$13.8Bn milk; is 'scratch your own itch' actually bad advice?
An investor once challenged me:
"'Scratch your own itch' sounds like great advice but, in reality, you’ll just build something that only you want. It won’t be VC-scale."
I wanted to argue back, but I realised I only had one essay from Paul Graham to point to and, frankly, some tired old anecdotes from the likes of AirBnB and Uber.
I didn't have any data.
So, over the last week, I painstakingly dug out the origin stories of the 100 highest-valued US unicorns crowned since 2012, and profiled how the founders of each came across the problem they ended up solving.
It took me a little minute. 🥵
Result: 67%
Of the 100 highest-valued US unicorns crowned since 2012 were started by founders scratching their own itches (i.e. solving their own problems).
A clear, two-thirds majority
Interestingly, many of these problems were:
- Encountered in the course of trying to build other things
- Ignored by lots of other people at the time
- Not obviously ‘VC-scale’ issues
Here’s one of my favourite examples from the set:
Alexandr Wang wanted to know when his fridge needed restocking without opening the door.
He started working on an AI product that could recognise when he was running low on milk, but struggled because of the lack of available training data.
Seeing that other AI developers were suffering from the same lack of data for other applications, he co-founded Scale AI with Lucy Guo.
Scale AI provides labelled training data for AI, and was last valued at $13.8 billion.
Of course, scratching our own itches doesn't excuse the need for market validation... This is about the invisible benefits conferred by solving our own problems as a starting point.
So far, the advice holds up.
Limitations
Granted, it's not the most fool-proof study (yes, survivorship bias is a thing). But I'm excited to go deeper on this in the future.
Wishing you lots of problems,
Alex