Ultra Low cost IoT, a breakthrough to innovation – disk91.com – technology blogdisk91.com – technology blog
Ten month ago, I was starting investigating the Ultra Low Cost Iot thanks to Sigfox given access to the experimentation they made in this area. The Ultra Low Cost IoT is defined by manufactured devices under or near the cost of $1.
Such level of price are enabling scalable deployment of devices and mass data capture, more than it has ever been able to made. And basically, it is not a question of device cost, it is a question of scalable use-cases where the higher device cost where a business model killer.
This is basically reducing the price of devices by a factor of 5x to 10x. No miracle, this is also reducing the device capabilities and concerns different use-cases than the one previously designed. In another word, Low Cost IoT is not classical IoT for a lower price, it’s a new IoT area where we find new use-cases, new business model and more generally is source of innovation.
At scale IoT
Before we talk about IoT, we had the M2M technologies, I’m characterizing M2M as based on technology needing a wired connectivity of having a short autonomy to support the communications. They are based on communication layers like, Ethernet, WiFi or 2G/3G/4G for the mobile version of it. To understand what has changed, let’s take the goods tracking example.
With the M2M technologies, we are tracking, since about 20 years, goods by installing tracking solution in trucks. The truck is a source of power and even for a price around 100€ you can track all what is transported by that truck. In addition the installation cost inside the truck are expensive. In theory this could be good enough. But in reality making the link between the trucks and the good is a complicated thing as the number of intermediate company and information system is large. During a single good transport you can have many operator involved and goods being at rest in multiple places.
With the rise of IoT, we have seen many use cases looking to track not the trucks but the goods directly to solve this system integration issue. PSA, Airbus, Michelin, most of the large companies have created their own solution for, in many cases, the same use-case. It would be interesting to understand why these big company with no real experience in IoT design have made X times the same hardware design but basically we can consider the fact there was a great business model making this quickly even on their own. In these IoT use cases, these company has decided not to track the truck, not to track each of the individual good but a group of goods within a single transport pallet or a sea container. This has became possible thanks to the two level of the IoT:
As a consequence IoT, for a TCO near to the M2M solutions, allows to change from a truck level tracking with a large and impossible integration complexity to a pallet level tracking with no integration challenge.
At scale IoT is the new revolution: thank to a device manufacturing cost around $1, it’s now possible to move from the pallet level tracking to the individual good tracking. This way, the traceability is never broken by any logistic operation like a transportation pallet change and the integration challenged is killed. Pallet tracking needs to know in what pallet a good is. This is a simpler but still existing integration challenge and this challenge can be high when the logistic is not integrated seamless. So tracking unitary goods is the target for any tracking solution and the previous way to do it where only compromise found to solve the TCO considerations. At scale IoT is the key point to start unitary good tracking.
Technology limitation in 2020
2020 is just the beginning of the At Scale IoT, we are like the first phone able to send emails before the rise of the 4G. IoT at scale is IoT inside each of the objects around us, each of the objects we own, each of the object we are using.
In 2020, the technology is still limited. For people looking for technical details, I’ve made deep dive post on the Ultra Low Cost technology proposed by Sigfox. For all of us, I propose this time to look at it in a business point of view in the coming chapters. Before, you may understand all what we can’t do with a such technology in 2020. As the technology is running on really low cost MCU (aka computers) these solution are only able to send 3 types of messages: (no data, 1 bit equal to 0 or 1 bit equal to 1). As an example you can’t report a temperature but you can report an over-temperature reached.
As I detailed tracking use-cases, Ultra Low Cost IoT is not currently reporting a GPS position but it can be positioned by the network. in certain condition, a precision under 100m can be obtained.
An ultra low power device can’t receive downlink (message sent by the network to it)
But Ultra Low power design can be alive for 1 year, potentially more, reporting 1 information on regular basis or detecting a specific event it has been designed for.
Is IoT at Scale a dream or reality ?
Technologies announced by network operator or detailed by a blogger may not be trustable technologies, some are announced years before being a commercial product.
IoT at Scale and Ultra Low Cost design is not part of these announcement, the technology is directly accessible with the existing components. Basically it has been built on quite old radio devices used for remote garage door controller. There is some engineering to be done as I personally spent a part of the last 6 month to design and industrialize a final product based on it. But it works well.
So, as a proof of market accessibility for IoT at Scale technologies, you can take a look to the first certified Ultra Low Cost design based product Sticktrackr I’ve designed, I’ve certified Sigfox READY and we are now mass producing at IngeniousThings.
Innovation in use-cases
Since the last 10 month I’m working on the device engineering, we also at IngeniousThings, worked on possible use-cases and associated business model for this technologies area. For certain of us, we are currently running POC with industrial companies, for other, we have created them and we are looking for industrial companies to experiment them on the field.
I’m going to review some of the use cases the IoT at Scale is now enabling.
Tracking goods
It’s not a surprise to come back on good tracking with Low Cost IoT. I’ll not detailed again why it makes sense as it has been made in the previous chapters. With Sticktrackr, we are able to track a single good, a letter, a parcel. We are able to detect when it has been opened. We can raise an a alarm on such event. A such device can be used in many use case and the more basic one is tracking good. The customer promise is to see the good moving, being able to locate it with a precision of a few kilometers and being aware of good opening. Depends on the reusability of the device on the precise use-cases, the price of one tracking can be not much more than a passive solution like the bar-code the postal operators are proposing.
Production planning optimization based on supply chain monitoring
Now, let’s imagine you put the device, not on a letter, parcel.. but inside a good a company is producing. What does it change ?
As a manufacturer, you need to pilot your production based on the downstream supply-chain state: the level of your stock, the level of your distributor stocks, the level of the consumer distribution stocks. Any one, you and your distributors, are providing a forecast signal… usually wrong. Companies are purchasing raw materials, organizing factory planning based on these forecasts.
At Scale Iot, is a breakthrough: by directly reporting information of location of any of the produced good up to 1 year after being produced, it allows to correct forecast by a data you directly get from the reality of the field. Are my produced goods stocked by the distributors ? or have they been bought by the final consumers ? Where are my consumers located ? in what area the sells are growing faster than in another … All these question now have simple answer, quantified, verified…
A such IoT sensor is fully autonomous, it can be included in any good, even a wood made table. IoT at Scale is not for high technology good, it address low tech goods, commodities.
Goods inventory
It has been a surprise for us, but the main use-case we have addressed up-to-now, is automated goods inventory. The question is to count goods in different storage places and to have a traceability of the movement of these goods between places.
Even is the network positioning, in regular basis, is a few kilometers, we have been able to improve this positioning to hundreds of meter thanks to the ability to extend the Sigfox network. This precision is good enough to such use-cases.
This technology is relevant even in an area where many technologies exists because this one doesn’t requires costly infrastructure, doesn’t requires costly manual operations and also works outside of the company taking benefit of the global Sigfox network.
Call to action
IoT in marketing is powerful, we have recently seen experiment coming from a car manufacturer to make call to action: you receive a postal advertisement with a button to press to get a trial. It seams that this experience has been a success even with an expensive device.
Not everyone sells cars, but with a low cost solution, you can extend this use cases to many more goods.
At IngeniousThings, we experimented to send letters to some of our customers and being able to immediately place a call when they opened the letters.
Hi, you just opened our brochure about our new product, we would like to know if you would get some more details about it…
Even a small company like the one we made is able to access a such advanced marketing campaign thanks to a very low unitary cost.
Amazon dash button is dead, long live to dash button
Amazon has a great commercial idea: offering a button to reorder your favorite product. After 3-4 years this service has been stopped. As part of the reason, the technical complexity, it worked with a smartphone, has to be connected to it. It’s hardware design was also a bit costly due to the complexity of the technology involved and the low number of devices produced/used.
Thanks to IoT at Scale, the use-case can now been redesign, improved, expended. Imagine a device inside your laundry, directly reporting a message to reorder it once you finished the first 50%. Seamless, with not need to be connected to anything else other than the global network around you. It is not a dream, it is a use case ready to me deployed today.
You want to let the user reorder something manually with a single clic, just do it, even your grand-grand-ma will be able to use it.
Classical use-cases
The previous use-cases looks like the most important to me because they are creating innovation, there was no way to make them at scale with the previous technologies.
They are not the only things we can make with Ultra Low Cost IoT, there are also things already existing we can make at a larger scale thanks to a lower device price:
There is no more limit to that list than your imagination. IoT at Scale is not many different than IoT, all existing use-cases apply to IoT at Scale with the scale differentiation.
Forest are already protected by IoT but price is limiting the monitoring to 1 hectare. Connected body thermometers already exists but for the price of 60€ only rich country have a couple of them deployed.
IoT at Scale is leveraging the business model thanks to a fine grained information and the scale effect.
Let’s conclude by the business models
Business models defines costs, benefits and money transfer between the different actors of the use cases. Business models depends on use-cases and detailing them would be a bit long for this post.
So I’ll only address the cost model currently existing on the IoT at Scale and point out what is currently limiting the expansion of this rising IoT.
Cost model is composed of 3 different layers:
To understand this, let’s consider the overall cost for creating an IoT device, you have design engineering, certifications… this is an investment around 100K€ you need to spread on each of the devices. If you make 10K devices, this is $10 par devices, so you $1 manufactured device will cost for real $11 each.
Now, if you want to design something generic with different sensors, different powering solutions … the cost of these extra components will double or triple the device cost. At this level of price any change, any de-optimization will have impact over 50% on the device price.
In my IoT experience, the only way to optimize the TCO is to make dedicated solution, hardware + software, engineered, maintained by one entity with the right level of expertise. It is basically the concept of SaaS. Most of the companies have taken that point about classical software (Salesforce, Office365…). This is particularly valid for IoT where the integration between hardware, telecom, infrastructure and software is a highly complex area where experience is needed and currently rare.
Low Cost IoT also means scale factor: just consider I’ll sell 1000 devices at $1. With a margin of 30%… that’s 300€. That’s usually less than the cost of the time spent to explain our customers how to solve their use-case with our solution and get an order from them. As the value is not in the hardware (In my point of view, as a reminder, it still true on regular IoT) it is on the associated service. And the right value is based on a shared business model and created value. What I mean is, there is no defined value for the service, it really depends of each use case. The service cost variability related to the scale is also really large as the software investment are usually higher than the hardware one (and usually under estimated).
Basically to make it simple, the service cost can go down to 0.3€ per device per year in condition of high dedicated or shared volumes. But is can also be a fixed price up to 100K€/year for low volume of devices depends on the degree of customization.
Basically the Sigfox offering is 2, 50, 100… 140 messages per day with 12 byte of data each. They are payed for a full year with a price from 6€ to 14€ (in France) You can have better deal over 1K subscription.
An Ultra Low Cost device is usually sending from 1 to 12 messages per days regarding the use cases. Due to the battery size related to the cost it make no sense to have more messages. Basically it means 2 messages is usually a bit low and 50 messages is always too much.
The 6€ to 9€ HT are about 3 to 5 times the price of the hardware. The solution total cost per device quickly reach the 10€-15€. At this level of price using a higher hardware design cost will only impact the TCO about 20 to 30%. It basically killing the innovation.
The use cases also have really different life cycles, some have a 1 year life, some other are only limited to 15 days. But the network connectivity is to be paid for a full year. We have made some engineering to solve these issues but some use-case are simply impossible to realize today due to the global year subscription and this impossibility to anticipate when the device will start to transmit it’s first message.
As a consequence, currently, in the business model related to the Ultra Low Cost design you have three tiers involved in the costing discussion: The customer, the service provider and the network operator. In my point of view, as a defender of the network neutrality, this is a problem.
Sigfox has announced 3 years ago Admiral Ivory, it was supposed to address such use-case, but up-to-now, I’ve not been able to get information and understanding about this offer. It think it was supposed to be a per message cost which makes sense for such use cases.
Considering 50 messages / day cost 9€ per year, it means 0,0005€ per messages. This kind of cost per message in an Ultra Low Cost business model could really make sense to unlock the innovation and making the rise of the IoT at Scale.
In terms of business model for Sigfox, as the number of Ultra Low Cost device is more scalable than standard devices due to the lower radio power it would also be a great deal at scale.
With the current pricing, we have the following situation:
This situation makes in this market a certain instability and currently slowing down the rise of the IoT at Scale. But in another hands, it’s also not a surprise to see the commercial offers being made after the technological rise because marketing needs to understand the use cases and business model applicable to the new technologies before being able to propose something.
I really hope Sigfox to quickly propose a subscription offering as innovative as this technology potential to create the breakthrough needed to make the IoT at Scale to come alive.