Automating something you should have eliminated makes waste efficient. It also makes it permanent, because a process nobody performs manually is a process nobody questions. Half the workflows I see in small businesses are careful automations of work that stopped being necessary two years ago.
So the order is eliminate, then simplify, then automate. Most people start at three.
The arithmetic almost nobody does
Automation has a setup cost. It is usually larger than it appears, because it includes the learning, the fixing when it breaks, and the time you spend thinking about it.
The calculation is simple. Time saved per occurrence, multiplied by how often it happens, multiplied by how long the process will stay the same. Compare that against setup time.
A five minute task done monthly saves an hour a year. If the automation takes three hours to build, it pays back in three years, assuming nothing changes. Nothing stays the same for three years in a first year business.
A five minute task done daily saves twenty hours a year. Same three hour setup, pays back in seven weeks.
The frequency matters far more than the duration, and people consistently get this backwards. They automate the annoying monthly task and hand key the daily one.
The question is not how much you hate doing it. It is how many times you will do it before the process changes.
What to automate first
In order, for a business of one.
1. Things that must happen and that you forget. Backups. Invoice reminders. Follow up on an unanswered quote. Renewal dates. These are worth automating even at low frequency, because the cost of forgetting is not the time, it is the consequence.
2. Data moving between systems you already pay for. Your booking tool to your calendar. Your payment processor to your accounting software. Your form to your customer list. Every one of these is a retyping you currently perform, and retyping is where errors enter.
3. The writing you do repeatedly. Not full automation. Templates. The quote follow up, the appointment reminder, the review request, the onboarding email. Build them from your four best real examples rather than from scratch.
4. Scheduling. A booking link removes an entire category of back and forth that produces nothing. This is one of the few automations that improves the customer experience rather than merely saving you time.
5. Payment collection. Automatic invoicing on a schedule, automatic reminders on overdue accounts, automatic reconciliation into your books. The gap between finishing work and having money is where cash problems begin, and most of that gap is administrative rather than customer caused.
That is the whole list for year one. If you have those five, you have removed most of the mechanical work available to remove.
What not to automate as a team of one
Equally important, and more often ignored.
- Anything a customer experiences as personal. An automated thank you reads as an automated thank you, and it is worse than nothing because it signals that you were not actually paying attention
- Anything where being wrong is expensive. Automated pricing, automated commitments, automated anything that goes out under your name without review
- Anything that changes frequently. You will spend more time maintaining the automation than performing the task
- Anything you do less than monthly. The arithmetic does not work, and you will have forgotten how the automation functions by the time it runs again
- Judgment. If the task requires deciding rather than executing, automation produces confident wrong answers at speed
Automating "send this reminder three days before the appointment" is correct. It is a rule, it runs constantly, and nobody is offended by a reminder. Automating "check in with clients I have not heard from" is a mistake, because the value of that message is entirely that a person thought about them. The first is a task. The second is a relationship, wearing the costume of a task.
The trap specific to working alone
There is nobody to tell you that you have spent three days building something that saves twenty minutes a week.
Automation is genuinely enjoyable work. It feels productive, it produces a visible artifact, and it postpones the harder work of selling or delivering. For a solo operator with no one reviewing how the time was spent, this is a real risk rather than a theoretical one.
Two guardrails.
Set a time budget before you start. If it exceeds the budget, stop and do the task manually while you think about whether it is worth continuing.
And do not build the second automation until the first has been running for a month. Most of what looks essential in week one turns out to be unnecessary once the process settles.
Start with the smallest useful version
The instinct is to design the complete system. Resist it.
- 01Pick the single task you performed most times last week
- 02Write down exactly what you do, step by step, as it actually happens
- 03Remove any step that exists only out of habit
- 04Automate the most mechanical remaining step, not the whole thing
- 05Run it for two weeks and note what broke or annoyed you
- 06Then extend it, or abandon it if it turned out not to matter
Step three is where the value is. Writing a process down honestly nearly always reveals a step that serves nothing, and removing it is free while automating it is not.
What good looks like after a year
Not a business that runs itself. A business where the same set of small things happen reliably without you deciding to make them happen.
Invoices go out. Reminders arrive. Bookings land in the calendar. Payments reconcile. Backups run. Follow ups happen on schedule rather than when you remember.
None of that is impressive. All of it is the difference between a week where you did the work and a week where you did the work and also chased six administrative loose ends.
The point of automating is not to remove yourself. It is to stop spending your limited attention on the parts that never needed it, so that the attention is available for the parts that do. If you want the broader version of that argument, it is here.