep testing risky assumptions with smoke tests, clickable prototypes, and concierge MVPs. Launch updates as fast as possible, measure their impact on your bottom line metrics, and react accordingly based on user feedback and data. Otherwise, you may need to repeat the entire process over again.
Steve Huffman, founder of Reddit, explains what the company did after launch:
“In the next six months, we created a ton of features. Only about 25% of them ended up lasting more than a day or two. There was stuff we thought was a great idea that didn’t work. If we had waited to include them in our launch, we would have waited until November and then wondered why they didn’t work.”
Further reading
If you’re interested in learning how to apply the practices mentioned in this story, the GV team has written a book called Sprint that explains how to go from idea to test in just one week. It’s a must-read for any entrepreneur.
