The Secret Driving the Internet of Things' Growth: Sensor Costs Declining 200X

How do you stimulate demand for a technology about whose utility and functionality there are still many questions?

Well it can't hurt to make it affordable and when the price of a - possibly the key component has dropped 200 times that makes for a compelling sales proposition.

For all of the chatter about managing you life via interconnected devices, the now cliched 'internet of things,' many consumers wonder about why they need it, what advantages it will deliver and what benefits it will bring to their lives - and what it will cost.

It turns out that the rapidly declining costs of the sensor make it all work are giving developers more time to experiment and therefore producing results that will be more interesting to consumers. JL

Thomas Frey reports in Impact Lab:

There will be challenges converting the volume of data into consistent and useful results and to creating a machine that can learn and gets more intelligent with time. But dropping the price of the sensor by 200X and making it available to a broad group of developers is a start.
Rapidly declining costs of sensors is one of the major factors in the growth of the Internet of Things (IoT) —the networking of the physical world within existing Internet infrastructure.
Sensors are critical to IoT. Consider a connected thermostat: Without motion, humidity and temperature sensors, there is no data that algorithms can use to set points tailored to a user’s behavior.
In some cases sensor costs have declined by as much as 100X over the past decade. One of those cases where a startup is attempting to drastically change the economics of sensors is in the area of near-infrared (NIR) spectroscopy. If that phrase sounds familiar, it may be because of all the attention SCiO is getting.
Designed and assembled in Israel by startup Consumer Physics, SCiO is an NIR spectrometer for consumers, set to roll out by summer 2015 for $249 per unit. NIR spectrometry detects the spectrum created from shining a light source at a given sample. That light spectrum—a so-called molecular signature—can be used to identify matter.
Typically, NIR spectrometers are found in university research laboratories and can run as much as $50,000. But Consumer Physics sourced hundreds of cheaper components globally and traded some sensitivity and accuracy of the sensor in order to get it to a consumer price point.
While we don’t know the full possibilities for the SCiO, initial applications center around plant hydration, pill analysis, and food. The SCiO can be pointed at a food item in order to notify a user of the calories as well as the protein, carbohydrate, and fat content. (SCiO can actually distinguish Pepsi from Coke.)
I spent some time speaking with Dror Sharon, the co-founder of Consumer Physics. While Sharon is focused on making sure that the initial applications—food, pills, agriculture—are working well and creating a solid user experience, he noted in a moment of candor, “Honestly, just making something a 100 to a 1000X cheaper is pretty cool in itself.”
As we gear up to guide a new IoT focused channel over at Gigaom Research, innovations in sensors are important because as more complex data becomes available, the potential to produce creative applications for that data grows. For SCiO (and other up-and-coming companies), that means paying a lot of attention to its developer community and its SDK because the more engineers that take its data and use it to build promising applications for that data, the better for Consumer Physics.
Creating a warehouse of data on matter and the physical world is a massive project. It also requires developers with expertise beyond coding. With that in mind, Consumer Physics has released an expert-level SDK that allows developers to download spectrum data without noise cancellation in the hopes that they will develop their own algorithms and tailor the data coming out of a SCiO to new and specific applications. Sharon says that almost daily the company is fielding requests for the hardware to be implemented in diverse applications. One smart-home request came from a blender-maker that wants to integrate the sensor so it could provide the nutritional content of a morning shake.
On the horizon, Sharon and I talked about what other-next generation sensors might either significantly decline in cost or be accessible to consumers in the next 5 to 10 years. We discussed ideas ranging from infrared camera sensors to 3D motion sensing to “digital noses,” or sensors capable of analyzing the air to high degrees of sensitivity.
For now, SCiO is an incremental step towards making a research-level sensor available to both consumers and hardware developers. There will be challenges converting the volume of data into consistent and useful results and to effectively creating a machine that can learn and gets more intelligent with time. But dropping the price of the sensor by 200X and making it available to a broad group of developers is a start.