Five questions to ask before you automate anything

The task costing you the most is usually the one you have stopped noticing. It is not the annoying one that gets your attention, it is the quiet 30-second job you do 40 times a day without thinking about it. So before you spend money automating anything, run it through these five questions. If a task cannot answer most of them well, leave it alone.

1. How often does it happen?

Frequency is what matters most, more than how long the task takes. A 2-minute job done 50 times a week is worth way more of your attention than a 2-hour job done once a month. When you are hunting for what to automate, count the reps first. The small daily thing almost always beats the big occasional one.

2. What does it cost when it goes wrong?

Some tasks are cheap to mess up and some are expensive. A mistyped internal note costs nothing. A missed filing, a lead nobody answered, a double-booked job, those cost real money or real trust. The more a slip costs you, the more a system that gets it right every time is worth, even for a task that only takes a minute.

3. Is it stuck in one person’s head?

If a task only works because one specific person knows how to do it, that is not really a task, it is a risk. The day they are out sick or they leave, the work stops. Automating that kind of thing does two jobs at once: it saves the time, and it gets rid of the single point of failure.

4. Is it the same every time?

This is the one people skip, and it is the one that sinks bad automation projects. A task is only ready if it runs the same way most of the time. If every case is a judgment call with different inputs and exceptions, a system will fight you. Sometimes the real first step is not automating the process, it is making the process consistent enough to automate in the first place.

5. Should it even exist?

A lot of the recurring work in a business is there just because it always has been. Before you automate a task, make sure it should be happening at all. Every now and then the best automation is deleting the step. Automating something pointless just means you do the pointless thing faster.

Where that leaves you

Run your recurring work through those five and a short list falls out: frequent, expensive to get wrong, stuck in one person’s head, consistent, and actually worth doing. Scoring your own tasks honestly is the hard part, because the ones costing you the most are the ones you have gone blind to. That is most of what we do on a first call. You bring the list, we ask the hard questions, and you leave knowing exactly which one to fix first.