The best first automation is a task that happens often, takes real time, hurts when it goes wrong and follows clear rules. Most owners know that instinctively. The trouble is that, in a busy week, the task that gets automated first is usually the one someone saw in a demo, or the one that annoyed the owner most recently. Neither is a reliable guide.
This post gives you a simple scoring method instead. In about an hour, you can list your repetitive tasks, score each one on four factors and pick your first three automations with a clear reason for each. It works on paper or in a spreadsheet, and it doesn't depend on which tools you use.
Key Takeaways
- Score each repetitive task from 1 to 5 on frequency, time per occurrence, cost of errors and rule clarity, then multiply the four scores.
- Multiplying (rather than adding) means a task that scores very low on any one factor drops down the list, which is what you want.
- A low rule-clarity score means the task needs fixing or standardising before it can be automated.
- Some tasks shouldn't be automated at all; they should be dropped.
- Pick three, define what success looks like for each, and measure before you build.
Why the first automation matters most
Your first automation does more than save time. It sets expectations. If it works and the team notices the difference, the next one gets an easy welcome. If it's fiddly, breaks in the second week or automates something nobody cared about, the team concludes that automation is more trouble than it's worth, and the next idea meets resistance.
So the first choice should be a confident one: a task with an obvious payoff, a clear set of rules and an owner who wants it fixed. That's what the scoring method is designed to find.
It also matters because automation has a cost that's easy to forget. Building is the visible part. Looking after it (updating it when prices, forms or staff change, fixing it when a connected tool changes) is the ongoing part. The well-known xkcd comic "Is It Worth the Time?" makes the arithmetic memorable: a chart of how long you can spend making a routine task more efficient before you've spent more time than you'll save over five years. A task you do once a month rarely justifies much effort. A task you do fifty times a day almost always does.
If you're new to the wider topic, our guide to business process automation covers the basics, including why you simplify a process before automating it.
List every repetitive task (the 30-minute exercise)
Before scoring anything, you need a list. Set aside 30 minutes, ideally with two or three people from different parts of the business, and write down every task that happens more than once a month and follows roughly the same steps each time.
Prompts that help people remember:
- What do you copy from one place to another?
- What do you send that's almost the same every time (emails, messages, quotes, reminders)?
- What do you check or chase on a schedule?
- What report or spreadsheet do you rebuild each week or month?
- What do customers ask for that you handle the same way every time?
- What happens when a deal is won, a job is finished or a new staff member starts?
Aim for 15 to 30 tasks. Don't filter yet. "Copy new web enquiries into the spreadsheet" and "send the weekly sales numbers to the owner" belong on the list just as much as anything that sounds technical.
Score each task on four factors
Give each task a score from 1 to 5 on four factors. Use the guide below so different people score consistently.
| Score | Frequency | Time per occurrence | Cost of errors | Rule clarity |
|---|---|---|---|---|
| 1 | Monthly or less | Under 2 minutes | An error is a minor nuisance | Every case is different; heavy judgement |
| 2 | A few times a month | 2 to 5 minutes | Internal rework | Mostly judgement, some patterns |
| 3 | Weekly | 5 to 15 minutes | A customer notices | Clear rules with frequent exceptions |
| 4 | Daily | 15 to 30 minutes | Lost money or a lost customer | Clear rules with occasional exceptions |
| 5 | Many times a day | Over 30 minutes | Legal, financial or reputational damage | Fully rule-based; anyone could follow it |
Then multiply the four scores. The result runs from 1 to 625.
Why multiply rather than add? Because a task that scores 1 on any factor is a poor first candidate, however well it scores elsewhere. A task done many times a day, taking over half an hour each time, with no clear rules (5 x 5 x 4 x 1 = 100) shouldn't beat an equally frequent ten-minute task that's fully rule-based (5 x 3 x 4 x 5 = 300). Adding would make them look much closer than they are. Multiplying rewards tasks that are solid across the board.
A few scoring tips:
- Score the task as it is today, not as it would be if it were tidier.
- Use "cost of errors" for what actually happens when it's wrong, including the customer's experience, not just the time to fix it.
- Be honest about rule clarity. If the person doing the task often says "it depends", score it 2 or 3.
Spot tasks that should be fixed or dropped, not automated
The scores will also tell you which tasks aren't ready for automation. Look for two patterns before you pick anything.
Low rule clarity (1 or 2) with high everything else. These are important, time-consuming tasks that nobody does the same way twice. Automating them now would mean encoding one person's habits and frustrating everyone else. Fix them first: agree the rules, standardise the inputs (one form instead of free-text messages, one price list instead of three), then score them again. Many of these become excellent automation candidates a month later.
Low value across the board, or a task nobody can explain. Some tasks exist because of a problem that went away years ago: a report nobody reads, an approval that's always granted, a spreadsheet kept "just in case". The cheapest automation is deleting the task. Before you automate anything, ask of every item on the list, "what would happen if we stopped doing this?" If the honest answer is "nothing much", stop doing it.
There's a third pattern worth noting: tasks where the human touch is the value, such as a personal call to a long-standing client or a conversation about a complaint. They may score well on frequency and time, but automating them risks the relationship. Automate the reminders around them, not the conversations themselves.
Pick your first three and define success
Sort the remaining tasks by score and look at the top five or six. Choose three, using the scores as your guide and two practical checks as tie-breakers:
- Is there an owner who wants this fixed? An automation without an internal champion tends to be ignored or quietly worked around.
- Can it be built with the tools you already have? A first automation that needs a new system and a long integration project isn't a quick win.
For each of the three, write a short success definition before anything gets built:
- The measure: what will change (hours per week, response time, error count).
- The baseline: where it stands today. Measure it for a week or two if you don't know.
- The target: what good looks like after a month.
- The owner: who checks it and who to tell when something looks wrong.
This takes ten minutes per automation and makes it obvious, a month later, whether the effort paid off. It also gives you a story to tell the team, which matters more than it sounds.
An illustrative scored example
Here's how the method might play out for an illustrative 25-person interior design and fit-out company in Bangkok. The tasks and scores are invented for the example, but they're typical of what a service business finds.
| Task | Frequency | Time | Error cost | Rule clarity | Score |
|---|---|---|---|---|---|
| Copy web and chat enquiries into the leads spreadsheet | 5 | 2 | 4 | 5 | 200 |
| Send appointment confirmations and reminders for site visits | 4 | 2 | 4 | 5 | 160 |
| Chase overdue invoices | 3 | 3 | 4 | 4 | 144 |
| Build the weekly sales and project report | 3 | 5 | 3 | 3 | 135 |
| Prepare custom design quotes | 4 | 5 | 4 | 1 | 80 |
| Send the monthly newsletter | 1 | 5 | 2 | 4 | 40 |
| Fill in the internal "job won" form nobody reads | 3 | 2 | 1 | 5 | 30 |
What the company might conclude:
- First three: enquiry capture (200), appointment reminders (160) and invoice chasing (144). All three are frequent, rule-based and visible to customers when they go wrong. Enquiry capture also feeds the other two, so it goes first.
- Fix before automating: custom quotes score high on frequency, time and error cost, but rule clarity is a 1 because each designer prices differently. The right move is to agree a standard pricing structure and quote template first, then revisit.
- Close, but not yet: the weekly report is a strong candidate for the next round, once enquiry data lands in one place and the numbers can be pulled automatically.
- Drop: the "job won" form scores lowest and, when asked, nobody could say who reads it. Replacing it with an automatic notification from the CRM, or simply stopping, saves time without building anything.
- Leave as is: the monthly newsletter is infrequent enough that automating its production isn't worth the effort right now.
Notice how the method does more than rank tasks. It surfaces a process problem (inconsistent quoting) that, left alone, would have kept the owner involved in every quote. That's often a hidden growth ceiling: work that can't be delegated or automated because nobody has written down how it should be done.
From scores to a shortlist
The point of scoring isn't precision. Two people might give a task a 3 or a 4, and that's fine. The point is to replace "whatever annoyed me most this week" with a quick, shared and repeatable way of deciding. Run it once now, act on your top three, and run it again in three months, when the list will look different because the first automations have changed how work flows.
If you'd like a second opinion on your scored shortlist, send it to us and one of our automation consultants will review it with you, including which tasks to automate, which to fix first and which to drop. If you'd rather hand over the build as well, our Automations, Integrations & Optimisations service covers that too.
If your top tasks turn out to be enquiry follow-ups, appointment reminders or replies to missed calls, Scalify, our all-in-one platform, covers those with built-in workflow automation, booking and missed-call text back. It's one option among several, and the scoring method works whichever tools you use.