The instinct at capacity is to add people. Often the constraint is a process that a second person will simply queue behind.
Small businesses rarely have clean historical data. You can still build a defensible forecast — you just have to be explicit about which assumptions carry the weight.
Two weeks, roughly thirty interviews, and a map of where work actually flows rather than where the org chart says it should.
When we map a process end to end, the delay is almost never in the doing. It is in the waiting between two people who each think the other has it.
Scaling a channel with negative contribution margin does not fix itself at volume. It gets worse, faster, and with more of your cash committed.