Best Construction Accounting Software for Small Contractors and Growing Firms
Nitin Mahajan
—
September 17, 2026
Fill out this form, and we will get in touch with you.
The expensive mistakes in industrial IoT are rarely made at the device. They are made months earlier, in a procurement meeting, when somebody picks a radio technology and a connectivity contract for a fleet of assets that will still be in the ground in 2035.
Here is the number that should sit in the middle of that meeting. According to Ofcom's guidance on the UK 2G and 3G switch-off, 3G networks in the United Kingdom are now switched off, and all mobile network operators have confirmed to the Government that they will switch off 2G by 2033 at the latest. Two of them have published start dates in 2029, and a third has said: "during 2030". If you commissioned a couple of thousand 2G-only sensors across twelve industrial sites in 2024, that is not a technology story. That is a capital request, a site access plan, a subcontractor, and a per-visit cost you will be arguing about internally for two years.
I have spent enough time in plant rooms and depot yards to be suspicious of connectivity guides that start with a technology list. The technology should fall out of the deployment, not the other way around. So this piece is organized around what the estate actually has to do: move things, meter things, watch machinery, follow vehicles, and reach places the carriers never bothered to cover.
An IoT connectivity solution is the whole path between a field device and the application that consumes its data, plus the commercial and administrative machinery that keeps that path working across an estate. It is not simply a SIM card, and it is not simply a data plan.
In practice, a buyer is purchasing five things at once, and most disappointment comes from only evaluating one or two of them:
An enterprise buyer usually starts at the first item. Operations teams, in my experience, end up caring most about the fourth and fifth.
Start with the deployment where things travel: pallets, containers, trailers, rented plant, high-value cargo. The requirement is intermittent position and status reporting from a device with no mains power, which will be in a different regulatory jurisdiction next week.
Technology selection falls out of that description fairly quickly. You need wide-area coverage that is not tied to a single national operator, which points at cellular with roaming or multi-profile capability. Private low-power networks are a poor fit here for an obvious reason: you do not own gateways along the route. Narrowband technologies can work for simple location reporting, but their availability is uneven once the asset crosses a border, and a device that camps on a network with no NB-IoT roaming agreement is a device you have lost track of.
The failure mode I have seen most often is a tracker that performed well throughout a domestic pilot and then went quiet as soon as the asset crossed a border, because the pilot never tested attachment behavior on a foreign network. Test the border, not the warehouse.
The other thing worth saying plainly: for moving assets, battery budget and reporting frequency are the same decision. Every additional daily report is a shorter service interval, and service intervals on mobile assets are the most expensive maintenance you will ever schedule.
Now the opposite deployment. The device does not move, it sends very little data, and nobody will visit it again unless something breaks. Water meters, gas meters, pressure sensors in a basement plant room, environmental sensors on a distribution network.
This is the natural territory of low power wide area technology, so the definitions matter.
LPWAN, or low power wide area network, is a family of wireless technologies designed to send small amounts of data over long distances from devices that must run for years on a battery. The trade is deliberate: you accept low data rates and high latency in exchange for range, building penetration, and battery life.
NB-IoT, or Narrowband Internet of Things, is a licensed-spectrum LPWAN technology standardized by 3GPP. 3GPP completed the NB-IoT core specifications into Release 13 in June 2016, and the design uses narrow 180 kHz carriers with peak Layer 1 rates in the region of 226 kbps downlink and 250 kbps uplink, which is why it suits stationary meters in awkward physical locations rather than anything interactive.
LTE-M, also called eMTC or Cat-M1, is the other licensed-spectrum LPWA technology from the same 3GPP release. It narrows the device receive bandwidth to 1.4 MHz and offers roughly 1 Mbps in each direction, and unlike NB-IoT, it was built with mobility and voice in mind, so it is the sensible choice when the device moves between cells or needs a speech path.
LoRaWAN is an unlicensed-spectrum LPWAN protocol that an organization can deploy privately by installing its own gateways, which is why it shows up on campuses, in utilities, and across industrial sites where the operator wants to own the network. It has formal standards recognition: the ITU published Recommendation Y.4480 in November 2021, describing a low-power protocol for wide-area wireless networks that is technically equivalent to and compatible with the LoRaWAN link layer specification.
The practical decision inside this pattern is usually not NB-IoT against LTE-M in the abstract. It is whether the meter sits three floors underground (favoring NB-IoT and its link budget), whether the estate crosses markets where one of the two has patchy operator support, and whether you are willing to own gateways to get a private LoRaWAN network with no per-device operator charge and no operator dependency at all.
One myth worth retiring: the idea that low power always means cheap. The radio module is cheaper. The site survey, the gateway backhaul, the network server, and the integration work are not, and on a private LoRaWAN build they land entirely on you.
Inside a factory, a substation, a pumping station, or a treatment works, the requirements change again. Now you have process control systems that predate the internet, a safety case, and an operations team that is genuinely and correctly nervous about anything that touches the control network.
SCADA, meaning supervisory control and data acquisition, is the class of industrial control system used to monitor and control physical processes across a plant or a distributed network of sites, typically through programmable logic controllers and remote terminal units reporting to a central supervisory application. Connectivity for SCADA is a different problem from connectivity for telemetry, because a monitoring link that drops is an inconvenience while a control link that drops can be a shutdown.
Two things tend to be true on these sites. Wired or private wireless carries the control traffic, and cellular earns its place as a secondary path, a backhaul for outlying assets, or the link to sites too small to justify a leased line. Cellular is also, increasingly, the resilience story for sites where the fixed line has one physical route into the building.
A private APN, or private access point name, is a dedicated network gateway on the mobile operator's core that routes your SIM traffic straight into your own network instead of out to the public internet. Devices on a private APN are not reachable from the open internet, which removes a very large class of scanning and exposure risk before you have configured anything else.
That definition sounds administrative, and it is, but it is also the single control that most changes the security conversation with an industrial operations team. Add IP allocation from your own ranges, a site-to-site IPSec tunnel, and a lock that binds each subscriber identity to a specific device IMEI, and you have something an OT security reviewer can actually approve. Without it, you are asking them to accept devices with public addresses on a network they are accountable for.
Vehicle deployments break the low power assumptions completely. There is a battery in the vehicle, so power is not the constraint. Throughput is. A telematics unit reporting position and engine data is a modest consumer, but the moment you add video, the profile changes by two orders of magnitude, and dashcam uploads become the dominant line on the bill.
This is where 4G LTE and 5G belong. High throughput, low latency, mature handover between cells at road speed, and enough uplink to move recorded footage without keeping a driver waiting at the depot. Narrowband technologies are simply the wrong tool: they were designed to avoid handover, and a truck is nothing but handover.
Two practical notes from deployments that went badly:
Pipelines, offshore assets, mining sites, remote pump stations, environmental monitoring in places nobody lives. There is no terrestrial coverage, and there is not going to be any.
NTN, meaning non-terrestrial network, is 3GPP's term for using satellites and other airborne platforms as part of a standard mobile network, so that ordinary cellular devices can attach through a satellite rather than a ground mast. 3GPP states that Release 17, with its ASN.1 frozen in June 2022, was the first release with normative requirements for NTN, and the accompanying work on NB-IoT and eMTC over NTN was split across Release 17 and Release 18, with the radio performance requirements only completed in the later release.
That timeline is the honest answer to "is satellite IoT ready?". The standards are in place, commercial narrowband satellite services are live, and the technology has moved out of the proprietary-terminal era. What has not changed is the economics: satellite messaging is priced per message or per small bundle, not per gigabyte, so it suits low-frequency telemetry and emergency reporting rather than anything continuous.
The sensible pattern for most industrial estates is not satellite instead of cellular. It is cellular where coverage exists and satellite as the fallback for the handful of sites and the handful of failure conditions where it does not, ideally with one device that can do both.
TechnologyDeployment pattern it fitsMain trade-off4G LTE and 5GFleet, video telematics, mobile plant, site backhaul, SCADA secondary pathHighest throughput and lowest latency, but highest power draw and highest data cost per deviceLTE-M (Cat-M1)Moving assets, wearables, trackers, devices needing voice or handoverGood mobility and moderate rates, but availability varies by market and it uses more power than NB-IoTNB-IoTFixed meters, basement and buried sensors, static telemetryExcellent link budget and battery life, but low rates, high latency, and uneven international roamingLoRaWAN (private)Campus, utility, and industrial site sensor estates you want to ownNo operator dependency and no per-device carrier fee, but you buy, install, and run the gatewaysSatellite and NTNRemote sites, offshore, pipeline, coverage-gapCoverage anywhere, but message-based pricing and constrained payload sizes
Coverage claims are the least reliable part of any connectivity proposal, because the word is doing at least three jobs at once. A provider can mean regulatory permission to operate, a commercial roaming agreement, or actual measured signal at your site. Those are not the same thing, and only the third one keeps a device online.
A comparison that survives contact with reality looks at:
The myth here is that a global provider gives you a uniform service level everywhere. Nobody does, because underneath every global offer sits a stack of bilateral agreements with national operators, each with its own terms. The realistic ask is not uniformity. It is transparency about where the service is strong, where it is thin, and what happens in the thin places.
These two buyers get bundled together in vendor material, and they should not be. An enterprise IT buyer is optimizing for governance, integration, and procurement scale. An industrial buyer is optimizing for a device that has to survive on a wall for fifteen years.
Where they diverge:
Data sovereignty is the principle that data is subject to the laws of the country in which it is collected or stored, which in an IoT estate determines where the traffic breaks out of the mobile core, where the platform holds device records, and which regulator can compel access. It becomes a live design question the moment devices in one country route through a core in another, which is the default behavior for most international roaming arrangements.
If your legal team has an opinion about where traffic terminates, raise it before the SIMs are ordered. Retrofitting a local breakout arrangement across a deployed estate is possible, and it is not fun.
This is where enterprise deployments actually get difficult, and it is the section most technology guides skip.
The complication is regulatory, not technical. Some countries restrict permanent roaming, meaning a foreign SIM is not permitted to sit on a domestic network indefinitely. Enforcement varies from a firm rule to an operator quietly deregistering devices after a period, which is worse, because it fails silently and unevenly across your estate.
India has published one of the clearest positions on this. In its recommendations on the use of embedded SIMs for machine-to-machine communications, the Telecom Regulatory Authority of India recommended that all communication profiles on any M2M eSIM fitted in an imported device on international roaming in India should be mandatorily converted or reconfigured into profiles of Indian telecom service providers within six months from the date of activation, having reduced an earlier recommendation of three years. TRAI's own reasoning is worth noting: it argued that remote SIM provisioning makes six months sufficient, because an Indian operator profile can be installed over the air.
That is the mechanism the market has settled on, and there are two versions of it.
Multi-IMSI puts two or more operator identities on a single card, with an applet on the SIM selecting the appropriate profile and the matching regional rate plan. Remote provisioning on an eUICC goes further and installs an entirely new operator profile over the air after the device is deployed. Both exist for the same reason: to let a device that was manufactured and shipped as one part number behave like a local subscriber wherever it ends up.
That is a procurement decision as much as an engineering one. When you are buying M2M SIM cards for an estate that spans several countries, the real question is whether one supplier can give you a single contract and a single invoice while still respecting each market's rules. One of the providers that has built around this constraint is Trafalgar Wireless, which supplies multi-IMSI M2M SIM cards carrying two or more IMSI profiles on one card, with an embedded applet that selects the best profile and the local rate, alongside eSIM and eUICC options whose operating profile can be pushed over the air once the device is already installed, all administered through one carrier-agnostic console spanning 500+ networks in 180+ countries.
Which brings us to the consolidation question. Estates almost always start fragmented, with one operator contract per country picked up as the business expanded, and they almost always end up consolidated on a single management platform. The reasons are unglamorous, and they are the reasons that actually matter to an operations manager:
The point of consolidation is not that the platform is exciting. It is that the platform is the only place where a multi-country estate is legible as one thing.
Assume two providers can both deliver the coverage you need. What is left is operational, and it is where deployments are actually won and lost:
It depends on whether the sensor moves and how buried it is. NB-IoT is generally the better fit for stationary sensors in difficult physical locations, such as basements, ducts, and metal enclosures, because it was designed for link budget and battery life rather than speed. LTE-M is the better fit when the device changes cells, needs firmware updates of meaningful size, or requires a voice path. Availability by country is often the tiebreaker, since operator support for the two technologies differs from market to market.
When assets sit outside terrestrial coverage, and the reporting requirement is low volume, or when a site is critical enough that a coverage failure has to have a fallback. Satellite messaging is priced and engineered around small, infrequent payloads, so it is well matched to alarm states, position beacons, and periodic telemetry from remote infrastructure. It is a poor match for anything that streams. Most estates use it for a minority of sites rather than as the primary bearer.
Through a single connectivity management platform that holds every SIM and eSIM in one inventory regardless of which underlying operator carries the traffic. The platform provides lifecycle actions, usage policies, alerting, diagnostics, and an API, so that provisioning and troubleshooting follow one process everywhere instead of one process per operator contract.
Estates that skip this step usually discover the problem when they try to audit how many active connections they are paying for.