What to automate first: a one week time audit

Most automation projects start in the wrong place: with a tool someone heard about. The better starting point is a notebook and one ordinary week. Before you spend a dollar on software, find out where your time actually goes.

Week one: write it down

For five working days, everyone on the team keeps a simple log of anything they do more than once. Not the real work of the business, but the work around it:

• Answering the same questions by phone or email
• Copying details from one place to another
• Confirming, reminding and following up
• Chasing payments, documents or approvals
• Pulling numbers together for a report

For each one, note roughly how long it takes and how many times it came up. Rough is fine. You are looking for patterns, not precision.

Score what you found

For every repeated task, multiply how often it happens by how long it takes. That gives you the hours it costs each week. Then ask two more questions: what happens when it goes wrong, and does the customer feel it?

Here is what a week might look like for a small service business. The numbers are an example, not a benchmark:

Task	Per week	Minutes each	Hours a week
Booking confirmations and reminders	40	3	2.0
Copying new customers into invoicing	15	6	1.5
Following up on quotes	10	5	0.8
Monday numbers for the owner	1	45	0.8

In that example the reminders win on time, and missed reminders also cost real money in no shows. That is the first candidate.

What makes a good first automation

The best first projects share four traits:

1. It happens often. Daily or weekly, not twice a year.
2. It follows a rule. You could explain it to a new hire in two sentences.
3. It is already digital. The information lives in a calendar, a form or an inbox, not on a sticky note.
4. A mistake is easy to catch. If it goes wrong, someone notices quickly and the damage is small.

Leave for later anything that needs judgment, happens rarely, involves upset customers, or depends on messy information. Those can be helped by automation eventually, but they make poor first projects.

Write the rule in one sentence

Before anything gets built, write the task as a rule: "When this happens, do that, unless this." For example: "When a booking is made, send a confirmation, then a reminder the day before, unless it has been cancelled."

If you cannot write the rule in one sentence, the task is not ready to automate. It needs a decision first.

Then decide what a person still checks. Good automation handles the routine and flags the exceptions: a failed payment, an odd request, a message that bounced.

Count the full cost

Setup is only part of it. Include the monthly software subscriptions, the time to look after it, and what happens when a connected service changes. Compare that with the hours you measured in week one. A task that costs two hours a week is worth roughly a hundred hours a year.

Measure, then do the next one

Run the first automation for a month, then repeat the week one log for that task. Did the hours drop? Did the mistakes stop? Did customers notice? Only then pick the next task on your list.

For ideas once the first one is running, see nine automations small businesses should steal from big companies.