Remote work is often described as location-free. In practice it is connection-dependent, and connections are not evenly distributed. Two people with identical skills, identical job descriptions and identical employers can have very different chances of working from home, depending on whether the nearest reliable signal is in their kitchen or at a petrol station twenty minutes away.
This piece looks at how uneven connectivity quietly decides who gets flexible work, what that costs the people left out, and what a business can do to widen the choice without pretending that a software setting can fix a tower that does not exist. It is a hypothesis about a South African pattern grounded in general connectivity research, not a measurement of any particular workforce.
Averages hide the edges
DataReportal's 2025 overview of South Africa reports internet penetration of roughly 79% of the population at the start of that year, and notes that mobile connections far outnumber people, partly because many connections carry only voice and SMS and some people hold more than one. That sounds like broad access. But a national average says little about the quality of the connection at a given address, and a person can be counted as an internet user while having a link that drops during a video call.
The International Telecommunication Union's global connectivity statistics make a related point: urban-rural gaps in connectivity are narrowing but persist, and coverage remains concentrated in higher-income countries. The GSMA's work on the sub-Saharan mobile economy examines the same divide between where networks reach and who actually uses them. Taken together, they suggest the right question is not whether a region has coverage on paper, but whether a person can run the work their job requires, reliably and at a price they can afford.
Why this matters for work
Many office jobs now assume video calls, cloud documents, remote desktops and secure logins. All of these are tolerant of a slow connection until they are not. A call that freezes for ten seconds is irritating at home and costly when it is a client meeting. A remote desktop session that drops forces a re-login and, depending on the system, may lose unsaved work. Load-shedding adds another layer, since power cuts can take out routers, towers and cellular backhaul at the same time as the lights.
The result is that flexibility is allocated, without anyone deciding it, to people who live where the infrastructure is good. A staff member who lives on a farm outside a small Karoo town or in a village in the Eastern Cape may be just as capable but is quietly treated as a person who must come into the office, or who cannot be given work that needs a stable link.
An illustrative case
Consider an illustrative example, not a real case: a Bloemfontein-based professional services firm of thirty people lets staff work remotely two days a week. A senior analyst lives on a smallholding about forty minutes from town. The only dependable service is a mobile data plan on one network, usable only in the yard, with a signal that weakens when it rains.
On her remote days she works in short bursts: downloading files in the yard, reviewing offline, uploading in the afternoon. She leaves video calls to colleagues and sends notes afterwards. Her manager, seeing her camera off and her replies delayed, quietly concludes she is less engaged. In the annual review, she is told she should be more visible. She is also the only person on the team who has to drive to a coffee shop to join an important call.
No one set out to treat her unfairly. The firm's remote working policy simply assumes a connection that she does not have, and measures her against it.
The human tax: access and optionality
The first cost is access. A person who cannot reliably connect has fewer opportunities in the first place: fewer roles they can apply for, fewer tasks they are given, fewer meetings they can attend. The second is optionality. Remote work is partly valuable because it lets people choose where to live, to be near family, to care for an elderly parent, or to live in a place they can afford. When the connection decides, that choice narrows to the places that already have good infrastructure, which tends to reinforce existing advantages.
The Moments piece on when digital-only leaves some customers behind makes a similar argument about customers, and why accessibility is also an operational question offers a way of thinking about it that applies to staff as well. Another angle is the cost to the business: a firm that cannot use the skills of people outside the main metros is choosing from a smaller pool than it could.
What could realistically change
Employers cannot build towers. But they can stop designing work as if every connection were the same.
- Ask what people actually have. A short, voluntary and private survey, covering the type of connection, typical speed and reliability, and the cost of data, gives a better basis for policy than assumptions. POPIA applies to what is collected, so keep it to what is needed.
- Separate the work that needs a live link from the work that does not. Reviewing documents, drafting and analysis can often be done offline and synchronised later. Real-time work such as client calls and remote desktop sessions may be scheduled for days when someone is in the office or at a better-connected site.
- Design meetings to tolerate poor links. Share an agenda and papers in advance, send written summaries, and accept audio-only participation as normal rather than as a lack of engagement.
- Support the connection where it makes sense. A data allowance, a second SIM on another network, a signal booster or a small power backup for the router can cost less than the lost productivity. Compare these options with the figures in the article on the risk of a single internet connection, which applies to staff homes as well as offices.
- Judge by results. Assess people on what they deliver rather than on whether their camera was on. This is a close relative of the idea in the quiet power of choosing when to respond.
The business outcome
A wider talent pool, lower attrition among people who would otherwise have to move or quit, and fewer unspoken penalties for staff whose circumstances differ. The business also gains a clearer view of which work genuinely needs a live connection and which was assumed to, which can save money elsewhere.
The human dividend
The dividend is optionality and autonomy: the ability to choose where to live and how to arrange the working week without being quietly excluded by a signal. Not everyone will want to work remotely, and some will prefer an office for good reasons. The point is that the option should depend on the work and the person's preferences, not on an accident of infrastructure.
The trade-offs
- Support has a cost. Data allowances and equipment cost money, and must be offered fairly to avoid new inequities.
- Some work genuinely needs a stable link. Not every role can be adapted, and being honest about this is better than pretending otherwise.
- Asynchronous work needs discipline. Written summaries and shared documents only help if everyone uses them.
- Employers cannot fix the network. Policy can reduce the penalty of poor connectivity but cannot remove it.
Why it matters
When remote work is available only to those with the best connections, geography becomes a hidden selection criterion. Businesses that look honestly at what each task needs, and adjust, gain access to people they were previously excluding, and give those people a fairer chance to show what they can do.
Sources and further reading
- Digital 2025: South Africa — DataReportal – Global Digital Insights (last verified 10 October 2026)
- Measuring digital development: Facts and Figures 2025 (last verified 10 October 2026)
- The Mobile Economy Africa 2026 (last verified 10 October 2026)
Related Mellorca reading
- The Employee Locked Out by a System Built for Someone Else
- The Friction of Signing Up for Something at Work
- The Apprentice Who Learns Workarounds Instead of Craft