Your sales team works nine to six. Your customers do not. They send a WhatsApp message at 11 pm after the kids are asleep, fill in the website form on Saturday morning, or ask a question from the car on Sunday. Most owners know that "something comes in overnight", but very few know how much, or how much of it later turns into revenue. Without that number you cannot decide whether you need someone on call, an after-hours assistant, or simply a rule that night enquiries are answered first thing in the morning.
This article gives you a step-by-step way to measure out-of-hours lead loss on your own data: which export to pull, how to tag the time windows, which three numbers to compare and how to turn the result into money. There are no borrowed industry percentages here — they rarely match a specific business.
What actually happens to an enquiry that arrives at 11:40 pm
"We reply in the morning" sounds reasonable, but it hides three different outcomes, each with its own cost:
- The customer waits and buys. Little or no loss. This is more common where decisions take a long time: education, property, B2B services.
- The customer waits but goes cold. In the morning they reply in one word or disappear after the first message. On paper the lead was handled; in practice it was lost. Your CRM records it as "no meeting booked".
- The customer goes to whoever answered at night. Your rep writes at 9:15 and the person has already booked elsewhere. This is the most expensive outcome and the hardest to see, because in reports it looks exactly like an ordinary "not interested".
So the question is not "how many leads come in at night" but how night leads perform further down the funnel compared with daytime leads. That comparison is the whole method.
Step 1. The export: which columns you need
Take at least 8–12 weeks. One month has too few weekends, and a single bank holiday or promotion can distort the picture. Export every inbound enquiry from your CRM, messaging channels and website form. If some channels never reach the CRM — a rep's personal WhatsApp, Instagram DMs — note that honestly; we come back to it under limitations.
| Column | Where it comes from | Why you need it |
|---|---|---|
| Date and time received | Lead creation time or the customer's first message | Tagging working vs out-of-hours |
| Channel | CRM source, chat type | The channel mix often shifts at night: messaging overtakes phone calls |
| Time of first meaningful reply | The rep's first real message or call | How long the lead actually waited |
| Reached the next stage? | Status: meeting, viewing, quote, booking | An early conversion signal that shows up faster than closed deals |
| Outcome | Won / lost / went silent | Final conversion |
How to define the moment of reply, and why an auto-responder does not count, is covered in our article on speed to lead and first response time. We use the same definitions here.
Step 2. Tag the out-of-hours windows
Add a "window" column and put every enquiry into one of four buckets. Four is better than two: weekday evenings and the middle of the night behave differently and need different fixes.
- Working hours — your real schedule, not the one on the website. If nobody answers between 1 and 2 pm, lunch is not working time.
- Weekday evening — from end of shift to midnight.
- Night — from midnight to start of shift.
- Weekends and public holidays — the whole day. Check which days your market actually treats as the weekend: in Saudi Arabia it is Friday and Saturday, in the UAE Saturday and Sunday, in the UK Saturday and Sunday plus bank holidays. Eid dates move every year with the lunar calendar, so tag them by the actual date, not last year's.
A quick sanity check: if your "working hours" bucket contains leads that waited more than two or three hours for a reply, the schedule in the spreadsheet does not match reality. Fix the boundaries before you go further.
Step 3. Three numbers that reveal the loss
A pivot table on the "window" column gives you everything. For each window, calculate:
- Share of enquiries. What percentage of total volume falls into evenings, nights and weekends. If it is under a tenth, a morning triage routine will probably solve the problem and you can stop here.
- Median wait. Not the average: one lead left over a long weekend ruins the mean, while the median shows the typical case.
- Conversion to the next stage and to a deal. This is the key figure. The gap between daytime and out-of-hours conversion is your loss in percentage terms.
Then the formula:
Lost deals for the period = out-of-hours enquiries × (working-hours conversion − out-of-hours conversion).
Lost revenue = lost deals × average deal value.
A worked example with illustrative numbers — replace them with yours. In a quarter you receive 600 enquiries, 170 of them in the evening, overnight or at weekends. Daytime leads convert to a deal at 12%, out-of-hours leads at 7%. The 5-point gap equals roughly 8 deals a quarter that could plausibly have been won. Multiply by your average deal value: that is the upper bound of what is at stake.
Note the words "upper bound". Part of the gap may be caused not by reply speed but by night leads being different — more accidental ad clicks, for example. How to test that is below.
Step 4. What it costs to cover the gap with people
The other half of the decision is the cost of covering out-of-hours time with staff: an on-call rep, weekend pay, shifted rotas. Our time-loss calculator helps here. It prices team hours rather than leads: people × hours per day × days per week × hourly rate. Add a process called "First reply in evenings and weekends" and be honest about the on-call hours. The "% routine" field is the share of those hours spent on standard questions — prices, location, availability, booking — rather than real negotiation.
Now you have two numbers side by side: lost revenue from step 3 and the cost of manual cover from the calculator. If the first is clearly smaller, on-call cover does not pay for itself and a disciplined morning triage is enough. If it is larger, the window is worth closing — and staff are not the only option.
What AI can handle and what stays with people
At night a customer rarely needs a full salesperson. They need a real answer and a clear next step. That is a job an AI agent connected to your messaging channels and CRM can do — provided it has up-to-date prices, services and available slots.
| What happens out of hours | What the AI does | What stays with a person |
|---|---|---|
| Question about price or availability | Answers from the price list, clarifies requirements | Individual discounts and non-standard terms |
| Wants to book | Offers free slots, records the booking | Rescheduling in special circumstances |
| Complex or high-value request | Collects the brief, promises a call at a specific time | The conversation and the proposal |
| Complaint or conflict | Captures the issue, hands it over flagged as urgent | The entire resolution |
There are limits. An agent cannot know what is not in its knowledge base, and it makes mistakes on rare questions, so hand-over rules to a human are needed from day one. We looked at where these systems break in practice in our piece on an AI assistant for Instagram DMs.
Where this calculation can mislead you
- Different traffic quality. Compare night and daytime conversion within the same channel. If the gap persists inside one source, reply time is the cause. If it disappears, the traffic mix is.
- Missing channels. If reps' personal chats are not in the export, your out-of-hours share will be understated — that is exactly where people tend to write in the evening.
- Small samples. With twenty night leads, a few points of difference is noise. Aim for at least a hundred out-of-hours enquiries, or a longer period.
- Seasonality. Do not compare nights in peak season with days in a quiet month. Use the same period for both windows.
The most reliable check is an experiment: cover the out-of-hours window for a month — with on-call staff or an automated first reply — and compare conversion with the previous period. That turns an estimate into a measured difference.
More on how enquiries and funnels work is in our Sales section. If you would like to go through your export together, get in touch: cost depends on scope and data, and the first conversation is about working it out.
FAQ
What if we have no CRM and leads live in WhatsApp or Telegram?
You can still do it. Export the chats, or manually go through the last eight weeks and log the customer's first message time, your reply time and the outcome. It takes a couple of evenings but gives a far more accurate picture than guessing.
What share of out-of-hours leads counts as large?
There is no universal benchmark. The practical test: if lost revenue from the step 3 formula is comparable to the cost of manual cover, it is worth acting on. If it is an order of magnitude smaller, a "night leads first" morning routine is enough.
Is an auto-reply saying "We'll get back to you in the morning" enough?
It beats silence, but the customer gets a promise rather than an answer. To test its effect, switch it on for a month and compare out-of-hours conversion before and after using the same calculation.
How often should we recalculate?
Once a quarter, or after a noticeable change: a new ad campaign, a new channel, a different schedule. Out-of-hours demand shifts along with your audience.