Four pillars for industrial-grade IoT: a solid foundation

In industrial IoT, a solution never fails all at once. It fails quietly, at the weakest point in the chain. And once that happens, trust is lost — often permanently. We often compare this to streaming a movie. If the battery of your tablet is low, the Wi-Fi is unstable, the app is slow, or the sound is distorted, you stop watching. In industrial IoT, the same principle applies — except the system must work continuously for seven years or more, without recharging, rebooting, or changing providers.

Over the years, we have learned that robustness is never accidental. It is designed, validated, and proven in the field. Batteries are a discipline on their own. They must last for years without intervention, under varying temperatures and loads. Small deviations in battery quality can have a massive impact on ROI at scale. From our perspective, selling trackers without pre-selected, pre-installed batteries matched to the use case is simply irresponsible.

Rugged casing is another example. IP and IK ratings are useful, but they are only a starting point. Real life does not happen in a lab. Are the seals integrated in the molds? Is the hardware assembled with controlled torque? Are flanges, brackets, foams, and so on, designed for long-term shock and outdoor life exposure? Is the product designed and manufactured by experts?

Connectivity is often oversimplified. ‘Global low-power roaming’ sounds reassuring, until uncontrolled network behavior drains batteries (yes, this happens), creates data gaps, or leads to unexpected communication costs. Connectivity must be validated in the field and governed over time, not assumed. You cannot just trust the catchy flyers of virtual network operators without proof of local operations on the actual operator networks. Make sure you ask for proof from the field, as the operational cost of missing this check can be significant.

Security completes the picture. Digitizing supply chains inevitably introduces new attack risks. If security is not designed end-to-end, vulnerabilities emerge — from cloneable tags to unwanted data leaks. False positives in logistics are not theoretical risks; they disrupt real operations.

Finally, data reliability. Using unreliable data to change processes is worse than not using data at all. Compare it as understanding climate impact by measuring the evolution of ice formation, from a satellite flying at 30.000 km/h. To get reliable insights from the field, you do require validation, pre-processing, post-processing, metadata, and consistency over time. No pilot should become operational without this foundation.

If there is one word that is abused most in IoT, it is scalability. After twelve years in the field, I have learned that scalability is not a design claim — it is an operational proof. Many solutions work perfectly in pilots. That is not surprising. Pilots are controlled environments: limited assets, motivated teams, extra attention from vendors. Reality starts when you deploy hundreds or thousands of assets, often executed by people whose job is logistics — not technology.

True scalability starts before installation even begins. Devices must arrive factory-ready: batteries installed, connectivity configured, firmware validated. Every manual step introduced at scale multiplies cost, error risk, and frustration. Installation itself is rarely just ‘mounting hardware’. It includes activation, confirmation, visual checks, configuration validation, and the certainty that the asset can leave the yard and remain unattended for years.

We learned this by facing these challenges early on — and then redesigning the process until installation became predictable, fast, and repeatable. Scaling also forces you to think about what happens later. You need to keep the fleet alive with minimal operational intervention for many years. At scale, even rare events happen every day. Devices will fail, batteries will age, and assets will be damaged.

If swaps are not anticipated and managed properly, operational costs quickly outweigh the original ROI. Finally, scalability does not stop at devices. Cloud systems are stress-tested when hundreds of thousands of assets send data asynchronously. Latency, congestion, and inconsistency show up quickly. Adding more servers may hide the problem temporarily, but it never fixes the underlying design, and it always translates into higher costs for you or just a bad user experienc,e which kills all adoption.

This pillar often raises eyebrows, but our customers insist on its importance. Industrial IoT is not a black box that can be handed over and forgotten. When things work, architecture matters. When things go wrong, accountability matters even more. We have seen situations where customers were sent from hardware supplier to network provider to application vendor — each owning a piece, none owning the outcome. That model might be acceptable in consumer technology. It fails in industrial operations.

Supply chains depend on continuity. When asset data drives planning, execution, or customer commitments, there must be one party accountable for restoring service — not explaining why it failed. There is another layer to this: longevity. IoT solutions live for many years. Teams change, priorities shift, and integrations evolve. Without documentation, training, shared learning, and experienced people, knowledge quietly disappears, and operational risk grows.

After twelve years, I am convinced that ‘someone picking up the phone’ is not about support friendliness. It is about operational maturity and ownership. You need to become part of an ecosystem of peers, including trained experts and other users sharing their experiences and daily tips and tricks, from installation to ROI calculations.

The final pillar determines whether value actually materializes. A solution that is technically excellent but not used has zero value. Adoption is not about convincing users. It is about removing friction until using the system feels natural. After decades in technology, I’ve seen how often “easy to use” becomes an empty promise. Most software is intuitive in decks, not in daily operations. That’s precisely why we made simplicity a design principle, not an afterthought. Our approach is validated by evidence: operators consistently describe the platform as “best-in-market and genuinely intuitive to use.” We’ve documented this in practice — including case studies showing how teams without specialized IT skills operate the system independently.

In addition, control plays a subtle but critical role. Organizations want to own their operational tooling. They want to configure, integrate, and evolve their solution without depending on the vendor or expensive service contracts. They want to provide their team with control over the tooling to become real domain “heroes”, focused on change and adoption. Integration is often where adoption is won or lost. Introducing yet another screen rarely changes behavior. Embedding reliable data into existing systems — TMS, ERP, BI — does. Make sure you check these boxes when you invest in an IoT solution.

Finally, AI reinforces this lesson. When data is clean and trusted, AI accelerates insight. When data is unreliable, AI simply scales confusion faster. Make sure you question well the AI capabilities and the quality of the AI integrations. Designing for adoption means respecting how operations actually work — and accepting that simplicity is the hardest feature to build.

Looking back at twelve years of industrial IoT, everything converges on one word: trust. Trust that the system works tomorrow. Trust that it scales next year. Trust that operations can depend on it every single day. We do not sell pure one-off hardware or disconnected software. We deliver an end-to-end solution, supported by a long-term subscription model aligned with our customers’ success. Our ambition is simple: to be the trusted long-term partner for industrial companies where process and supply chain visibility is critical for growth and competitive advantage. In industrial operations, reliability and predictability win. We are proud to build systems that deliver exactly that.

About the author, Kristoff Van Rattinghe

Kristoff is CEO and co-founder of Sensolus. Since its foundation, Kristoff has built up in-depth knowledge about transport and logistics chains. He understands better than anyone how the Internet of Things offers a solution to various challenges in the industry and how companies can ultimately save (a lot of) money.

Before Sensolus, Kristoff has built up a track record in setting up new business units, building project teams, and successfully translating ideas into products in the ICT sector. In addition, Kristoff, as a board and unit manager, has provided important added value in the marketing of software applications for renewable energy.