Our missed-call math, with the arithmetic shown

Search "how much does a missed call cost a plumbing business" and you won't have to scroll far for a number. Usually a specific one — $500, $800, sometimes four figures — stated flat, no working shown. I went looking for where those numbers come from once, for a different post, and mostly what I found was silence where a methodology should be. Nobody publishes the three numbers they multiplied. They just publish the product.

So here's mine. Not because it's more rigorous — it isn't, it's the same three-number multiplication everyone else is doing. The difference is I'm going to show you all three numbers, where they came from, and everywhere the model is wrong on purpose.

The formula

It's genuinely this simple, and that's sort of the point — there's no hidden model, no proprietary weighting, nothing you couldn't do on the back of an invoice:

missed calls per week → times your close rate → times your average job value → times 4.33 (average weeks in a month) → that's your modeled monthly loss.

Four numbers, three multiplications. You could build this in a spreadsheet in ninety seconds, and you basically already have — it's the same arithmetic you'd use to answer "what's a slow week cost me," just aimed at the calls that never got answered instead of the ones that did.

I built it into a page you can actually touch — the missed-call calculator — but the tool isn't the finding. The formula is. Handing you the output alone would be worth less than nothing, because the output changes depending on numbers only you know.

Running it for a one-truck plumbing shop

I need starting numbers to walk this, so I'm using the defaults my plumbing page ships with — not because they're true of any real shop, but because they're the ones sitting behind the calculator right now and you can go check every step against the live page yourself.

8 missed calls a week. That's the assumption baked into the default slider — not a count of anything, a starting point you're meant to drag until it matches your own phone.

35% close rate. Of the calls somebody actually answers, a little over a third turn into booked jobs. Also a default, also draggable.

$450 average job value. A blend across drain cleaning, water heater swaps, the occasional bigger repair.

Now the arithmetic, one step at a time:

8 missed calls × 35% close rate = 2.8 jobs a week that would have closed, if every one of those calls had reached someone.

2.8 jobs × $450 a job = $1,260 a week in modeled lost revenue.

$1,260 × 4.33 (the average number of weeks in a month — more on that number in a second) = $5,456 a month.

Subtract my own $199 flat price, since that's what the calculator does next: $5,257 net, on paper, is what the model says a shop this size gives up by not answering versus paying for something that does.

I loaded the actual calculator before writing that number down. Sliders untouched, defaults as shipped: it reads $5,456 and $5,257, same as the arithmetic above. If you land on a different number, drag a slider — that's what they're there for.

Two numbers on my own site, for the identical shop

Here's the test of everything I just said, and it happened on its own before I could set it up on purpose: a post already live on this site runs this same shape of math for the same kind of shop — 8 missed calls a week, $450 average job value — and lands on "about $5,800 a month." This post says $5,456. Two different published numbers, same company, same scenario. You're exactly the reader who'd go check, so you'd have found that gap before I got around to mentioning it. I'd rather get to it first.

It isn't a mystery once both posts are open side by side. That post used the same 8 calls and the same $450, and put the close rate at "somewhere in the 30-40% range" — then, instead of carrying the fraction through, it rounded straight to a whole job: "of those 8 missed calls, maybe 3 would've become jobs." This post keeps the fraction: 8 × 35% = 2.8 jobs, not 3.

Run that one rounding choice both ways and you get the entire gap. 2.8 jobs × $450 × 4.33 = $5,456 — this post's number. 3 jobs × $450 = $1,350 a week × 4.33 = $5,845.50, which that post then rounds down again to "about $5,800." Every other input is identical. The whole ~$390 difference is one decision: whether 2.8 gets carried as 2.8 or bumped up to a round job before the rest of the multiplication runs.

That older post called its own math "deliberately rough," and it was right to — nobody has 2.8 jobs, they have jobs, and the fraction is a modeling convenience either way. But look at what the rounding did on the way out the other end: nudging the jobs figure from 2.8 to 3 is about a 7% change to that one number, and it moved the final monthly answer by roughly $390 — from a step that doesn't even read like a decision while you're making it. Nobody sat down and chose to add $390 to the estimate. Someone rounded 2.8 up to a number that reads better in a sentence, and the multiplication carried that rounding all the way to the bottom line.

That's not a defect in either post specifically. That's the finding. A rounding choice buried in the middle of a calculation doesn't announce itself in the output — you just get $5,800 or $5,456, and neither number, read on its own, tells you a decision got made in between. The only way to catch it is to have both chains open at once and walk them step by step, which a published total with no published steps never lets you do.

I'm not going to tell you which of $5,456 and $5,800 is the "real" one, because neither is. They're the same rough model, run with one different rounding choice, and your actual number depends on your actual close rate — not on either one of mine.

The 4.33 is the tell

Every other number in that chain, you supplied, or I supplied as a starting guess for you to correct. The 4.33 is different — nobody chose it, nobody measured it, it's a constant. There are 52 weeks in a year and 12 months, so 52 ÷ 12 = 4.33 weeks per month, on average, because months aren't all the same length and neither are the weeks that fall inside them.

I'm calling that out specifically because it's the kind of thing that's easy to slide past as if it were data. It isn't. It's an averaging convention doing a small, honest job: turning a weekly number into a monthly one without pretending every month has exactly four weeks in it (none of them do) or exactly 4.33 (none of them do either — some have four, some have five, 4.33 is just where they land on average). That's the whole function of that number. It's not measuring your business. It's calendar math.

What this formula deliberately does not count

This is the part that actually matters more than the dollar figure, so don't skim it.

It assumes every missed call was a job you'd have won at your normal close rate. Real call volume includes wrong numbers, suppliers, telemarketers, and people who were never going to book regardless of who answered. The model doesn't know the difference between those and a burst pipe. It treats all 8 the same.

It has no idea some of those callers try again. A missed call in this formula is gone, permanently, the moment it's counted. In reality some fraction call back an hour later, or the next morning, and reach you fine. The formula can't see that — it counts the miss once and moves on.

It uses the same close rate for missed calls as for answered ones. That's almost certainly wrong, and wrong in the flattering direction for this number. A caller who reaches a live answer immediately is probably a warmer lead, on average, than one you're extrapolating backward from a call you never actually had. The formula can't tell — it just applies the one close-rate number you gave it to both groups equally.

It flattens every job to one average dollar figure. Real weeks mix an $89 faucet fix with a $3,000 repipe. Averaging that into $450 is useful for a rough estimate and useless for any specific missed call, which was either the $89 one or the $3,000 one, never the average of the two.

It has no ceiling. However many jobs the formula says you lost, it assumes you had the trucks and hours to do every single one of them. If you're already booked solid, some of those "lost" jobs were never getting done regardless of who picked up the phone.

It doesn't know about seasons. A plumbing shop's call volume in a hard freeze looks nothing like a slow week in October. Eight missed calls a week is a flat assumption across a year that isn't flat.

Put plainly: this is a model, not a measurement. It's not a measurement of anyone's business, including mine — I don't have customers running through it yet, so there's no real-world number to check it against, only the arithmetic.

Where the "1 in 4" comes from

My plumbing page states one outside figure you didn't supply yourself: that plumbers miss around 1 in 4 calls. That number isn't mine. It traces to Invoca's Home Services Lead Conversion Benchmarks Report for 2026, which reported roughly a 74% answer rate for plumbing calls specifically — the unanswered remainder lands close to 1 in 4.

Two things worth knowing about that figure before you lean on it. First, Invoca sells call-tracking software into this exact market, so it's vendor data about the vendor's own customer base, not an independent academic study — the best evidence I could find, not a disinterested one. Second, it's an industry-wide figure, not a measurement of your shop specifically, the same way the $450 average job value above isn't a measurement of your job mix. You can and should replace 8 missed calls a week with whatever your own week actually looks like.

The point isn't that my number is better

It isn't, particularly. It's a formula same as any other "cost of a missed call" figure floating around this category — four inputs and three multiplications, same shape as the ones with no working shown.

The difference is you can see all four inputs, check the multiplication yourself, and argue with any one of them. A number that comes with its assumptions attached is a number you can push back on. A number that shows up alone, with a dollar sign and no ingredients, isn't wrong because it's dishonest — it's unfalsifiable because there's nothing in it to disagree with.

Go run your own week through the calculator. Change the calls, the close rate, the job value — none of the three are mine to know, they're yours. Whatever number comes out the other side, you'll have watched exactly how it got there, which is the one thing most of what you'll read on this topic won't give you.