Let me start by telling you, the $1 to $20 sensor COULD happen if produced in quantities of 100,000 units with limited sensing requirements (very basic sensor) and short-range radio.
Unfortunately, at this time the IoT industry is a ways away from achieving the $1 sensor.
Time and time again we have heard from clients that they thought the $1 sensor was their anticipated price, and at best the sensor would be $20 dollars each. Hence, this article was written to explain what goes into building a sensor and why we have a lot more work to do before achieving this.
What goes into building a sensor?
Step 1: (Creating a sensor)
Once a company decides to build a sensor, they must first start with building the team. This is a significant step and investment as product development cycles are anywhere from 12 to 24 months before a product is viable for resale. Industrial, mechanical, electrical, and firmware engineers have to be employed for that period and paid before work can begin on that sensor. Not to mention the manufacturing and supply chain issues that come with high volume electronic products. These upfront costs are required before you can start to develop a comprehensive sensor that encompasses all the requirements for your intended market. At the minimum, you can expect to pay $200,000 towards the development of your first sensor.
Step 2: (Injection mold)
The second step is the injection mold, which is a capital-intensive investment depending on the volume of sensors you intend to manufacture. A plastic injection mold can cost between $5,000 to $10,000 and will yield up to 3000 to 5000 units. An injection mold built of metal will cost between $10,000 to $20,000 but it will help you produce north of 30,000 units. The key here is, can you sell enough sensors to recoup your investment?
Step 3: (Choosing the sensing capabilities)
Once a company has agreed on a product design that will be suitable for the marketplace, they must identify what they want to be sensing along with each additional element they wish to measure. Any add-ons will cost more money. For the purpose of this example let’s use our Water Sniffer/Water Sensor. After months of market research, we identified that we want the following from our sensor:
Water detection (This is the primary value of the sensor yet again we needed both the flood contacts and the rope sensing to ensure it would fit all business cases)
Tamper detection (This is as important to identify theft and vandalism which is more common than you think)
Temperature sensing (This was important for clients worried about frozen pipes)
A power button (This was important to simplify the installation process)
A buzzer (This was important to us as a second line of defense to the dashboard and also to ensure we knew if the sensor was “live” once we pressed the power button)
Step 4: (Finding the right network for the product and/or solution)
Once a company has identified what sensing capabilities they want to achieve and what design they like, they will need to consider how the sensor will communicate with their application/dashboard. In this area, there is a range of chipsets and networks with varying limitations. For the purpose of this exercise let’s combine chipsets with networks and assume each network only has one kind of chip. For the record, many other networks exist but this paints a decent picture for the audience.
Bluetooth: Amongst the most cost-effective chip
Sigfox: A great option for certain long-range applications where the network provider manages the network cost for an annual fee. This chipset has a cost double the price of the Bluetooth chip.
LoRa: Another great option depending on its application where the solutions provider must manage his or her own network. This chipset comes in at approximately 3 times the cost of the Bluetooth chip but offers more versatility.
Behrtech: A new up-and-coming option that has many benefits for solutions providers in the proptech business who may not understand networks. The price point here is closer to 5 times the cost of a Bluetooth chip.
This is another important milestone a company needs to surpass as they build their sensor. Many nodes are not designed to be able to change chipsets after the company has landed on a specific chip. This further limits the company’s market segment/opportunity.
In our case, we’ve chosen to spend a little more in step 1 to ensure we can be agnostic to the network our client wishes to work with however, our most common network and chipset is LoRa.
Stay tuned for another article about network comparisons!
Step 5: (Choosing how long you want the batteries to last and how many you will need)
At this stage, the company must choose if they believe the client will see value in evaluating the total life cycle of the product or if they will air towards the most cost-effective product.
In many cases, companies will compromise on the power to help reduce costs for the client. However, in our opinion, this is the worst decision. The best products will have the most amount of energy to last as long as humanly possible. This is another additional expense for those who choose to view things the same way we do.
Simultaneously, the company will have to look at offering its sensor with or without batteries. At a minimum, a decent 1.5V battery costs approximately $2 each so you can do the math.
In our case, we’ve opted for three 3.6V batteries with a cost of approximately $3 each as we see this as a game-changer for product versatility and for the life expectancy of the sensor.
Step 6: (The PCB)
We cannot forget the board that connects the sensors in Step 2 and the batteries identified in Step 4. This is effectively the brain of the sensor and the most capital-intensive piece of the sensor.
Step 7: (Certifications)
After the node has been assembled and has been tested, the certifications come in. The water sensor then needs to be certified FCC & IC if you want to sell it across North America. The total investment for the certification is approximately $10,000.
After the hardware development cost comes:
Minimum inventory requirements and the costs to hold this inventory.
Cost of firmware development and software development where both are vital to the deployment of a comprehensive solution.
Manufacturing and assembly costs.
As you can tell by now, the capital investment and the recurring inventory cost of producing a sensor is quite significant. Thus, it will require immense market momentum and substantial purchase orders to help drive the market towards the $1 sensor, or even to the to $20 sensor range.