Only two apps, Facebook and YouTube, occupy more of our time than our need to know our location and proximity to others. As humans, our need to connect socially and our need to be entertained are dominant instincts. But location – and its relative cousin proximity – has been one of the most important components of information exchange since the discovery of the first fruit tree (“Eve, where did you get that apple?”). Latitude and longitude have been guiding travelers since the 3rd century BC. Look at phone numbers – +1-xyz-555-1234 – the first three fields are dedicated to location. The US and Soviet Union each put up large constellations of satellites so that we can know where we are in the world at all times. A recent survey found that 83 percent of smart phone users say that location services are critical to their app experience – not to mention the value of user location to ad networks.
But the Internet of Things (IoT) is everywhere – it’s ubiquitous. If we have 50 billion connected devices and trillions of sensors we will be awash in both things and streams of data. We talk about big data and the cloud in which it lives. Although the cloud is geographically undefined, even cloud location has importance at times; just ask anyone working on a digital health application in the EU. But even if we do not care where the data we pull from the cloud is stored, that data has less value if we don’t know the location of its origin – its source in the physical world.
Consider one of the earliest and biggest markets in the M2M/IoT space – fleet management. The fleet management market, over $15B a year and growing at nearly 25% year over year, was originally based on the simple concept of managing of the location of fleet assets. Although additional vehicle health and driver behavior sensors now provide a wealth of data on the mobile assets, GPS (Global Positioning System) is still the center of the universe; pardon my Copernican pun, for fleet applications. Fleet management is the IoT poster child for location creates value for data streams.
I was recently introduced to a fleet management platform company called Geotab. Geotab is interesting to me for two reasons. First, its hardware premise is based on an after-market device to retrofit vehicles with a cellular transmitter for the data from those vehicles – including geolocation. This type of device is not new. I personally tried out such a device to test the value proposition of knowing more about my car and behavior change via the Hawthorne Effect per the Progressive “rate sucker” commercials. But after-market and retrofit approaches are critical to the Industrial IoT because so many assets, Intel estimates more than 85%, are not yet connected and do not generate data. Second, and this is what really caught my attention with Geotab, is the application Marketplace they have facilitated from vehicle data streams. I am a big believer in platform business models and was excited to see this developing in an IoT space. I counted nearly 50 applications using the data streaming from the Geotab OBD (On Board Diagnostics) device and most of those apps leveraged the identity of the vehicle and its location.
I get excited about this because my thesis for the Internet of Things is that every “thing” needs to know where-it-is, who-it-is, and how-it-feels. I did not say that we need to know where every “thing” is; we need the “things” to know where they are. They need to know who they are and where they are so that they can include this critical information in every data stream they generate. The value of that data stream is then immediate because it is authenticated and situationally located.
Think about the data that you use in your everyday life. When was the last time you looked at a data set that you did not associate with a location? Virtual machine performance – you know the exact location of the server upon which that machine lives. Network performance – you know the identity and location of every router and switch. Patient data – you associate it with both the person and the geographic location of the person for action. Manufacturing data – you know the location of every machine from which every data set came. The value of almost all information is increased by the knowledge of the location of its source.
I tested my premise by searching out leading application of big data. The applications of big data are already as diverse as IoT applications, but I did find an interesting list of emerging big data applications compiled by a data scientist. I summarized the applications and the analysis value delivered:
|Big Data Application
|Big Data Billboards
Define and justify its pricing model for advertising space … using sophisticated GPS, eye-tracking software, and analysis of traffic patterns.
Collect data and input from users phones to compile data for health studies.
|Big Data and Foraging
Municipal tree inventories, foraging maps and street tree databases to provide an interactive map to tell you where the apple and cherry trees in your neighborhood might be dropping fruit.
|Big Data on the Slopes
Help ski resorts understand traffic patterns, which lifts and runs are most popular at which times of day, and even help track the movements of an individual skier if he were to become lost.
|Big Data Weather Forecasting
Taps into sensors already built into Android phones to crowdsource real time weather data … fed into predictive models.
|Yelp Hipster Watch
Map then plots the locations for the reviews in red. The darker the red, the higher the concentration of that word … and ironic tee shirts and handlebar mustaches.
|Even Big Data Bras?
Using big data to help women find better fitting bras.
In all seven applications the location of the source of data is at the root of the value derived by the analytics.
But here’s the rub. I was talking to Matt Neegard of cloud platform company Exosite the other day. I asked Matt what percentage of the devices connect to their cloud presently have known locations. “All of them,” Matt said. “What percentage of those devices knows where they are?” I asked. Matt thought for a minute — “About 5%.” I suspect that for connected things-without-wheels this percentage probably holds across most applications. But what happens to the value of the data streams from these devices when someone moves them? What happens when the business that owns the devices changes ownership? What if, heaven forbid, the location look up table is lost during the transition of the IT systems? How do you direct action from data analytics if you don’t know the geographic veracity of the source?
The fleet management industry has been an early benefactor of the IoT because the things in its focus can move – they all have wheels – and knowing the location of those things had obvious value. The market and industry evolved to the point where almost all vehicles have a standard “who-am-I-how-am-I” data port – OBD. But what about all the things-without-wheels? How do we trust the data streams from these devices if they cannot tell us where they are or when they get moved? Many are beginning to address this problem with technologies like RFID and beacons. But as Matt pointed out, few are changing their “location point-of-view” to that of the things.
Teenagers don’t have wheels but they move around in things that do. When I purchased my OBD II device, a consumer product called Zubie, one of the value propositions advertised was the Family Plan; better known as “knowing where your kids are with the car at all times.” At the time all of my kids had moved out and bought their own cars so that wasn’t a big appeal to me. But I suspect it is for many parents, particularly those whose kids learn how to turn off “Track my iPhone.” But just like knowing the location of your loved ones on a Saturday night gives you comfort, knowing the location of your data streams is critical to your state-of-mind regarding the health of your business.
So the question stands – do you know where all your data streams are?