Start With the Number, Not the Tool

Most automation projects start the same way. Someone sees a tool. Maybe a competitor uses it. Maybe a vendor called. Maybe a friend at another company swears by it.

Then the question becomes “how do we use this?” instead of “what are we trying to fix?”

That’s backwards. Start with the number.

Why the number comes first

A number does three things a tool can’t.

It tells you if the problem is worth solving. If your no-show rate is 4%, a new reminder system probably isn’t your best bet. If it’s 18%, it might be.

It tells you when you’re done. “Better lead follow-up” never ends. “Respond to web leads in under an hour” does.

It tells you if it worked. Without a baseline, every project feels like a success to the person who bought it and a waste to everyone else.

What a good number looks like

A good outcome number is:

Specific. “Average time from web form to first call,” not “customer experience.”
Measurable today. You can pull it from data you already have, even if it takes some digging.
Owned by the business. It’s something the owner or ops lead would brag about if it moved.
Movable in weeks. Not a five-year metric. Something you can see change this quarter.

Examples from service businesses

Here are the kinds of numbers we see moving in Twin Cities service businesses:

HVAC: average web-lead response time, or the share of leads never called.
Clinics: no-show rate, or open slots filled from a waitlist.
Accounting firms: hours spent on month-end client reports.
Commercial cleaning: missed or late shifts per month.
Agencies: days from discovery call to proposal sent.

Notice none of these mention software. They’re business results. The tool comes later.

Measure the baseline before you touch anything

This is the step everyone wants to skip. Don’t.

Pull two or three months of data. Match web forms to call logs. Count the no-shows. Time the month-end reports. It’s usually messier than you expect, and that mess teaches you something.

In our illustrative HVAC story, the baseline showed average response time was about 6 hours. But the bigger finding was that 30% of leads were never called at all. Nobody knew that until someone counted. That one fact changed what got built.

Then, and only then, weigh the fixes

Once you have the number and the baseline, list every possible fix. Not just the tool you saw. All of them. Sometimes the best fix is a process change. Sometimes it’s a small automation. Sometimes it’s the big tool after all.

Weigh each one on two things:

Impact. How much would it move the number?
Effort. What will it really take, including getting people to use it?

Our free Impact vs. Effort plotter walks you through this in about a minute. It splits effort into build and adoption, because adoption is the part most teams miss.

A simple test for your next idea

Before your next software purchase, try finishing this sentence:

“We want to move ______ from ______ to ______.”

If you can fill in all three blanks, you’re ready to pick a fix. If you can’t, you’re shopping for a tool, not solving a problem.

For a lot of Minneapolis small business automation projects, that one sentence is the difference between a fix that sticks and another subscription nobody uses.

Start with the number. The tool will be easier to pick, and much easier to judge.