What should a small business automate first?

Start with the task your team does the same way every week, that involves no judgement, and that somebody currently does by hand. That is almost never marketing, and it is almost never your website.

The instinct is to automate the thing that feels most modern. The task that actually pays is the one nobody talks about, because everyone has quietly accepted that it takes an hour every Monday.

#How do you know a task is worth automating?

Three questions. A task needs all three, not two.

1. Does it repeat on a schedule? Weekly, daily, every time an order comes in. If it happens twice a year, leave it alone.
2. Does it follow the same steps every time? If the answer starts with "well, it depends", you are looking at judgement, not process.
3. Is somebody doing it by hand right now? Automating something nobody does yet is a product decision, not an efficiency one.

The arithmetic is simple enough to do on paper:

Thirty-five hours is most of a working week, spent on copy and paste. That is the number that makes the case, and you can measure it this afternoon without buying anything.

#Which tasks should you leave alone?

Anything where being wrong is expensive and being slow is cheap.

• Pricing decisions. The rules change faster than you will maintain them.
• Anything customer-facing that needs an apology. An automated apology reads as an insult.
• Hiring. Every shortcut here is a way of not reading the application.
• The task that only breaks once a quarter. You will spend more time maintaining the automation than the task ever cost.

If a person has to check the output every time, you have not automated the task. You have added a step to it.

#What does the first automation actually look like?

Smaller than you expect. A good first build usually has one trigger, one transformation and one destination.

Take the order-to-spreadsheet example. Something happens, you reshape it, you put it somewhere:

{
  "event": "order.created",
  "order_id": "EJU-4417",
  "customer": "Adeyemi Ltd",
  "total": 148000,
  "currency": "NGN"
}

That payload arrives, a few fields get renamed, a row appears in the sheet. No dashboard, no AI, no new software for your team to learn. The person who used to copy those numbers finds them already there on Monday morning.

#Build it for the failure, not the happy path

The happy path takes an afternoon. What separates an automation that survives from one quietly abandoned in March is what happens when the payload arrives malformed, duplicated, or at three in the morning.

Ask before you build: when this breaks, who finds out, and how? If the answer is "the customer", build the alert first.

#How long before it pays for itself?

Take the hours you measured, multiply by what that hour actually costs, and compare it against the build. Most first automations of this size are worth it inside a quarter. If yours needs a year of arithmetic to look sensible, you have picked a task that is too small, or a tool that is too big.

The second one is faster. So is the third. The expensive part is never the building, it is working out which forty minutes to take back.