What one lost customer a week costs a local business

Most owners never see the customer they lose. The phone just does not ring, and there is no way to know who called someone else instead.

This post gives you a way to put a dollar figure on that customer. You bring your own numbers, and we will show you the math.

What's the formula?

Your average job value times 52 is what one lost customer a week costs you in a year.

Average job value is what a typical customer pays you for one job. To find it, take last year's revenue and divide by the number of jobs you did. If you do not track jobs, a rough guess from a stack of recent invoices is fine.

The full version looks like this: lost customers per week × average job value × 52 = revenue lost per year.

Why multiply by 52?

There are 52 weeks in a year. One lost customer a week adds up to 52 lost jobs by December.

If your jobs are big and infrequent, use months instead. Lost customers per month × average job value × 12 works the same way.

What if my jobs vary a lot in size?

Split your work into two or three types and run the formula for each. A plumber might run it once for service calls and once for water heater replacements.

This matters because the big jobs are usually the ones customers research before they call. Losing one large job a month can cost more than losing a small call every week.

Two worked examples

These are examples with made-up numbers. Swap in your own.

Example 1, a pest control company. Say a typical first treatment brings in $150. One lost customer a week is $150 × 52 = $7,800 a year.

Pest control customers often sign up for a recurring plan, though. If that customer would have stayed on a plan, the real loss is several times bigger, and the simple formula undersells it.

Example 2, a roofing company. Say an average replacement job is $9,000. A roofer probably does not lose one of those every week, so use the monthly version: one lost job a month is $9,000 × 12 = $108,000 a year.

Run your own numbers before you read further. Most owners are surprised by what comes out.

What does the formula leave out?

It leaves out everything that comes after the first job. A lost customer also takes the repeat work, the review they would have left, and the neighbor they would have sent your way.

If customers come back to you, multiply by how many times a typical customer hires you. Either way, treat your number as the floor.

Where are those customers going?

Some of them are asking an AI tool who to call. When they do, the AI usually answers with names.

In our Florida HVAC AI Visibility Study, ChatGPT named a specific HVAC company in all 149 of its searches. Across the five AI tools we tested, 96.9% of the companies named showed up on only one tool.

So if a customer asks ChatGPT and you are not on its list, they get someone else's name. They never see yours, and you never know they asked.

The same pattern showed up in every Florida study we ran. In our roofing study, 97.2% of the roofers AI named showed up on only one of five tools. For pest control it was 96.7%, and for dentists it was 94.7%.

How do I know if I'm losing customers to AI?

You cannot see it in your call log. You can see it by asking the AI tools yourself.

Type the question your customer would type, like “best roofer in [your city],” into ChatGPT, Gemini, and Perplexity. Write down who comes up, and check whether you do.

Or run our free scan, which checks five AI tools for your business in about 60 seconds.

Is it worth fixing?

Compare your yearly number to what it would cost to close the gap. Plenty of the fixes are free and take an afternoon: cleaning up your Google Business Profile, asking for reviews that mention the job, and making sure your website says what you do and where.

If your number is big and the gap is real, paid help can make sense. If your number is small, do the free fixes first.

Pair the number with what you find when you check AI. If AI names you on most tools, your lost customers are probably leaving for other reasons, like slow callbacks or price. If AI names you on none, visibility is a likely leak.

Either way, start by finding out whether the gap exists. The free scan tells you that.