Running on data Activity trackers and the Internet of Things


The fitness tracker on your wrist may be the most evident sign of the Internet of Things, but it is just one node resting on top of communications, analytics, policy, and even behavioral infrastructure. Host Tanya Ott talked to Tom Davenport, independent senior advisor to Deloitte LLP, about the future of IoT.

The Press Room
Subscribe on iTunes
Listen to more podcasts
View the collection
Read the full article.
Receive IoT insights


TANYA OTT: This is The Press Room, Deloitte University Press’s podcast on the issues and ideas that matter to your business today. I’m Tanya Ott and today we’re talking about the Internet of Things… all of those devices we use everyday that sense things like movement and temperature and collect data that can help us be more efficient, safer, healthier. They can help us if the makers of those devices know how to use the data.

TOM DAVENPORT: “I don’t think that most businesses have in mind exactly how are things going to work when we get all of this data. Even if I’m selling consumer-oriented devices they’re not sure how consumers will use it or what will they care about.

TANYA OTT: There are about 20 billion Internet of Things devices in the world. Everything from fitness trackers to smart thermostats to car backup sensors.

TANYA OTT: In fact, when you load up your car for a late summer vacation this year you’re getting into a big hulking steel box filled with Internet of Things things. These days the average car as about 200 sensors, says Tom Davenport. He’s a professor of Information Technology and Management at Babson College, a researcher at MIT, and a senior advisor to Deloitte LLP.

He says our devices are pretty good at sensing things… figuring out just how far away your bumper is from your neighbor’s mailbox. Or counting how many steps you took on your morning walk today. But, Davenport says, that just the first step.

TOM DAVENPORT: I mean all of this information is not terribly useful if left at the level of the device. And so if we want to do any control, coordination, monitoring, we typically have to get it somewhere else and that becomes somewhat challenging in part because of the number of devices and many of them have different data formats. So it’s been estimated that if the average car has 200 sensors there’s almost as many data formats. And so if we like to find out everything that’s going on in your car, for the driver or for someone who’s regulating traffic, you’ve got a huge data integration job in order to make sense of them all.

TANYA OTT: So, is this kind of like the old quandary of Beta versus VHS for video or more recently, MP3 versus M4A for audio? This is a barrier that we have in adapting uniform standards?

TOM DAVENPORT: That’s part of it and in a way it’s, you know, as if we had 20 or 30 different standards for video tape format instead of just two because if anything the number of standards is proliferating and there are lots of standards-oriented consortia out there. Those are proliferating as well and so we’re not really converging at this point.

TANYA OTT: How significant a barrier is this standards issue?

TOM DAVENPORT: You know in the past it was a big barrier and in order to sort of find the answer to that question I consulted a friend who’s at MIT and who was at the head of one of the consortia to do radio frequency identification devices, RFID, which was arguably the first real internet of things network. It didn’t actually operate via the internet, but it did exchange a lot of information. These have been put on a lot of, you know, shipping pallets and cartons and boxes and so on. I asked my friend, ‘how long did it take to come up with a standard to identify one of these devices?’ and he said ’15 years.’ Now his view is that maybe we can accelerate that integration in the current environment b/c we have cloud computing now. You can sort of do a translation in the cloud that would connect one different device up to another. But it’s all rather sobering, I think, and suggests we’re not going to be pulling all this information together very quickly.

TANYA OTT: If a dominant vendor doesn’t emerge another option might be that lots of different folks come together across industries to set standards or that regulators of some kind of governing body mandates standards. So far, though, that hasn’t happened.

TOM DAVENPORT: The other problem is there are so many different device types to what might work for a fitness tracker might not work nearly as well for a heat and pressure and vibration sensor in a jet engine. So I think it’s unlikely we will ever have just one standard.

TANYA OTT: Davenport says there are cloud service providers who translate data from one form to another, usually within a particular sector. But then there’s question of what to do with that all of that data.

TOM DAVENPORT: Analytics are how we make sense of data and make sense of how we need to take action on it. And thus far we’re not terribly far along in that in that space. Most of the time with sensor data you tend to see, if you see a bar chart of Internet of Things data that’s suggesting that we’re not terribly far along. That’s a pretty rudimentary approach to analytics. But at some point, say in fitness trackers, we might see some predictions as to you know if you keep on walking or not walking at the level where you are you’ll contract type-2 diabetes within five years or even comparative analytics where you’re better than 92% of the walking population or the FitBit wearing population. But we don’t really have too much of that yet. It’s all quite rudimentary and what you see if you upload your activity tracker data to your smart phone or to your laptop that’s the kind of thing that you see Across the Internet of Things in general.

TANYA OTT: So the first example that you cited was the idea of predictive, which is almost like when I go to my bank website and I can put in what my assets are right now and what I’m putting away and what my retirement plan looks like and it predicts out for me, ‘you can be kind of comfy in retirement’ or ‘you might be in trouble.’

TOM DAVENPORT: Exactly. You can even get now on many of these financial sites the percentage likelihood you will run out of money assuming you live to age 95 or whatever. And we don’t have that sort of thing with Internet of Things data yet on any website I’m aware of and it’s not surprising because we’ve had financial services data for a long time and it took them quite a long time to come up with that kind of predictions. So it usually takes them a while before companies get around to this kind of analytics phase. 1

TANYA OTT: And these are all what we call nudges. You know your sensor telling you ‘hey, sit up straight’ or ‘you’re almost to 10,000 steps’ or ‘if you don’t get more sleep you’re going to have a real deficit problem by the end of the week.’ What kind of evidence is there that these nudges actually work in changing behavior?

TOM DAVENPORT: You know just as it’s early days for the Internet of Things it’s early days for behavioral economics and there’s a fair amount of academic research suggesting that these kinds of nudges can be effective, but we don’t know how broadly to interpret the results. All of the research is quite narrow. So you know, it might work in one energy-related setting but not work in health and fitness. We just don’t know yet.

TANYA OTT: Davenport says some companies have begun assembling teams of data scientists to help them get smarter about what data is collected and how it’s used. That’s the final step in the process…. What he calls cognitive action.

TOM DAVENPORT: Once you’ve pulled all of the data together and analyzed it you have to start changing how you generate energy, changing how you manage traffic patterns within a city, and because it involves politics and behavior and organizational change that tends to be the toughest thing and the last thing that organizations typically address. I don’t think that most businesses have in mind exactly how are things going to work when we get all of this data. Even if I’m selling consumer-oriented devices they’re not sure how consumers will use it or what will they care about. If I’m talking about activity trackers I have to think, well, how will the health care establishment respond? Will people want to include this data in electronic medical records at a hospital? And who might do that at a hospital? Will insurance companies start charging less if you’re more active? There’s now one example of an insurance company in the United States that is offering discounts if your fitness tracker suggest that you’re being more active. But that took a while to offer that – fitness trackers have been out for almost a decade now and it’s still a pretty early stage plan by just one company. And if we’re talking about traffic, how to start changing the behaviors of drivers that’s a pretty complex phenomenon and even in Singapore where they’re quite used to legislation that changes the behavior of citizens they were not quite ready to start passing laws that controlled traffic based on all of the data they were getting from their Internet of Things system.

TANYA OTT: Tell me about that system in Singapore. I’m intrigued about how they were thinking of using the Internet of Things and sensors in order to change traffic patterns.

TOM DAVENPORT: Tell they already have done sort of the first three layers. They have sensors all around the city. They use data from taxi cabs, they’re putting sensors in the streets. And they manage to pull a lot of it together and they have a relative rudimentary I’d call it descriptive analytics display – a dashboard – of what’s happening around the city at a centralized location in terms of traffic activity. But beyond that you have to start talking, well are we going to close off certain streets if it gets too congested? Are we going to give people certain economic incentives or disincentives to drive certain streets? You know Singapore does charge for driving with in the city, but it’s a single charge for anybody who passes into a certain part of the city – congestion charging, it’s sometimes called. You’d like to get a little bit more granular and more precise about maybe you would have lower charges for streets that are not terribly congested and we would know that based on the data, but it just gets very complex and confusing to citizens and so I think it will be a while before we start to use all of that data for really sophisticated cognitive action.

TANYA OTT: Have you driven in Singapore? That sounds really …..

TOM DAVENPORT: I have driven in Singapore and you get used to there are these big gantries over the streets that say you’re entering the congestion charging area – and then you get charged from an account. It’s not that different from being charged by your car device when you go over a toll bridge but they’re going to start taking away those gantries now so it will be a little less clear and they’re thinking about differential pricing at different times on different streets, but they haven’t adopted that yet.

TANYA OTT: You’ve driven in Singapore… what would that experience be like if you had to be, well if I take a right here it’s going to cost me more than if I go up a mile and take a left?

TOM DAVENPORT: Well I think at some point you will have maps and GPS systems that will take that into account, just as have systems that tell you if you take a right here that will add 10 minutes to your trip. We now can sort of plan the time that a trip takes pretty well with GPS devices and online maps. I think at some point we’ll be able to plan the most economic route. It might just tell you ride your bicycle … or walk.

TANYA OTT: Tom Davenport says one concern underpinning all of this is security. Many of us have been swept up in credit card data breaches in the last few years. But did you know you car can also be hacked?

TOM DAVENPORT: There have some very well-publicized incidents of people hacking into car sensors, and as you’re driving by they can make the doors unlock and make the windshield wipers go on and so forth. Some evidence that you could hack a jet engine on an airplane. It’s a little bit harder to do from a bandwidth standpoint because they’re further away and not something that you’re going to do through Bluetooth, but that obviously becomes quite scary if you think about the dangerous implications of what we might do here. So we need to be much better than we are at protecting all of this information and the ability to control a complex machine from afar.

TANYA OTT: It’s a lot of think about. But Davenport says if companies break it down and think through all of the steps there’s a lot of opportunity and untapped possibilities in the quickly expanding Internet of Things.

TOM DAVENPORT: You want to be thinking about the big picture and not just think, oh wow, let’s put a sensor in one of our products that we’re selling here. You want to think: where is this going? How am I going to integrate the data? What kinds of analytics might I ultimately do on it? And what kinds of actions would I be able to take if I had all of this?

TANYA OTT: Tom Davenport and his co-author John Lucker of Deloitte Consulting LLP lay our more thoughts in their article “Running on Data.” It’s available at

TANYA OTT: I’m Tanya Ott for The PressRoom, a production of Deloitte University Press. Be sure to subscribe and let us know what you think about the issues we’re discussing. Leave us a comment on the podcast page, email us at or tweet us @du_press.

This podcast is provided by Deloitte LLP and its subsidiaries and is intended to provide general information only. This podcast is not intended to constitute professional advice or services of any kind. For additional information about Deloitte LLP and its subsidiaries, go to