Doing Things That Don’t Scale Was the Strategy All Along
Founders treat manual, unscalable work as a phase to escape. The ones who win often treat it as the moat.
"Do things that don't scale" has been repeated so often it has lost its teeth. Most founders hear it as permission to hack together a messy launch before building the real, automated machine. They treat the manual phase as an embarrassment to escape as fast as possible. That is the part they get wrong.
The unscalable work is where you learn
When you onboard the first hundred users by hand — answering every email, fixing every problem personally, watching people use the thing over a call — you are not just being scrappy. You are doing research no dashboard can give you. You hear the exact words people use, the place they get stuck, the reason they almost gave up. Automate too early and you lose that signal forever.
The founders who skip straight to automation often build a very efficient pipeline for the wrong thing.
Manual service is a feature, not a stopgap
For some companies the "thing that doesn't scale" never goes away, because it is the product. Concierge onboarding, a founder who answers support personally, a human in the loop where competitors have a chatbot — these are not failures to scale. They are the reason customers stay. The moat is precisely the thing your bigger competitors are too efficient to do.
You do not have to automate everything just because you can. Sometimes the manual touch is the differentiator, and removing it to save an hour costs you the loyalty that hour was buying.
Knowing when to let go
The discipline is not "never automate." It is automating only after the manual version has taught you what to build, and keeping by hand the parts that are doing real work on retention. Do the unscalable things long enough to understand them. Automate the parts that are pure overhead. Guard the parts that are secretly your edge.