The number that gave it away
Our audit asks ten questions and returns a monthly figure for what missed calls cost you. Someone with nine or more trucks, an average job over $5,000, and a person answering the phone got this back:
You are leaking about 13 jobs a month. That is somewhere around $62,400 to $104,000 a month, on your own numbers.
Read that again. A shop where a person answers the phone was told it was losing six figures a month. If that were true, the business would not exist.
Here is the arithmetic it ran:
- 400 calls a month, estimated from the fleet size
- times a 15% miss rate, giving 60 missed calls
- times 85% who never call back, giving 51 gone
- times 1 in 4 who would have booked, giving 13 jobs
- times an $8,000 average job
Every step is a plausible-looking multiplication. The output is nonsense.
Where the error actually is
The question that sets the miss rate is this one:
Someone calls at 8pm on a Saturday. What happens?
That is an after-hours, weekend question. It is a good question. The problem is what we did with the answer: we applied it to the entire month's call volume.
So a shop with a person on the phone from eight to six on weekdays, whose line goes to voicemail at 8pm on a Saturday, answers "it goes to voicemail" and gets modelled as missing 55% of every call it receives, all month.
Saturday evening is a few hours out of roughly seven hundred in a month. The model treated it as all of them.
That is not a tuning problem. It is a category error, and no amount of relabelling the output fixes it.
Four of the five numbers were ours, not yours
The copy under that figure used to say the numbers were "on your own numbers," and that two of them were industry figures and the rest were yours.
That was backwards. Here is the actual provenance:
| Number | Where it came from |
|---|---|
| 400 calls a month | Ours. Derived from how many trucks you said you run |
| 15% miss rate | Ours. Derived from the 8pm question, then misapplied |
| 85% never call back | Ours. An industry figure we had not sourced properly |
| 1 in 4 would have booked | Ours. An assumption |
| $8,000 average job | Yours, and even then it is a band, not a figure |
One number out of five came from the person taking the audit. The tool then described the result as their own numbers.
What we changed
Three things, this week:
- A perfect score now returns no loss. The old code picked a "weakest link" by sorting three equal scores, which always returned the first one. Answer every question perfectly and it told you your website was the problem. It now says there is no obvious gap, because that is the honest answer.
- The copy states who each number belongs to. Four of five are ours. It says so.
- The figure is labelled an illustrative estimate, not a measurement of your business.
What we have not fixed yet
The miss-rate scoping. There are three honest options and each has a cost:
- Scope the rate to after-hours volume only, which is what the question actually measures. Most accurate. The headline number drops a long way.
- Ask a second question about weekday daytime handling and blend the two. More accurate still, one more step in the funnel.
- Stop showing a dollar figure until the tool collects real inputs. Most honest, and the weakest hook.
We are deciding between them. Until then the number is overstated and the page says so. We would rather publish that than quietly tune a constant until the output looks reasonable.
How to get a real number in twenty minutes
You do not need a calculator. You need four numbers you already own.
Total inbound calls. Your phone carrier's portal has this. Log in and look at the last full month. Most business lines show answered and unanswered separately.
Missed calls, split by time. The same report. Pull out the ones inside your working hours. Those are the expensive ones, because a caller at 2pm on a Tuesday expects a human and a caller at 9pm on a Sunday often does not.
Your booking rate. Out of ten quoted jobs, how many become work? You know this one. Do not use an industry average.
Your average invoice. Last month's total revenue divided by jobs completed.
Then multiply only the calls you missed during hours you were open by your booking rate by your average invoice. That is your number. It will be smaller than any calculator's, and it will be real.
What to do with it
If the figure is small, good. Do not buy anything. Check again next quarter, or after a busy season.
If it is large, the cheapest fix is almost never software. It is usually one of these, in this order:
- Decide who answers the phone, by name, and tell them.
- Turn on voicemail-to-text so a missed call at least becomes a readable note.
- Add an automatic text back on a missed call, so the caller knows they were seen.
- Only then look at anything that costs a monthly fee.
We sell number three and number four. We still think you should do one and two first, because they are free and they work.
A note on statistics in this industry
The 85% figure above, and the 62% unanswered figure that appears on a lot of marketing sites including parts of ours, trace back through agency blog posts that cite other agency blog posts. We are currently working back through every statistic on this site to its original source. Anything that does not survive that check is coming off the page.
If you find one we have missed, tell us and we will fix it.