Telecom expansion is often sold with a clean promise. More towers. Wider coverage. Faster speeds. Better connectivity. It sounds simple, as if the problem is mainly one of investment and ambition. Build enough infrastructure and the network will reach everyone.
I do not think it is that simple. In difficult terrain, telecom expansion is not just a technology problem. It is a geography problem. Mountains, forests, islands, valleys, dense cities, informal settlements, coastal weather, land access, power supply, road networks and user behaviour all shape whether a network succeeds or fails.
Coverage on a map can look impressive while real service remains poor. A signal footprint can cover a district in theory, yet fail inside homes, along roads, in valleys, behind ridgelines, or during bad weather. This is one of the quiet problems in telecom planning. The official map can say connected. The lived experience can say otherwise.
That gap matters because digital infrastructure is no longer optional. It underpins banking, education, logistics, healthcare, emergency response, work, commerce and public administration. When coverage is weak, people are not just inconvenienced. They are excluded from the systems that increasingly define economic life.
Complex terrain changes everything. A flat urban grid and a mountainous rural district are not the same planning problem. Yet telecom strategy is too often discussed in broad national terms, as though population coverage percentages tell the whole story.
They do not.
Radio signals behave differently depending on frequency, distance, elevation, obstruction and surface conditions. Hills block line of sight. Valleys create shadow zones. Forest canopies absorb and scatter signals. Dense buildings reflect and weaken coverage. Heavy rain can affect higher-frequency links. Remote islands create backhaul problems. Desert regions create maintenance and power challenges. Mountain roads make construction expensive and slow.
This is why a tower placed in the wrong location can be a very expensive disappointment. It may satisfy a coverage model at a high level, but fail to serve the actual pattern of demand on the ground. It may reach ridgelines but not villages. It may cover road corridors but not homes. It may serve tourists but not farmers. It may provide signal strength without enough capacity.
There is a difference between theoretical coverage and usable connectivity. I think that distinction is where serious telecom planning begins.
The global digital divide is often described as if it is closing in a smooth, inevitable way. More people online. More mobile subscriptions. More 5G. More devices. The direction is positive, but the pattern is uneven.
The International Telecommunication Union estimated that in 2024, 83 percent of urban dwellers globally were using the Internet, compared with only 48 percent of rural populations. It also estimated that 1.8 billion of the world’s 2.6 billion offline people lived in rural areas. That is not a small residual problem. That is the geography of exclusion. 
The numbers matter because they remind us that the last mile is not really the last mile. In complex terrain, it can be the hardest mile, the most expensive mile and the least commercially attractive mile. A telecom operator can achieve strong national coverage metrics while the remaining unserved areas are precisely the places where infrastructure is hardest to build and demand is hardest to monetise.
That is the uncomfortable commercial reality. The people most in need of connectivity are not always located where the investment case is easiest. Mountain villages, dispersed agricultural communities, islands, border regions and low-income settlements often fall into the gap between social need and commercial return.
This is where spatial intelligence becomes essential. Not because it magically solves the economics, but because it shows the real shape of the problem.
Traditional telecom planning has often focused on signal propagation. Where will the signal travel. What frequency should be used. How much coverage can a tower provide. Where are the dead zones. These questions are necessary. But they are not sufficient.
A technically efficient network can still be commercially poor if it does not align with real demand. Demand is not just population density. It is mobility, income, device ownership, business activity, school locations, health facilities, transport routes, tourism flows, agricultural markets, industrial sites and seasonal patterns.
A mountain valley with a small permanent population may look weak on population density alone. But if it contains a major road corridor, hydropower project, border crossing, tourist zone or mining operation, the demand profile changes. A coastal area may look quiet on census data, but become highly active during tourist seasons. A rural market town may serve as the digital and commercial hub for a much larger surrounding area.
This is why I dislike crude coverage thinking. It asks where people live, but not enough about how they move, work, trade and communicate.
Real demand mapping should combine population data with mobility patterns, economic activity, transport networks, institutional locations and service usage. The aim is not simply to cover land. It is to connect life as it is actually lived.
Population density is useful, but it can mislead. A dense settlement may produce strong demand, but it may also contain low purchasing power, difficult building access or high interference from existing infrastructure. A low-density area may look unattractive, but may contain high-value users, strategic assets or critical public service needs.
Telecom expansion fails when planners treat density as destiny.
In Southeast Asia, for example, demand is rarely uniform. Major cities produce heavy data consumption, but rural and peri-urban regions often contain fragmented but important demand clusters. Industrial zones, ports, fisheries, plantations, universities, hospitals, border checkpoints and tourism areas can create demand that does not show up in a simple residential density map.
This is where location intelligence becomes commercially useful. It allows operators to separate empty land from under-recognised opportunity. It also helps governments understand where market forces alone will not deliver adequate service.
I think the best telecom expansion strategies are built from two questions, not one. Where can the signal go. And where does the signal matter most.
People notice towers. They do not notice backhaul. But backhaul is often the real constraint.
A cell tower is only useful if it can connect traffic back into the wider network. In cities, fibre routes and existing infrastructure may make this relatively manageable. In remote or complex terrain, backhaul can be expensive, fragile and difficult to maintain. Fibre may require difficult trenching. Microwave links may need clear line of sight. Satellite may provide coverage but at higher cost or lower capacity. Roads may be poor. Power supply may be unreliable.
This is why telecom expansion cannot be planned through tower placement alone. You need to map roads, rights of way, power access, existing fibre, microwave visibility, terrain elevation and maintenance logistics. A site with excellent radio coverage may still be poor if it is hard to power, hard to connect and hard to repair.
There is a brutal practicality here. Networks are not built in PowerPoint. They are built in mud, heat, rain, forest, concrete, paperwork and politics.
A spatial model that ignores backhaul is not a network model. It is a coverage fantasy.
5G adds another layer of complexity. It has enormous potential, but it also makes geography more important, not less.
Higher-frequency 5G bands can carry large amounts of data, but they generally have shorter range and weaker penetration through obstacles. That makes them powerful in dense urban and industrial environments, but more challenging in rural or mountainous areas. Lower-frequency bands travel further and penetrate better, but they do not provide the same capacity. The right answer depends on geography and demand.
GSMA reported that mobile technologies and services generated around 5.8 percent of global GDP in 2024, equal to about $6.5 trillion in economic value. In Asia Pacific, GSMA estimated that mobile contributed $950 billion to GDP in 2024 and could reach $1.4 trillion by 2030 as 5G becomes more widespread. These are large numbers, but they also raise the stakes. If mobile infrastructure is now a major economic platform, poor spatial planning becomes an economic drag, not just a technical weakness. 
I think this is the point often missed in 5G debates. The issue is not only speed. It is industrial geography. Ports, factories, logistics hubs, smart cities, mines, farms and energy systems all have different connectivity needs. A 5G rollout that treats them as generic coverage zones will underperform.
The real opportunity is targeted deployment. Put high-capacity infrastructure where it creates measurable economic value. Use lower bands and alternative technologies where broad coverage matters more than peak speed. Align the technology with the terrain and the demand.
That sounds obvious. It often is not done well.
When people hear complex terrain, they often think of mountains or jungles. But dense cities are complex terrain too.
Urban signal environments are full of obstruction, reflection and congestion. High-rise buildings create canyons. Underground spaces need dedicated systems. Transport hubs produce extreme peaks in usage. Wealthy districts may consume huge amounts of data. Informal areas may have high population density but poor fixed infrastructure. Commercial zones empty at night. Residential districts surge in the evening. Stadiums, airports and shopping districts create temporary demand spikes.
A city is not one market. It is a mosaic of micro-markets.
This matters for site placement. A tower or small cell location should not be chosen only by geometry. It should be chosen by the interaction of geometry and behaviour. Where do people gather. Where do they move. Where do they complain. Where do dropped calls cluster. Where does mobile payment activity concentrate. Where are emergency services most dependent on reliable coverage.
The best urban telecom planning is less like blanket coverage and more like nervous system design. It must understand pressure points.
There is also a human issue that sits beneath the technical one. Connectivity is not meaningful if people cannot afford to use it.
A network may reach a community, but adoption can remain low if devices are expensive, data plans are unaffordable, digital skills are weak or local services are limited. This is why coverage statistics and usage statistics diverge. A person may live under a signal and still remain effectively disconnected.
That matters for demand modelling. If planners assume that coverage automatically creates demand, they may overbuild in some places and under-support others. Real demand is shaped by income, education, language, age, trust, device access and service relevance. A farmer may need weather alerts and market prices more than entertainment streaming. A small business may need payment reliability. A school may need stable broadband at specific hours. A clinic may need resilient connectivity during emergencies.
This is where I think telecom companies and governments need to be more honest. The purpose is not merely to paint the map with coverage. The purpose is to create usable connection. Those are not the same thing.
A serious GIS approach to telecom expansion brings the layers together.
It maps terrain, slope, elevation and line of sight. It models signal propagation by frequency and infrastructure type. It identifies shadow zones and interference risks. It overlays population distribution, settlement structure, income indicators, transport routes, public facilities, commercial centres and mobility patterns. It assesses fibre access, power availability, road access and maintenance feasibility. It compares capital cost with expected demand and public value.
Then it produces something useful: a ranked expansion strategy.
Not just where towers could go, but where they should go. Not just where coverage is weak, but where weak coverage matters most. Not just where demand exists today, but where demand is likely to grow.
This is the difference between mapping and decision intelligence. Mapping shows the world. Decision intelligence helps you act within it.
Imagine a mountainous province with scattered villages, a main road through a valley, a growing tourism area, several schools, a district hospital and weak existing coverage. A basic planning model may place towers to maximise geographic area covered. That might produce a clean-looking coverage map.
A better model would ask more difficult questions. Which villages are in radio shadow. Which road segments have no emergency signal. Which schools could become public connectivity points. Which ridgelines provide the best line of sight. Which locations have power access. Which sites can be reached in the rainy season. Which tourist areas generate seasonal data demand. Which backhaul option is realistic.
The final plan may cover less land on paper but serve more real need. That is the kind of trade-off that separates intelligent planning from cosmetic planning.
Telecom expansion in complex terrain is not only about telecoms. It affects national development, regional competitiveness and social inclusion.
For governments, it supports education, healthcare, emergency response and digital public services. For operators, it improves capital efficiency and reduces wasted infrastructure spending. For businesses, it opens markets and improves logistics. For communities, it connects people to opportunity.
But the work has to be grounded. Terrain matters. Demand matters. Backhaul matters. Affordability matters. Maintenance matters. Politics matters.
I think the organisations that get this right will be those that stop treating connectivity as a blanket and start treating it as a spatial system. The aim is not to cover the most empty space. The aim is to connect the right places, in the right sequence, with the right technology, at the right level of resilience.
That requires evidence. It requires local context. It requires the humility to admit that the map is more complicated than the slogan.
Telecom expansion sounds futuristic, but much of it comes down to old realities. Hills. Roads. Weather. Power. Money. People. Distance.
The signal may be digital, but the problem is physical.
That is why complex terrain exposes weak planning so quickly. It punishes assumptions. It reveals whether the network has been designed around real geography or around convenient averages. It shows whether demand has been understood or merely estimated.
I would not judge a telecom expansion plan by how impressive its coverage map looks. I would judge it by how well it matches the lived geography of the people and systems it is supposed to serve.
Because in the end, connectivity is not about reaching land.
It is about reaching life.