Most people know the small movement. The phone comes out of a pocket on a Sunday afternoon, the thumb opens the mail app, and the eyes run down the list. Nothing needs to be answered. It is a check, not a task. Then the phone goes back, and some part of the person stays in the list.
This is a common enough scene that it hardly registers as a scene at all. It is worth slowing down, though, because the question it raises is not really about discipline or etiquette. It is about where work lives.
In many businesses, the inbox is not only a place where messages arrive. It is the task list, the approval queue, the record of who promised what, and the only place a client's request exists until someone acts on it. When that is true, checking the inbox is not idle habit. It is the only way to find out whether something is about to go wrong.
What the data shows, and what it does not
Microsoft's Work Trend Index has published telemetry on how working patterns spill past the working day. In its analysis of what it calls the infinite workday, Microsoft reports that nearly 20% of employees who are actively working at the weekend check email before noon on Saturday and Sunday, and that over 5% are back in their email on Sunday evening after 6pm. The average worker in that data sends or receives more than 50 messages outside core business hours.
The same body of research found that people spend a large share of their working time communicating rather than making things. In an earlier Work Trend Index report, Microsoft found that 57% of time in its Microsoft 365 data went on communicating, through meetings, chat and email, and that 68% of surveyed people said they did not have enough uninterrupted focus time in the working day.
Those figures describe users of one vendor's tools across many countries. They are not a measurement of South African small businesses, and they do not tell us why any individual opened their mail on a Sunday. What they do establish is that after-hours inbox activity is a pattern, not a quirk, and that a great deal of the working day is spent moving information between people rather than acting on it.
The inbox as an unofficial system of record
Consider an illustrative case. A Pretoria engineering supplier with about twenty-five staff runs its quotes, supplier queries and customer complaints through two shared mailboxes and the personal inboxes of three senior people. There is a job-tracking spreadsheet, but it is updated at the end of the week. If a customer emails on Thursday afternoon to say an order is wrong, the only record of that problem, until Monday, is a message sitting unread or half-read in someone's inbox.
The operations manager in that business might look at her mail on Saturday morning. Not because anyone asked her to, but because there is no other way to know whether the Thursday complaint was picked up. The inbox is holding a commitment, and she is the only safeguard against it being dropped.
This is the mechanism worth naming. When the inbox is the only place an open item exists, three things follow:
- Ownership is implicit. Whoever the message reached is assumed to own it, even if they were only copied in.
- Status is invisible. A message that has been read looks the same as one that has been dealt with, unless someone remembers.
- Memory becomes the control. The only thing preventing a dropped commitment is a person carrying a mental list of what is still open.
That mental list does not switch off when the laptop closes. This is the inferential step, and it should be stated as one: it is reasonable to think that a person who is the sole holder of open commitments will check on them out of hours, but no study cited here measures that link directly. It is a hypothesis grounded in how the work is arranged.
The pattern is familiar from Mellorca's work on why shared inboxes become operational bottlenecks and on processing routine requests through email chains. Those pieces look at the cost to throughput. This one looks at what the same arrangement costs the person carrying it.
The human tax: presence
The obvious tax is time. Ten minutes on a Saturday morning is not nothing, especially when it happens every weekend.
The less obvious tax is presence. A person can be at a child's sports match, at a braai with friends, or simply resting on a couch, and still be partly occupied with whether the Thursday complaint was handled. The check itself may take a minute. The unresolved question can stay with them for much longer. Some people will recognise this; others will not, and they may genuinely be untroubled by a quick look at their mail. Neither experience should be assumed for everyone.
What can be said with more confidence is that the arrangement places the burden on whichever people happen to be most conscientious. The person who checks is often the one who cares most about not letting a customer down. The system relies on that care and gives nothing structural back.
What could change
The interventions that help are mostly about giving work a home outside the inbox, and they are not all technical.
- Turn requests into records. A complaint, quote request or approval that arrives by email can be logged as a ticket, task or case with a named owner and a status. That can be done with a simple shared tracker or a helpdesk tool; the important part is that the item exists somewhere visible.
- Make ownership explicit. Being copied on a message should not mean owning it. A rule about who picks up what, and a short daily check of unassigned items, removes much of the guesswork.
- Give exceptions a path. If something urgent does arrive after hours, there should be an agreed route for it, such as an on-call number, rather than an expectation that everyone watches the inbox just in case.
- Remove what does not need to be email at all. Routine approvals and status requests are often better handled through a form or a workflow, as discussed in why chasing people for updates is a systems failure.
None of this requires a large software project. In the illustrative engineering supplier, moving customer complaints into a shared tracker with a named owner and a Friday afternoon review would mean that, by the weekend, the operations manager can see whether the Thursday complaint was picked up without opening her mail. The check either becomes unnecessary or becomes a deliberate choice.
Trade-offs
Moving work out of the inbox is not free of cost or risk.
- A new tracker can become one more place to check, which simply moves the habit rather than ending it.
- Some people feel more in control when everything sits in their own mailbox. Taking that away can feel like losing visibility, at least at first.
- Logging every request takes a little time. If the tool is clumsy, people will route around it and the inbox will quietly become the real system again.
- Some people prefer reading mail at odd hours because it suits the shape of their lives. A redesign should make that a choice, not remove it.
The aim is not to forbid anyone from opening their inbox at the weekend. It is to make sure nobody has to.
Why it matters
The business case for giving work a proper home is usually made in terms of response times, dropped requests and duplicated effort. Those are real. But there is a second outcome that rarely appears on a report: the person who used to be the safety net can stop being one.
That is the dividend in this moment. Not a promise of rest or happiness, which no tracker can deliver, but something narrower and more defensible. When open work is held by a system that anyone can see, the people who were holding it in their heads can put it down. Their attention can stay where they are. A Sunday afternoon can be about whatever they choose it to be about.
This sits alongside the argument made in Friday evening should start on Friday evening, which looks at work that overflows its hours. The inbox is one of the quieter routes by which that overflow travels. Finding where commitments really live, and giving them a better home, is one of the more practical places to start.
Sources and further reading
- https://www.microsoft.com/en-us/worklab/work-trend-index/breaking-down-infinite-workday (last verified 10 October 2026)
- https://www.microsoft.com/en-us/worklab/work-trend-index/will-ai-fix-work (last verified 10 October 2026)
Related Mellorca reading
- The Report Someone Rebuilds by Hand Every Friday
- The Meeting That Could Have Been a Shared Record
- Why Holidays Still Need a Laptop in the Suitcase