Korva is a scroll of the internet’s best ideas.
Two failure modes
Product work has two failure modes. One is iterating forward and throwing spaghetti at the wall with no sense of direction. The other is writing long roadmaps and strategy documents for the next two or three years, which gives up software's advantage over every other industry: learning. A/B tests let a team update its understanding and change course as it goes.
The accordion
An accordion makes no music while it is being stretched open. The sound comes when a key is pressed and the bellows push back in, and then the player pulls it out again for the next progression. Each pull out is the question of what to do given what was just learned, and each push in is building the next thing.
Live commerce example
Live commerce shows the motion. A seller can hold up AirPods, start them at a dollar, and buyers have all the information they need, with no listing, which takes about 3 minutes to make. Zoom out, and new buyers will expect search to work, which it cannot if nobody knows what is being sold until after it sells. Requiring listings fixes that, but zoom out again: at 3 minutes per listing, sellers sell far fewer things per hour, so that fix is out too.
Why can software teams learn faster than others?
Software can show two versions of a product to random groups of users and measure what each group does within days. A car or a building cannot be rebuilt for a sample of customers that cheaply. Ron Kohavi, who ran experimentation teams at Amazon and Microsoft, describes in his book Trustworthy Online Controlled Experiments how large tech companies run thousands of such tests each year. The method is usually called an A/B test.
How does an accordion actually make sound?
Inside an accordion sit small metal tongues called free reeds, each mounted over a slot. When air is pushed past a reed, it vibrates and produces a note. The bellows move air in both directions, and pressing a key lifts a pad that opens the path to the matching reed. With no key pressed, no air flows through a reed and the instrument stays silent while it is stretched open.
Where did the smallest-possible-test idea come from?
The term minimum viable product was coined by Frank Robinson around 2001. Eric Ries, author of The Lean Startup in 2011, made it widely known. The idea is to build only enough to test one assumption with real users, then use what comes back to decide the next piece.
More from Lenny's Podcast: Product | Growth | Career
