What you don’t do compounds too
Startup pop culture has it totally wrong about launching a product. Most of early stage should have little to do with building. It is the least effective investment of time, unless there is no faster way to get feedback.
Early stage is about finding unorthodox ways to not build something until the data shows its needed. Building off pure “conviction” is nothing but gambling. Wild successes of that kind exist, so do lottery winners.
These little “things not done” compound just like many thoughtful features that make great later-stage products. But instead of magical UX they create a sense of tangibility that resonates with target customers — which shrinks feedback loops.
Of course, show not tell, and no demo beats first-hand experience. But many assumptions can be ruled out by a simple discussion. Workflows proven useless as clickable prototypes. Ideally your MVP is just implementing what customers already know and want to buy from you.
If such approach was more widely adopted, we would hear the word “pivot” way less, and product-market fit wouldn’t take years to find. The only exception is the genuinely hard stuff like bio, self-driving etc which is in fact venture-backed research, not startups.
To the point that “it is often much quicker to just build”: it is never. But it is nearly always easier to start doing than to formulate clearly what and why are you doing. What is “quick” to build could just as well be expressed in a line or two of text and / or a visual.
Another way to think of the bias for not building: in a week’s time, there is no way one can build something of significance and collect enough data to validate or course-correct, even with unlimited resources. Without building though it’s not that hard.
So the minimal possible feedback loop is much shorter if you don’t build. It follows, counterintuitively, that to maximize pace of progress you need to postpone building for as long as possible, which is until all quicker sources of feedback are depleted.
Originally a tweetstorm
