Coverage and resilience: designing dependable mobile connectivity for critical applications

The UK’s mobile networks have improved considerably after sustained investment. But Ofcom’s latest research shows that no single network performs consistently everywhere, and which one is strongest changes from place to place. For life-safety and other critical applications, that makes dependence on a single network the real risk. Multi-network connectivity, which can use whichever network performs best in each location, and resilient designs that recover automatically when one of them or a core network fails, give these devices a dependability that single-network connections cannot match. Coverage and resilience, designed in from the start, are what keep them connected.

Why does critical IoT need multi-network connectivity?

Critical IoT devices often operate in fixed locations where users cannot move to find a better signal. Because mobile performance varies by network, building, route, congestion and time of day, relying on one operator creates avoidable risk. Multi-network connectivity gives devices access to more than one available network, while resilient designs such as DualCore/rSIM add automatic recovery when the active path fails.

In June 2026, Ofcom published new research on the state of mobile service, alongside a discussion paper, Connectivity you can count on.

Mobile connectivity in the UK is good, and it is getting better. Since 2020, the mobile networks have invested around £10 billion to expand capacity and coverage, and 4G and 5G are now widely available. For most people, most of the time, mobile simply works, and it now underpins how we communicate, travel, pay and run businesses. That progress is a good starting point, and it is also why the remaining gap matters.

The headline is not that mobile is poor, but that it is not yet consistent enough on each network:

The clearest signal is not one weak network but inconsistency between them. In many places one network clearly outperforms the others, and which network that is changes from place to place.

For a consumer, patchy mobile performance is mostly a frustration, something to work around by moving to find a signal, switching to Wi-Fi, or trying again later.

For a telecare provider, a payments estate, an alarm installer, an IoT service provider or an infrastructure operator, it is a design risk, because the device has to keep working in one fixed location, at a point in time, whatever any single network does there.

Ofcom uses a practical benchmark for good performance: download speeds of at least 5 Mbit/s, upload speeds of at least 1.5 Mbit/s, and latency of no more than 50 milliseconds, the level needed for everyday tasks such as video calls and streaming. It proposes that networks should meet this standard at least 90 per cent of the time.

Today the UK sits below that mark. Based on crowdsourced data, good mobile performance is still achieved below Ofcom’s proposed 90% benchmark, with results varying across the UK nations and placing the UK towards the bottom third of comparable economies. The point is not that the networks have underdelivered, given the scale of recent investment, but that the remaining challenge is about consistency in the real world rather than headline coverage on a map. The practical implication is not that mobile is weak, but that depending on one network leaves a device exposed to that network’s worst day in that exact spot.

Ofcom identifies four types of location where performance most often disappoints: busy town and city centres, shared indoor public buildings, trains, and rural areas. No single operator reached the 90 per cent benchmark in any of the city centres examined. In shared indoor places such as shopping centres, stations and hospitals, performance was weaker and less predictable, particularly during busy periods. In rural areas, it declines steadily as locations become more remote.

One pattern matters for how organisations respond. Performance varies considerably both by location and by operator. This location-by-location variation is the fundamental logic behind multi-network SIMs, which is why they work so well across a wide range of sites and use cases at scale. The practical lesson, therefore, is that connectivity for a critical application cannot assume that a single network will be the strongest everywhere a device is deployed.

Trains are a different and slightly odd case. Across 24 segments of key railway lines in England, Scotland and Wales, the best-performing network delivered good performance on just 42 per cent of tests and the others on between 17 and 21 per cent, so performance fell short of Ofcom’s good performance threshold on between 58 and 83 per cent of tests. On-board Wi-Fi, provided by train operators, performed well just 1 per cent of the time. The issue is that even the best-performing network falls well short, which points to a structural problem of signal and capacity along the line rather than something competition between networks can resolve. That is why Ofcom is clear that competition between networks alone will not fix train connectivity, and that infrastructure-led measures such as track-to-train systems and satellite connectivity will matter. In shared indoor locations, dedicated indoor mobile infrastructure also has a role, with government, landlords, local authorities and industry all part of the wider solution. Ofcom has opened a consultation on these issues, inviting responses by 29 July 2026.

Ofcom’s accompanying consumer research helps explain how people actually judge their service. Most do not separate coverage from performance. They assess a network by whether it works when needed, and value consistency as much as, if not more than, raw speed. Small, predictable problems are usually tolerated, and connectivity is taken for granted until something goes wrong.

That experience is not evenly shared. Small business owners, carers, people with health needs and those without reliable home broadband depend on mobile connectivity more heavily and feel the impact more acutely when it falters. These are precisely the users served by life-safety and business-critical systems, from telecare and lone-worker devices to payment terminals and connected estates, where an unavailable connection at the wrong moment is more than an inconvenience.

For life-safety, healthcare, telecare, payments, transport and critical national infrastructure applications, a weak connection cannot be worked around the way a consumer would, by moving to find a signal or trying again later. A device has to connect reliably from the fixed location where it is deployed, using whichever available network can provide suitable performance there. For many IoT and critical applications the bandwidth required is lower than Ofcom’s consumer benchmark, but the need for availability and predictable recovery is often higher.

Ofcom’s findings point to two things that have to come together. A coverage map is only the beginning: it shows whether a network may be present, not whether a device will keep working when performance varies by building, route, operator, time of day or congestion. Coverage determines whether a network can be reached at all. Resilience determines whether the application keeps working when the best available network changes. For critical applications, the principle is to design for the real-world location where the device must operate.

• The first is coverage. In those rare locations where there is no usable signal at all across all three networks, the answer is infrastructure: denser networks, dedicated indoor router systems powered by higher gain external antennae, track-to-train solutions and satellite, along with the network investment expected to follow recent UK mobile market consolidation. This determines whether any network can be reached at a given location, and it is part of the route to closing the structural gaps on trains and in the most remote areas.
• The second is resilience. Where a usable signal does exist, but performance differs by network and from one moment to the next, the answer is to avoid depending on any single operator. This is where managed critical connectivity differs from an ordinary consumer SIM: the aim is not simply to attach to a network, but to give a device more than one viable path to service and to manage what happens when the path it is using degrades. This is the approach CSL takes. Our multi-network and roaming connectivity draws on the networks available in a given location, and CSL’s IoT SIMs use unsteered roaming, so a device is free to use them rather than being steered onto a single or preferred mobile operator.
• Resilient SIM technology takes this a step further. CSL’s patented rSIM® uses a DualCore design, holding two independent multi-network profiles on a single SIM. It monitors its own connection and, if the active network stops working, is designed to switch automatically to the second core, without manual intervention. For applications where an interruption carries real consequences, that built-in failover is the difference between a connection that recovers on its own and one that simply stays down.

This difference is easiest to see in practice. Consider a telecare alarm in a rural home. If the resident’s primary network has a coverage not-spot at the property but a second network reaches it, a single-network device is exposed, whereas a device that can use whichever network is available stays connected. The same logic applies to a card payment terminal at a busy event, where handsets may show 4G or 5G yet struggle to connect at peak times, and the ability to use another available network with usable performance can help keep the terminal taking payments. Where no network reaches at all, that becomes a coverage question rather than a resilience one, which is why the two measures matter together (and where satellite systems come into the equation).

For installers, service providers and anyone deploying connected devices across varied UK locations, the research is a prompt to weigh both coverage and resilience from the outset, rather than assuming a single network will be good enough everywhere. A few questions help:

• Which networks actually perform well in the places my devices or assets sit, indoors as well as outdoors?
• What happens to the connection when the strongest network there is congested or interrupted?
• Can the device reach more than one network, or is it tied to a single operator?
• How would I know if a device dropped offline, and how quickly would it recover?

Ofcom’s free Map Your Mobile tool, which has now passed a million visits, is a useful place to start for performance by location. Its performance view is averaged across a whole postcode district, though, so it indicates the likely picture rather than what a device will get at one fixed spot. From there, designing for both a usable signal and the resilience to keep using it is what turns good mobile connectivity into connectivity that critical applications can count on. For critical communications it is important to measure on site (see CSL’s signal analyser). If you are responsible for connected devices across multiple sites, CSL can help you assess where single-network dependency creates risk, and where our multi-network or resilient SIM connectivity could strengthen performance across your full estate.

• Signal struggles on trains widespread, Ofcom research finds (Ofcom, 3 June 2026)
• Connectivity you can count on (Ofcom discussion paper, 3 June 2026)
• Consumer experiences and understanding of mobile network quality (Jigsaw Research for Ofcom)
• Map Your Mobile (Ofcom coverage and performance checker)