Customer service & automation
How much do missed customer requests cost? How to make a realistic estimate
An honest way to calculate it: first count what you know for certain, then make your assumptions visible and work out a low, middle and high scenario. With formulas, a step-by-step example and a seven-day count.
Claudia Academy · Bits PleasePublished October 16, 2026 9 min read
An unanswered call or a WhatsApp message left sitting can be a customer who buys elsewhere. But how much does that really cost you? A number you read somewhere says little about your own business. This article is for business owners who want an honest estimate, based on their own numbers. You will learn the difference between what you can count and what you assume, a simple formula, where estimates often come out too high and how to collect your own numbers in seven days. With a complete step-by-step example in SRD.
What you can count and what you assume
A realistic estimate starts with a strict separation. Some numbers you can count. Others are assumptions: reasonable estimates of what would have happened. Both are useful, as long as you don't mix them up.
| Number | Type | Where it comes from |
|---|---|---|
| Number of unique requests that went unanswered | Fact (counted) | Your call history, WhatsApp and a week of tracking |
| Average order value | Almost a fact | Your cash register, invoices or sales records over a recent period |
| Share that was a real sales opportunity | Assumption | Derived from the requests you did answer |
| Share that would have bought with a timely response | Assumption | Derived from your own answered requests, with a margin |
The first number is an observation: this many people got no answer. The outcome of the whole calculation is not. It remains hypothetical revenue: what could have come in if you had responded on time. Nobody can say for certain what a customer who got no answer would otherwise have done.
The formula, part by part
Hypothetical gross revenue per week
Missed unique requests × % possibly qualified × % that would have bought with timely follow-up × average order value
This is revenue, not profit, and not a guaranteed amount.
What each part means:
- Missed unique requests: people or questions, not calls. If the same customer calls three times, that is one request. Take out what you know for certain: spam, suppliers, wrong numbers and customers who later ordered through another channel after all.
- % possibly qualified: with a missed call you often don't know what the caller wanted. This percentage estimates what share was a real sales opportunity. Base it on the new requests you did answer: what share of those were about something you sell?
- % that would have bought with timely follow-up: the most uncertain number. Look at how many of your answered, qualified requests became an order or appointment. Use that as your starting point and also run the numbers with a lower value, because you don't know whether the missed customers were just as serious.
- Average order value: the revenue from one type of order divided by the number of those orders, over a recent month. Take the type of order that comes in by phone and WhatsApp, not your total revenue if that is made up differently.
Converting to a month
Amount per week × 52 ÷ 12
A year has 52 weeks and 12 months, so a month has about 4.33 weeks on average. If you multiply by 4, your estimate comes out about 8% too low.
Rough contribution (optional)
Hypothetical gross revenue × gross margin in %
Gross margin = (revenue − cost of goods or other direct costs) ÷ revenue. Fixed costs such as rent and wages are not included, so this is not net profit either.
Why that last step? Because an SRD 500 order doesn't earn you SRD 500: you also have to buy ingredients, parts or materials. If you don't know your gross margin, ask your accountant, or skip this step and explicitly call the amount revenue.
Where estimates often come out too high
The calculation itself is simple. The pitfalls are in what you put into it. Watch out for these five:
- Repeat callers. Someone who calls four times is one customer. If you count calls instead of people, you count the same opportunity several times.
- Capacity. If you were fully booked at that moment or the kitchen was full, an extra request might not even have been able to become a sale. Only count requests you could actually have served.
- Type of business. Someone who wants a takeout meal easily orders somewhere else. A customer who is used to their hairdresser, or whose car is already in your shop, is more likely to call back. So a missed request does not become a lost sale equally often in every business.
- Unknown outcome. You don't know what the customer did next. Maybe they came by, sent a WhatsApp message later or ordered through a colleague. Compare missed requests with your orders and conversations from the days after: anyone who came back doesn't count.
- One week is one week. A holiday or an unusually busy or slow week skews the picture. Note what was unusual and repeat the count later if you are in doubt.
Step-by-step example: a restaurant in downtown Paramaribo
Worked example
Example (fictional): a week of missed requests at a restaurant
A restaurant downtown takes takeout orders, delivery orders and reservations by phone and WhatsApp. For seven days, the owner records every request. All the numbers below are sample data, not benchmarks.
| Count | Number |
|---|---|
| Calls and messages in total | 163 |
| Unique requests after merging repeats | 118 |
| Of these, never answered | 29 |
| Of those, later ordered or came by after all (checked) | 4 |
| Of those, definitely not a customer (supplier, sales rep, wrong number) | 3 |
| Missed unique requests used in the calculation (29 − 4 − 3) | 22 |
For the assumptions, the owner looks at the 89 requests that did get an answer. Of those, 19 were from existing customers about an ongoing order; he leaves those out. Of the 70 answered new requests, 49 were about an order or reservation: 70%. Of those 49, 31 actually became an order or reservation: about 63%. Because the missed customers may have been less serious or may already have ordered elsewhere, the owner uses 45% for the middle scenario. The average value of a phone or WhatsApp order last month was SRD 450, calculated from the cash register.
Middle scenario per week
22 × 70% × 45% × SRD 450 = SRD 3,118.50
Intermediate steps: 22 × 0.70 = 15.4 requests with a real chance; × 0.45 = 6.93 orders; × SRD 450 = SRD 3,118.50.
Middle scenario per month
SRD 3,118.50 × 52 ÷ 12 = SRD 13,513.50
With a gross margin of 35% (fictional), the rough contribution is about SRD 4,730 per month.
Six of the 22 missed requests came in on Friday evening, when the kitchen was already running at full capacity. So in the low scenario, the owner uses 16 requests instead of 22.
| Scenario | Calculation per week | Per week | Per month | Rough contribution per month (35%) |
|---|---|---|---|---|
| Low | 16 × 55% × 30% × SRD 450 | SRD 1,188 | SRD 5,148 | SRD 1,802 |
| Middle | 22 × 70% × 45% × SRD 450 | SRD 3,119 | SRD 13,514 | SRD 4,730 |
| High | 22 × 80% × 60% × SRD 450 | SRD 4,752 | SRD 20,592 | SRD 7,207 |
How does the owner read this? Not as “we are losing SRD 13,514 a month,” but as a range based on his own assumptions. Even the low scenario is reason enough for him to organize callbacks better. The count also showed when things went wrong: mainly between noon and 1:30 PM and on Friday evening. That is where the solution starts.
Whether an amount is a reason for you to change something depends on what the solution costs, in time or in money. More on that under “What to do with the outcome.”
How to collect your own numbers in seven days
Without your own counts, every calculation remains a guess. Tracking for a week takes a few minutes a day and gives you all the numbers for the formula.
Get the template
Use the customer contact audit or your own list with the same columns: date, time, channel, contact type, subject, status and outcome.
Record every request for seven days in a row
Phone, WhatsApp, email and walk-ins. At the end of each day, copy the missed calls from the call history of every phone that customers call.
Mark repeats
If you recognize the same number or the same question, set that row to “Repeat message” so that you count people instead of calls.
Record status and outcome
Answered, answered late or unanswered, and what it led to: an order or appointment, information, not interested, open or unknown.
After a few days, check who came back
Check whether missed customers later ordered, sent a WhatsApp message or came by after all. Take those out of the missed requests.
Calculate your two percentages
Divide the answered new requests about something you sell by all answered new requests: that is your % qualified. Divide the number that became an order or appointment by those qualified requests: that is the basis for your conversion rate.
Get the average order value from your records
Take a recent month and the type of order that comes in by phone and WhatsApp.
Work out three scenarios
Enter your numbers in the Missed requests calculator, or use the formulas above. For each scenario, write down which assumptions you used.
A full explanation of this count, with an example and a checklist, is in lesson 1: map your customer contact.
Customer contact audit (spreadsheet)
Record every customer request for seven days. The Summary tab automatically counts your unique requests, the response rate and the average response time.
What to do with the outcome
- Compare it with what a solution costs. What does it cost to be easier to reach at busy times, in staff hours or in money? Put that next to your low scenario, not the high one.
- Look at the pattern, not just the amount. On which days and at which hours were requests left unanswered, and through which channel? That is where the first improvement lies.
- Measure again in the same way. After a few weeks with a new way of working, track seven days again, with the same definitions. Only then can you compare the outcomes fairly.
- Show the assumptions. Anyone who sees the amount should also see how it was calculated.
The first improvements don't have to cost extra staff: fixed check times, a callback process and one shared follow-up list. How to approach that is explained in improving customer service without hiring more staff.
Checklist for an honest estimate
- I count unique requests, not calls.
- I have taken out spam, suppliers and customers who came later after all.
- My percentages come from my own answered requests.
- The average order value comes from my cash register or invoices.
- I have accounted for times when I was fully booked anyway.
- To get a monthly amount, I multiply by 52 ÷ 12, not by 4.
- I call the outcome hypothetical revenue: not profit and not a loss.
- I have a low, middle and high scenario, with the assumptions alongside.
Does your count show that mainly standard questions or calls at busy times are left unanswered? Then Bits Please can set up Claudia to take over suitable parts of that, such as taking calls and recording requests, depending on the package and integrations. The calculation and the improvements in this article are separate from that.
Frequently asked questions
Why do I use unique requests and not missed calls?
Because one customer who calls three times buys once at most. If you count calls, you count the same opportunity several times and your estimate comes out too high.
Is the calculated amount my loss?
No. It is hypothetical gross revenue: some of those customers would never have bought, some came back later anyway, and revenue is not profit. It is an estimate to support decisions.
How do I get a conversion rate if I have never tracked it?
Track for a week which answered requests became an order or appointment. That gives you a first, rough basis. Also run the numbers with a lower and a higher value, because one week doesn't tell you everything.
Why do I multiply by 52 ÷ 12 and not by 4?
A month has about 4.33 weeks on average. If you multiply by 4, your estimate comes out about 8% too low; 52 ÷ 12 gives the correct average.
Does this calculation also work for a salon, clinic or garage?
The method works for any business that receives requests, but the assumptions differ. A regular customer, or someone whose car is already with you, is more likely to call back than someone who wants to order something quickly. So always use your own numbers and three scenarios.
Free downloads for this article
Missed requests calculator
Work out with your own assumptions what unanswered requests could hypothetically be worth. Every formula is shown; nothing is stored.
Customer contact audit (spreadsheet)
Record every customer request for seven days. The Summary tab automatically counts your unique requests, the response rate and the average response time.