The message arrives at 19h40 on a Thursday, or at 10h on a Saturday, and just sits there. By Monday morning, the customer has already resolved it elsewhere. It’s not a lack of effort from the team: it’s a schedule designed for office hours, not for when the customer actually writes.
In this article
The cost of uncovered hours
In many WhatsApp sales operations, between 20% and 30% of new messages arrive outside business hours. Let’s use 26% on 300 monthly contacts: 78 conversations start when no one is watching.
If these 78 are only answered the next day, their closing rate plummets. Suppose it drops from 12% to 5%: that’s 3.9 sales instead of 9.4 — 5.5 lost sales per month. At R$ 2,400 per ticket, that’s R$ 13,200.
Compare that with the cost to cover. A Saturday on-call agent, four hours, four Saturdays per month, at R$ 25 per hour, costs R$ 400. Extending coverage at the edges — half an hour before opening and one hour after closing — is usually handled with internal rotation, at no extra cost.
You don’t need to cover 24 hours. You need to cover the hours when your customer writes — and you discover those hours with data, not opinion.
The hourly volume map
Before changing the schedule, build the map. Export or record, for two to four weeks, the arrival time of each new message. Create a simple table: rows with two-hour intervals, columns with days of the week, and the number of contacts in each cell.
Three patterns almost always appear:
Lunch peak. Between 11h30 and 13h30 usually concentrates a high volume — precisely when part of the team is away. This is the most common coverage gap and the easiest to fix with rotation.
Evening wave. From 19h to 22h, mainly in businesses aimed at end consumers. People can only handle personal matters after work.
Saturday morning. For residential services, renovations, health, beauty, and automotive, it is often the second busiest period of the week.
With the map in hand, the decision shifts from "we need to serve more" to "we need someone from 19h to 21h and on Saturday until 13h". One more question worth answering: what is the closing rate by time slot? Sometimes the evening volume is high but quality is low — and then the decision changes.
Continue reading for free
Fill out once and unlock this and all other blog articles. We don’t charge anything — we just want to know who we’re writing for.
Takes 20 seconds. Prefer to talk directly on WhatsApp?
Three scheduling models
Choose based on your volume and team size.
1. Edge rotation (small team, up to 3 people). No one changes shifts; just one person per day opens 30 minutes earlier and another closes 1 hour later, rotating weekly. Covers the gap at the start and end, where most delays happen, at no extra cost. It’s the first step for almost every company.
2. Staggered shifts (medium team, 4 to 8 people). Instead of everyone from 9h to 18h, two blocks: 8h–17h and 11h–20h. The operation gains three hours of coverage without hiring anyone, and the lunch peak is covered by design. Requires an explicit handover rule between shifts.
3. On-call with reduced scope (night and weekend). One person on standby, with a clear and limited task: respond, qualify, and schedule — no closing deals, no price negotiation. Small scope allows cheap on-call and avoids the person needing full autonomy.
In any model, the queue needs to be unique and have a visible owner: without this, "everyone sees and no one takes responsibility" — the problem addressed in multiple agents on the same number.
Shift handover
The gap between two shifts is where conversations get lost. A five-minute routine solves this, and it needs to be written down:
Whoever leaves leaves the queue clean. No conversation without status. Each is marked as resolved, waiting for customer, or waiting for service.
Whoever leaves notes what needs continuity. Two lines per case: what the customer asked for and what the next step is. Without this, the customer retells the story — and retells it with irritation.
Whoever enters confirms receipt. An internal message of "I took over the queue" avoids the ten-minute gap where no one considers themselves responsible.
If your team transfers conversations between people, the transfer needs to carry the context — name, what was discussed, what is pending. Transfer without context is the polite version of telling the customer to start over.
What to write outside business hours
Automatic replies outside business hours are not decoration: they hold the customer for a few hours. But most companies write the useless version — "We received your message, we will get back to you soon."
The version that works has three parts: confirms receipt, says when someone will return, and asks for something that speeds up service.
"We received your message! Our team serves Monday to Friday, from 8h to 18h, and Saturday until 13h. You’re next in line tomorrow morning. To speed things up: tell me in one line what you need and, if it’s a quote, the city — so I can come prepared with the answer."
Three details make a difference. Give the exact time of return, not "soon." Ask for information, because waiting becomes prep work. And if there is a real emergency channel, inform what it is — in emergency services, this prevents the customer from looking for someone who is available now.
A rule to avoid embarrassment: the automatic message should not promise what the schedule does not deliver. If you say "we respond within 1 hour" and respond in four, you create an expectation that works against you. The limits of what to automate are in service automation.
When to hire and when to just reorganize
The question always comes in the same format: "Do I need another agent?" Before hiring, three checks usually restore capacity at no cost.
Check workload distribution. It’s common for one agent to answer twice as many as another due to a queue without rules. In this case, the problem is distribution, not team size.
Measure time spent on repeated questions. If half the conversations start with the same five doubts, saved quick replies and simple triage return hours per week — without removing people from service.
Look at the rate of reopened conversations. A customer who returns three times with the same issue indicates incomplete service, and incomplete service consumes more time than well-done service would.
If after these three corrections the response time remains above target during peak hours, then the hiring calculation is justified — and it’s usually favorable: one more agent costs less than the revenue lost in an uncovered time slot over an entire quarter.
How to know if the schedule is right
Four numbers, viewed by time slot and day of the week:
First response time by slot. This indicator reveals the gap. Overall average hides it; hourly breakdown shows it.
Percentage of contacts without response. Usually concentrated in one or two specific slots.
Closing rate by slot. If evening conversations close well, the on-call pays for itself; if they close poorly, maybe the right move is to reinforce business hours.
Volume per agent and per shift. Concentrated overload is what worsens response time in mid-afternoon.
Review the map every quarter. Seasonality, new campaigns, and channel mix changes shift the peak — and schedules that aren’t reviewed age quickly. These and other indicators are detailed in service metrics.
Scheduling mistakes that cost sales
Everyone eating lunch together. Concentrates the worst response time exactly at the peak.
No owner in the queue. Without a named responsible, the message waits while everyone thinks someone else answered.
Generic automatic reply. Without a return time, it holds no one.
On-call without scope. An on-call agent without defined autonomy stalls the conversation and annoys the manager on Sunday.
Covering hours without demand. Scheduling someone for 6h in the morning because "a message might come" is spending where there’s no return. The map prevents this.
Advertising outside covered hours. If the campaign runs 24 hours and service is 9h–18h, you pay clicks to the competitor.
Not reviewing. The schedule set last year doesn’t reflect the new channel you activated this month.
Scheduling is not an HR issue: it’s a commercial decision. Covering the right three or four hours per day usually pays more than any budget increase — because it recovers conversations the company already paid to receive and simply let go cold.
Frequently asked questions
How many messages arrive outside business hours?
It depends on the audience, but in operations selling to end consumers it’s common for 20% to 30% of new contacts to arrive at night, during lunch, or on weekends. The way to find your number is to record the arrival time of each new message for two to four weeks and create a table with two-hour slots by day of the week — the map usually reveals three peaks: lunch, from 19h to 22h, and Saturday morning.
Is it worth having WhatsApp on-call coverage on weekends?
It’s worth it when Saturday or Sunday volume is relevant and conversations on those days close at a rate similar to weekdays. A four-hour on-call agent per Saturday, at R$ 25 per hour, costs about R$ 400 per month — an amount paid back with a single sale in most service businesses. The on-call should have a reduced scope: respond, qualify, and schedule, without negotiating price or closing contracts.
What to write in the automatic message outside business hours?
Three parts: confirmation of receipt, the exact time someone will respond, and a request for information that speeds up service. For example: "Received! We serve Monday to Friday from 8h to 18h and Saturday until 13h. To speed things up, tell me in one line what you need and the city." Avoid "we will get back to you soon" and never promise a timeframe the schedule can’t meet.
Centralize service without changing your number
Connection via QR or official API, single queue with defined owner, customer history that stays with the company, and sales funnel on the same WhatsApp panel.