run many cheap experiments fast
run many cheap experiments fast. you can't know in advance which one opens the next door.
what to do
Run a lot of small, cheap experiments quickly instead of betting everything on one plan. Each result opens a door you could not see before you ran it. The job early on is to find which door opens.
for example
At a launchpad workshop I ran in 2026, the teams that found traction were the ones shipping many small tests and reading the results honestly. This is the adjacent possible that Steven Johnson describes, borrowing the idea from the biologist Stuart Kauffman: at any moment only certain next steps are reachable, and you find them by trying. One cheap experiment reveals the next.
the failure it prevents
Months spent perfecting a single bet that was never going to work, with no cheap tests run that might have shown you the one that would.
when this doesn’t apply
Once an experiment clearly works, stop running scattershot tests and pour everything into the thing that is pulling.