The hockey stickPhase 3 · Wiring

Automate It, Hire For It, or Leave It Alone

Not every annoying task is worth automating. Some are better handed to a person, and some are best left exactly as they are. Here is a one-page test before you wire anything up.

4 min read

The two-question gut check

For most tasks, two questions settle it:

  1. 1
    Do I do this the same way, over and over, every week?
  2. 2
    If the computer got it wrong, would it cost me real money or a customer?

A clear yes to the first and a manageable no to the second is your green light. When the answers conflict, run the full score below.

The five-factor score

Score each factor 1 to 3. Add them up.

Factor123
Frequencyrarelymonthlydaily or more
Time per runa few minutestens of minutesan hour or more
Error costcosmetic, easy to undosome reworkmoney, compliance, lost customer
Rule-clarityneeds judgmentmostly rulesclear rules, few exceptions
Stabilitychanges weeklychanges a few times a yearunchanged in a year
  • 12 to 15: automate. High-frequency, stable, rule-based work is the textbook case.
  • 8 to 11: maybe. Run the break-even math below first. Often the answer is a partial automation with a human checking the output.
  • 5 to 7: leave it alone, or hire a person.

The break-even rule

Automate only if the setup time, plus ongoing upkeep, is clearly less than the hours it will save in a year, and ideally it pays back within twelve months.

A few honest anchors:

  • Value the time you save at about 1.25 to 1.4 times the wage, to cover benefits and overhead. Then subtract maintenance and any subscription, or you will overstate the win.
  • A task done weekly that saves five minutes can justify a lot of one-time setup. A task done once a year almost never does, no matter how annoying it is.
  • The first third of the tasks you automate tend to deliver most of the time savings. Start with the highest-frequency, most stable one and stop chasing the long tail.

The "leave it alone" list

Do not automate:

  • Rare or one-off tasks. The build costs more than it saves.
  • Processes still changing. If it shifts every few weeks, you will spend all your time editing the automation. Wait until it settles.
  • High-judgment work. Interpretation, strategy, and creative calls belong to people.
  • Relationship and high-stakes moments. Hard customer issues, negotiations, and complaints need a human.

One nuance Auto-Phil stresses: do not automate is not do it yourself forever. A task that is steady and high-volume but needs judgment is often the case to hire a person, not to wire it up. Automation fits steady and rule-based. A hire fits steady and judgment-heavy. Leave-it-alone fits rare or fast-changing.

From Auto-Phil

Auto-Phil helps owners decide which annoying tasks are worth automating, which belong to a person, and which are best left alone. The company runs a one-page test before anything gets wired up, so effort goes where it actually pays back.

When you want a hand

Skip the guesswork on your own setup.

Thirty minutes, no pitch. Tell us the work you do and we will tell you the next move that actually fits your shop.