What to Automate First When You Finally Have Time
A simple way to rank what to automate in a small business: how often the task happens, times what it costs you when it goes wrong.
The first thing most owners automate is the thing that annoys them most. That's usually the quarterly thing. The spreadsheet you rebuild every three months. The year-end inventory count. The vendor form nobody can find.
Automating it feels fantastic. It also saves you about an hour a year.
There's a better way to pick. Rank every repetitive task by two numbers you already know in your gut: how many times a week it happens, and what it costs you when somebody gets it wrong. Multiply. Start at the top. It isn't elegant, but it puts your attention where the money actually leaks.
The two questions that do the sorting
How often does this happen? Not how often it feels like it happens — how many times this week. A salon confirms maybe sixty appointments a week. A restaurant takes a few hundred orders. You renegotiate your insurance once a year.
What does one mistake cost? A missed appointment reminder costs you an empty chair and the stylist's wage. A wrong order costs you the food, the remake, and sometimes the customer. A late supplier payment costs you a phone call and some embarrassment.
High frequency plus high cost is where you start. Low frequency is where you stop, no matter how satisfying the fix would be.
The quarterly task has another problem beyond the math. You do it so rarely that you'll forget how the automation works before you need it again. You'll end up debugging your own system at the worst possible moment. Tasks you do daily keep their own automation honest — if it breaks, you find out within hours.
First: confirmations and reminders
These sit at the top for almost every local business, because they happen constantly and failure is expensive and invisible until it's too late.
For appointment businesses — salons, clinics, repair shops — that's the reminder before the booking. A no-show doesn't just cost you the service. It costs you the slot you turned someone else away from. You can't fill a 2pm chair at 1:55. We've written about why no-shows are a reminder problem rather than a customer problem, and the short version is that most people who miss appointments meant to come.
For restaurants and shops, it's the order confirmation. Somebody orders online. Do they get a message back, instantly, with what they ordered and when it'll be ready? If not, they call to check. Now one order has cost you a phone call during service, and the person answering is reading the ticket aloud instead of running food.
Build this first because it runs on its own schedule forever, it touches every single customer, and when it works, your phone gets quieter.
What "built" looks like in practice: the booking or order system sends a text at a set interval — 24 hours out for appointments, immediately for orders, with a second message when the order is ready. Text, not email. People read texts within minutes and email whenever. There's more on picking the channel if you're deciding between the two.
Second: the review request
Reviews are the thing every owner knows they should ask for and nobody remembers to do at the moment it would work. The moment is right after the service, while the person is still happy and still holding their phone.
Automate it as a message that goes out a few hours after an appointment or a day after an order, with a direct link to your Google review page. Not your website. Not a form. The actual review box, one tap.
This ranks second rather than first because a missed review request costs you less, immediately, than a missed reminder. Nobody notices a review that didn't get written. But it happens just as often as the reminder does, and the cumulative effect on how you rank in local search is real. It compounds while you sleep.
One rule: the automated ask is the backup, not the replacement. The in-person ask still works better, and we've covered how to make it without feeling pushy. Automate the follow-up for everyone you didn't catch in person.
Third: the handoffs between your systems
This is the one owners skip because it doesn't feel like automation. It feels like plumbing.
Every place where a human retypes information from one screen into another is a daily, high-cost failure point. Online orders retyped into the POS. Bookings copied from an email into a paper schedule. Delivery app tickets entered by hand during a rush. Each retype is a chance to drop a modifier, fat-finger a phone number, or lose a ticket entirely.
These are expensive mistakes that happen dozens of times a day, which puts them near the top of the ranking — they only land third because the fix usually requires connecting two pieces of software, and that's a bigger job than turning on text reminders. Stop retyping online orders into your POS walks through what that connection actually involves.
If you only have one afternoon, do the reminders. If you have a budget and a few weeks, do the plumbing — it's the one that lets you add volume without adding staff.
Fourth: reporting
Now automate the numbers. Not before.
Reporting ranks low because a late report costs you very little in the moment. You can always pull the number tomorrow. But it does happen weekly, so it earns a place on the list — just not the first place.
The automation here is simple: have your systems email you the same handful of numbers at the same time every week, so you stop deciding which report to open. Three numbers every Monday morning is the version we use. Sales, labour as a percentage of sales, and the cost of what you bought. Everything else can wait for your accountant.
The mistake is building a dashboard first. Dashboards are enormously satisfying to build and almost never change a decision. A three-line email you actually read beats a beautiful dashboard you check twice.
What to leave alone
Anything you do less than monthly. Year-end. Licence renewals. The annual menu reprint. Put these on a calendar reminder and do them by hand. Automation has a maintenance cost, and an annual task will never pay it back.
Anything that's really a judgment call. Responding to a bad review, deciding whether to comp a table, approving a shift swap. You can automate the alert that tells you it happened. Don't automate the decision.
Anything you haven't done manually enough times to describe out loud. If you can't write down the steps, you can't automate them, and whoever you hire to build it will guess. Run it by hand for a month first.
The order, in one list
- Confirmations and reminders, by text, automatic, every customer
- Review requests, a few hours after the visit, direct link
- Data handoffs between your ordering system, your POS, and your calendar
- A weekly numbers email with three figures in it
- Everything else, by hand, on a calendar reminder
Work down that list and stop when you run out of time. You'll have covered the tasks that happen hundreds of times a month before you touch the ones that happen twice a year.
Most of this is wiring that already exists inside tools you're paying for — booking software, your ordering system, your phone. If you'd rather not spend the evening figuring out which settings do what, our services page lays out what we set up and what we hand back to you to run yourself. Either way, the ranking holds: frequency times cost, highest first.