Originally published on LinkedIn in February 2021; reviewed and updated for this site.
Because IoT solutions are complex and made up of a number of connected parts, they demand more attention at the design stage. Not paying enough attention, and falling for the common mistakes, can delay or cripple a project.
As connected products become more common — with just about everything now available "connected" — ideas with poor value propositions are being weeded out. The high-value opportunities that remain tend to involve increasingly complex systems.
These are the fundamental questions worth answering before you develop one.
1. Connect, or not?
Technology adds cost to a traditionally unconnected product, and communications technology in particular brings both one-off and recurring cost. It is almost expected these days that a new product will be connected, or that an existing one will have a connectivity layer added, but it is important to understand the business case and the value. That layer adds cost to the product, possible monthly subscription costs, and an initial and continuing stream of expenditure on development and lifecycle support.
Without a clear logic of business value, the product will struggle to survive, and may join the ranks of good products that no longer exist. Solid research before starting, and an honest evaluation of that research, will tell you whether the system you are considering makes business sense.
2. Pick the right platform
When adding intelligence to a product that was not connected before, many startups select hobbyist-grade boards. The trouble is that developer platforms are not suitable for large-scale deployment. If the device proves successful and generates serious demand, production cannot scale, because you cannot source thousands of that type of board.
Off-the-shelf platforms are useful for proof of concept and as platforms for software developers — but do not confuse a proof-of-concept system with one that is production-ready. Many hobbyist platforms do not have the certifications a real product needs, and as any experienced hardware developer will tell you, a development system is not a high-volume extensible platform. Source only components and modules that will be available in the volumes you anticipate, appropriately costed, now and in the future.
3. Design with regulations in mind; do not add them afterwards
Regulatory testing is another important part of any IoT design effort, and the requirements and certifications must be factored into the design. Because they are connected, IoT products must be tested for radiated emissions and susceptibility. If they plug into an outlet, conducted emissions and susceptibility come into play too. In some countries the cellular carriers must also perform testing to certify your product for deployment on their infrastructure. Depending on how you implement the cellular technology, that step could take months and be very costly — or indeed impossible.
Selecting pre-certified components can drastically cut the time and expense, but do not count on it entirely. Pre-certified parts are more expensive, and they radically reduce the headaches involved in getting certification later.
Lastly, where and when do you plan to sell? It is not practical, particularly for a startup, to expect worldwide sales of fully certified devices on the same calendar date. While most safety, communications and cellular standards are similar, pick your target countries carefully and think about international rollouts over a period of time. Many countries are grouped under a common set of requirements, which helps — but only if you have planned for it.
4. Design for security at the start
Security needs to be built into the design process, not added as an afterthought. It is a must-have, not a nice-to-have. The number of connected devices is astounding and increasing rapidly; the potential for a breach is enormous and the results can be devastating. Attackers scan for poor or misconfigured security.
Consider end-to-end security mechanisms, end-to-end data encryption, access and authorisation control, and activity auditing. A security chain is only as strong as its weakest link, and low-end, poorly protected endpoints are a frequent point of entry when they are not carefully and intentionally secured.
5. Get the right product development team
Many startups spend a lot of time researching the contract manufacturer who will assemble the product, and not enough thinking about their design team. Sometimes a contract manufacturer will offer their own design engineers at very low or subsidised cost — occasionally in order to lock in manufacturing rights. Unless you are a tech giant placing initial orders for hundreds of thousands of units, do not expect a high-quality design and engineering team to be responding to your needs on that basis.
Nothing burns through a budget, and time, as quickly as redesigning a product that does not meet the required functionality, uses the wrong or outdated technology, or gets deployed with poor quality. Problems like that can destroy a company.
6. Design for the user
The person using the device directs its activity and controls its acceptance, connected or not. Sometimes there is no Wi-Fi, so it makes sense for a product to have some functionality even when it cannot reach other systems. By making the device usable when disconnected you create a product where the user stays in control and remains productive even in areas of poor connectivity. That matters to customers who are uneasy about the connectivity inherent in IoT products.
Whatever you do, do not believe the cellular operator's hype. Connectivity will be an issue and cannot be guaranteed.
In summary
The progress of complex connected devices is exciting, but the basics still apply. You need a sound business case, as with any investment. Solid project management matters as much as avoiding the mistakes above when taking a leading-edge device from inception to the manufacturing floor. Selecting engineers with domain knowledge, technical skill in the key technologies and good communication skills is critical. And staying within budget and meeting deadlines is what turns a plan into a product.