Productized operations vs hiring another VA.
Most “VA-shaped problems” are actually workflow-shaped. Hiring a person to repeat a task you haven’t defined doesn’t fix the task — it just hides it.
The test
Before hiring, ask three questions about the work:
- Can the input be defined precisely?
- Can the desired output be defined precisely?
- Are the rules in between mostly stable, or are they actually judgement?
If the answer to the first two is yes and the third is “mostly stable,” you don’t have a hiring problem. You have an undocumented workflow.
What VAs are actually good at
VAs are excellent at exception handling — the cases where rules don’t cleanly apply. Where they fail (and where founders quietly burn money) is on rule-bound, repetitive tasks where “the assistant has to remember” what the standard is. Memory is what workflows are for.
The hybrid model that works
Run the rule-bound work as workflows, and use a VA — full-time or fractional — as the human surface on top. They handle exceptions, edge cases, and the small percentage of tasks the system flags. You stop paying a person to copy-paste, and you start paying them to think.
The cost question
A capable VA at 20 hours a week costs more than the operations setup that replaces 80% of their original task list. Run the math on the actual repetitive work — order exports, ticket triage, refund decisions, reporting — and the comparison usually isn’t close.
Hire people for judgement. Build workflows for everything else.