The good thing about failure is that it allows us to learn, so I suppose it’s a good thing that I’ve failed quite a bit, especially at my last startup. One of the things I eventually learned is that startups are to be treated as a science experiment. But first, let me describe a tremendously dangerous mentality that all startup founders have.
First, you come up with an idea which you think is pretty good. You talk to your friends and family, and everyone seems to think its a good idea. Thus, you jump into development of this idea, wholly believing in the premise that this is a good idea…after all, you wouldn’t be developing it if it wasn’t. But your product takes a while to develop, months go by and you don’t want to release a shoddy product, so you work on perfecting it – bugs could kill you, scalability could kill you, bad UI could kill you…
The problem is, your idea sucks. Or at least it’s untested and the market hasn’t had time to tell you whether or not it’s a great idea. At my last startup, we fell into this trap, and it took 2 years before the first beta version was released to the world. Think about that – developing in a vacuum for 2 YEARS before there was any market feedback. We released a stable, mature, scalable product which promptly received very little usage. I brought hundreds of thousands of visitors to our site through PR, advertising, SEO and word of mouth, yet active user growth stagnated, and churn was nearly 100%. We felt pretty ridiculous.
That’s when I realized that we had developed a technology with no awareness for our market, and by the time we realized our mistake and launched into full scale market research, it was nearly too late.
I’ve said it before, but Peter Drucker has a great quote about how business has only two functions: marketing and innovation. And one without the other fails.
As such, the most important component of any early stage startup is to: 1) determine if your technology can be built and 2) determine if people want your product. The end result is what most people call product/market fit.
Startups as a Science Experiment
Every science experiment has 5 components:
You will use each of these steps to test your market and technology.
Research: What products are currently available. How are people using them? What problems are they having? What needs are unmet? You will establish a broad understanding of the market and technology options available.
Problem: What specific problem will you attempt to solve?
Hypothesis: Here is when you establish the who/what/when/where/why/how of your startup. It must be testable and more importantly, measurable. At my last company we should’ve said something like “lawyers who research case law need annotation tools to generate reports for their clients (and are willing to pay for this possibility)”.
Experiment: If you have an enterprise startup, this is when you hit the pavement and start talking to your end users before you’ve written a line of code. Determine if your particular technology implementation is valuable to them. Do they have this problem? Do they KNOW they have this problem? Have they actively tried to find a solution? If you can answer yes to all three of those questions, you’ve hit the jackpot. If you have a consumer play, the easiest thing to do is build what you consider to be a minimum viable product (buggy as all hell and literally only a way to capture data), then release it. Follow the creed, “release early, release often”. This is not a time to be a perfectionist, this is a time to collect data, so if you find yourself building a product to do anything other than prove the technology works, STOP.
Results: Compile all of your data, look it over, and see if you need to collect anything else. What does your data tell you? Did you prove or disprove your hypothesis? If you’ve disproved it, hypothesize and test again.
Moral of story, stop building for perfection and start testing your market.