Missed calls are the quietest line on a restaurant's books. Rent shows up on a lease, food cost shows up on invoices, but the caller who rang out during Friday service shows up nowhere at all. This post walks through the arithmetic in the open: what the inputs are, where they come from, what the model gets wrong in both directions, and what fixing the problem costs in 2026.
The four numbers that drive everything
- Calls per week. Pull this from your carrier's call log rather than guessing; most owners guess low because they only remember the calls someone answered.
- Miss rate. The share of calls that ring out, hit voicemail, or get a busy signal. It clusters brutally: calls arrive exactly when your team cannot answer, so a 25% overall miss rate often hides a 50% miss rate during the dinner rush.
- Share of missed callers who buy elsewhere. The figure that circulates in the telecom and answering-service industry is that about 85% of callers who reach voicemail do not call back. Treat it as an industry estimate, not a law of physics; we will stress-test it below.
- Average phone ticket. Takeout and delivery phone orders commonly run $25 to $40. Use your own POS average for phone and pickup orders if you have it; we use $35 as the default.
None of these four is precise for your restaurant until you measure it. The point of the exercise is not decimal accuracy; it is finding out whether your leak is a rounding error or a salary.
The worked example, step by step
Take a neighborhood takeout spot: 100 calls a week, 25% missed, 85% of missed callers lost, $35 average ticket.
- Missed calls: 100 x 25% = 25 per week
- Callers lost to competitors: 25 x 85% = 21.25 per week
- Lost revenue: 21.25 x $35 = $743.75 per week
- Monthly: $743.75 x 52 / 12 = $3,223
- Yearly: $743.75 x 52 = $38,675
Those are the default inputs in our free missed-call calculator. Every field is editable and the math runs entirely in your browser, so you can swap in your own numbers in about a minute.
Three restaurants, three answers
The same formula produces very different losses depending on volume, miss rate, and ticket size. All three rows below keep the 85% assumption.
| Scenario | Calls per week | Miss rate | Avg ticket | Lost per month | Lost per year |
|---|---|---|---|---|---|
| Quiet cafe | 60 | 10% | $22 | $486 | $5,834 |
| Neighborhood takeout | 100 | 25% | $35 | $3,223 | $38,675 |
| High-volume pizza shop | 180 | 30% | $38 | $7,558 | $90,698 |
Stress-testing the 85% figure
The "85% of voicemail callers never call back" claim has circulated in industry material for years without a rigorous public study behind it, which is exactly why it should not be swallowed whole. So cut it down and rerun the takeout example:
- At 50% lost: 25 x 50% x $35 = $437.50 a week, or about $1,896 a month and $22,750 a year.
- At 25% lost, meaning three of every four missed callers dutifully call you back: about $948 a month and $11,375 a year.
Even the most charitable assumption leaves a four-figure monthly leak on a moderately busy phone. The conclusion survives the skepticism.
What the model leaves out, in both directions
The model understates in three ways. It counts one lost ticket per lost caller, but a regular who switches shops takes a year of Fridays with them. It ignores catering and group orders, which arrive by phone and are the largest tickets most independents see. And it ignores the review a frustrated caller writes after ringing out twice.
It also overstates in ways an honest analysis should name. Not every call is an order: suppliers, wrong numbers, and spam inflate raw call counts, so apply the miss rate to order-intent calls if you can separate them. Some missed callers reorder on your website instead, which is a saved sale the model counts as lost. And if your kitchen is already at capacity on Friday night, an answered call may shift revenue to a slower hour rather than add it. The honest read: your true number is probably not $3,223, but measured against your own inputs it is very unlikely to be near zero.
What fixing it costs in 2026
Option 1: staff the phone. Twenty hours a week of rush-hour phone coverage at $15 an hour - a typical counter wage in most markets, well above the federal minimum - is $300 a week, roughly $1,300 a month. It is the most human option, and the most expensive, and it still misses the 9 p.m. Sunday caller and the sick day.
Option 2: an answering service. Human answering services typically bill $1 to $2 per operator-minute, so 200 minutes of call traffic runs $200 to $400 a month. Most take messages and reservations well, but few will complete a modifier-heavy food order end to end, which is the call that actually carries revenue in a restaurant.
Option 3: AI answering. Disclosure: this is the category our product is in. AskSiya answers 24/7 from $19 a month for 60 minutes of calls ($79 for 200, $199 for 600), takes complete orders with modifiers, reads totals back with tax, prints kitchen tickets, and speaks 100+ languages. It does not integrate with your POS yet - orders land on a dashboard order board - and the wider market has strong alternatives, which we compare honestly in our restaurant AI roundup. Full pricing is on the pricing page.
The bottom line
Count your calls for one week, apply your real miss rate and ticket size, and put a defensible discount on the 85% figure. For most restaurants the result lands between $500 and $7,500 a month, while every fix on the menu costs between $19 and $1,300. Whatever you choose - staffing, a service, or software - the expensive option is the ringing phone nobody answers.