A stopwatch before a quote
Replacing Manual Business Processes With Software: Time the Task Before You Touch It
How long does it take you to send an invoice? You think you know. You do not, and neither does anyone selling you software. Two weeks with a sticky note will tell you which chores are worth replacing, which are only annoying, and which should simply stop.
November 15, 2026 · 9 min read
The short answer
Before replacing a manual process with software, time it for two weeks: count how often it happens, how many minutes each time and what errors it causes. Then try deleting it, batching it, using a template or renting a tool. Only then price a build. A chore that eats 20 minutes a workday is about 87 hours a year; one that takes 10 minutes a week is rarely worth any software at all.
On this page
The chore you hate is rarely the chore that costs you
Owners pick what to automate the way people pick what to complain about. The loudest chore wins. The monthly report you resent takes 40 minutes, 12 times a year. The two-minute text confirming tomorrow's appointment, sent nine times a day, gets no complaints at all, and by our arithmetic it comes to about 78 hours a year if you do it five days a week. One of those is a burden and the other is an annoyance. Your feelings have them backward.
Sellers are glad to work from your feelings. An estimate of "hours saved" built on your guess is a number with a decimal point and no evidence behind it. The cure is cheap and slightly embarrassing: a sticky note on the phone, a tally and two weeks of mild self-awareness. You will also get a by-product more valuable than the number, since most people who time a process notice at least one step that nobody needs.
There is no special skill needed. You do not need a consultant, a process map or the phrase "digital transformation," which has its own page and a raised eyebrow. You need a pencil. If you want a quick first estimate on phone and follow-up time, the automation savings calculator works in about a minute. Your own timings will always beat it, because they are yours.
A two-week time log that survives a busy week
The usual advice is to log your whole day, and nobody with a truck full of parts will do that past Wednesday. Log only three to five chores you suspect. Keep the note where you do the chore. Round to the nearest minute and do not agonize. A rough number from real days beats a precise number from memory by a wide margin, and two weeks is long enough to include a bad one.
- 1
Name the chores in plain words
"Copying leads into the job software." "Chasing unpaid invoices." "Sending the same photos to the insurance adjuster." If it takes a paragraph, you have named two chores.
- 2
Tally each time, with minutes
One line, one chore: 6, 4, 11. Include the hunting for the file and the switching between apps. People leave those out, and those are often the biggest part.
- 3
Mark who does it
You, a helper, the bookkeeper you pay by the hour. A minute of yours and a minute of theirs are worth different amounts, and that decides the sum later.
- 4
Note the damage
Each time something goes wrong because of the chore, write it down. A missed invoice. A quote sent to the wrong person. A lead who called someone else. Errors tend to outweigh the time.
- 5
Mark the thinking
Put a star by any instance where you had to use judgment. A chore with many stars is not a chore. It is your job, and software will not be doing it.

What those minutes are worth, and what they are not
Now turn tallies into a yearly figure. The table uses the Bureau of Labor Statistics' May 2025 median of $24.36 an hour for bookkeeping clerks as a neutral yardstick. All of the arithmetic is ours. The last column compares each row with $575, the floor of our web app and integration price, which depends on complexity, as of November 2026. It also assumes the tool removes every minute and costs nothing to run. Neither is likely, so treat that column as the most flattering case possible.
| Hours a year | At $24.36 an hour | Time to reach $575, best case | |
|---|---|---|---|
| 10 minutes a week | About 9 | About $211 | About 2.7 years |
| 30 minutes a week | 26 | About $633 | About 11 months |
| 20 minutes each workday | About 87 | About $2,111 | About 3.3 months |
| An hour each workday | 260 | About $6,334 | About 5 weeks |
Before you get excited, here is the awkward question. Who actually gets that money? If you pay a bookkeeper by the hour and those hours go away, it is a real saving, on the invoice, next month. If the chore is yours, nobody writes a check. A saved hour is only worth money if you sell it to a customer, or spend it on something you would have paid for anyway. Sleep counts, though not on the spreadsheet. Be honest about which case you are in.
If you employ someone for this, wages understate the cost. The BLS employer cost release for June 2026 splits private-industry compensation 70% wages and 30% benefits, a figure covering all private workers rather than small firms. At that split, a $50,670 bookkeeping wage implies about $72,400 in total employer cost, by our own estimate. That number is for a full-time hire, which is not what most companies with fewer than 10 people have. It is still a useful sanity check on a vendor who claims to replace a whole person.
Six things to try before anyone writes code
Software should be nearly the last thing you try. Work down this list in order and stop as soon as the problem shrinks to something you can live with. The first three cost nothing and often handle most of the time on your log. Automating a chore you could have deleted is the most expensive way there is to do nothing.
- 1.Delete it. Ask who reads the report. Ask what happens if the confirmation text does not go out for a week. Sometimes the honest answer is nothing, and you have just saved 100% of the time at no charge.
- 2.Batch it. Twenty interruptions become one 15-minute block at 4 p.m. Invoices sent Friday instead of whenever. The switching can cost more of your day than the task itself.
- 3.Template it. Saved replies, a standard estimate, a checklist. Your phone can expand two letters into a whole paragraph today. Nobody needs to be paid for this.
- 4.Hand it off. A few paid hours a week from someone else can beat any tool, particularly for work with judgment in it. Software does not call a customer back with sympathy.
- 5.Rent a tool. See the tasks a very small business can automate and automating without hiring. Zapier's Free plan offers 100 tasks a month as of November 2026, enough to test an idea.
- 6.Then, and only then, build. A small connection or tool, for the slice that the first five steps could not remove.
Automating a chore you could have deleted is the most expensive way to do nothing.
Sorting your log: hand over, assist or keep
With your log in hand, sort every remaining chore into one of three piles. The dividing line is how much a mistake costs and how much judgment it takes. The stars you drew in the log will do most of the sorting. Almost nothing belongs in the first pile for the reasons a sales deck gives you. The word "AI" does not move a chore from one pile to another by itself.
| Looks like | Examples | What to do | |
|---|---|---|---|
| Hand over | Same steps every time, data already typed once, a mistake is cheap to fix | Lead into CRM, paid invoice into the books, appointment reminder | A connection, live the same day |
| Assist | Same structure, needs a human look, a mistake reaches a customer | Estimates and invoices, review replies, follow-ups | Machine drafts, person approves |
| Keep | Judgment each time, a mistake costs a customer | Complaints, pricing a strange job, firing a customer | Do it yourself and enjoy it |
The middle pile deserves more respect than it gets. Removing most of the typing while keeping a person's eye on the outcome is often the best deal available to a very small company. Our AI automation page covers where a model can help, and the back-office screen is where this tends to live once it grows. Also, watch out for one thing in this pile: a review step that gets rubber-stamped after a month. If you stop reading the drafts, you have moved the chore to the first pile without deciding to.
Measure again a month after the switch
Nearly nobody does this, which is why so much software outlives its reason for existing. A month after a new tool goes live, use the same sticky note for one week on the same chore. Software tends to swap one chore for another: you no longer retype leads, but now you check that the sync worked. That is often still a good trade. You should know whether it is.
Signs it worked
- The minutes on the note really fell
- The error type you logged has stopped happening
- You stopped checking it every morning
- Your helper uses it without being told
Signs it did not
- You now do the old way and the new way
- You have a new spreadsheet to track the tool
- Time fell, but errors moved somewhere less visible
- Nobody can say what it costs each month
When your log says a build is worthwhile, the rest is small. What moves a web app's cost explains the quote, the workflow automation service handles jobs that need no screen, and AI automations cover the drafting pile. If the chore is answering the same chat and booking questions, Kairos, our AI widget, is from $575 once to build, then $99, $199 or $349 a month by conversations. The section for very small businesses maps the rest. We will not promise savings, because they depend on minutes we have not seen. Read us your log at (210) 346-0848 and we will tell you which line we would touch, including none.
Questions owner-operators ask
Two weeks is enough for most small companies, as long as it includes a normal week and a messy one. Chores tied to monthly events, such as billing runs, need a full cycle. You are not after precision. You want to know whether this is a 10-minute-a-week problem or an hour-a-day problem, and two weeks of tallies will make that obvious.
Choose those that happen often, follow the same steps each time, and copy information someone already typed once. Moving leads into a CRM and marking invoices paid are typical. Ignore how much you hate a chore and look at the log. Leave anything involving judgment or an upset customer for last, or keep it permanently in human hands.
Multiply the minutes per occurrence by occurrences per year, divide by 60, then multiply by the hourly cost of whoever does it. For your own time, use what you could actually bill, and only if you could really sell those hours. Add the cost of errors separately, and subtract the tool's monthly cost. Be skeptical of the result if it depends on perfect adoption.
Yes. It can spread a mistake faster, hide a failure until someone complains, or lock in steps you should have removed. It also adds a thing to look after. That is why you simplify the process first, keep a person reviewing anything customers see, set up an alert on failures and write down how to do the work by hand.
Sometimes it postpones a hire, but we will not promise that. Software can absorb repetitive copying. It will not return an upset customer's call or decide which job comes first. If your log shows most of the time going to judgment and conversation, you probably need a person, even part time. A tool may then make that person more useful.
Keep reading
- The list, with the catchesTen Tasks a Very Small Business Can Automate, With the Cost and What Goes WrongTen business tasks a microbusiness can automate, each with the cost as of November 2026, what setup takes, what goes wrong, and which to leave alone.
- Small, and yoursCustom Software for Microbusinesses: Small Tools That Retire the SpreadsheetDashboards, portals and integrations from $575, for owners who run the business from a phone and a spreadsheet, and the rented tools to try first.
- One owner, one phoneWeb Design for One-Person Businesses: What You Need and What to RefuseWebsite design for one-person businesses: the short page list, what to refuse to pay for, and fixed prices from $1,500, or when a builder plan is better.
- The build sideSmall Business Website Development: Who Holds the Keys When the Builder LeavesWebsite development for independent contractors: who owns the code, domain and accounts, what custom means at $1,500, and what you must maintain yourself.
- Probably not. Sometimes yes.iOS App Development for Small Businesses: When an iPhone App Earns Its KeepWhen an iPhone app makes sense for a very small company, what from $3,750 buys, Apple's fees and rules as of 2026, and what to try before you build.
- Two stores, one budgetMobile App Development for Small Business: iPhone, Android or a Web App FirstiPhone and Android apps for a small business: web app from $575, native from $3,750, Google Play and Apple fees as of 2026, and what to build first.
- Everything in MicrobusinessSee the section
From the Insights desk
- “We'll Fix It Later” Is the Most Expensive Sentence in Web DesignIt is said with the best intentions around week three of every website project. But later is not a time — it is a place where work goes to accumulate interest, and it always arrives on a Saturday.
- What a Fast Website Is Actually Worth — in DollarsSpeed is a sales number wearing an engineer's costume. Every second your site makes a visitor wait, a share of them leave — customers you already paid to attract. How a fast website turns into dollars, with the math for your own traffic.
Read us your two-week log
Send the chores, the minutes and what went wrong. A person replies with which ones to delete, which to hand to a rented tool, and what any small build would cost.
Which of these sounds like you? (pick any)
Rather talk now? Call (210) 346-0848.