Jun 6, 2018 | 2 min read

Podcast #13: What the Future of Tech Holds – Parsing 2018 Predictions with Duncan Stewart of Deloitte

Duncan Stewart is the author of Deloitte’s annual Tech, Media and Telecom Predictions, which provides specific forecasts around tech trends and provides extensive strategic food for thought for businesses in all industries.  Our conversation covered the predictions regarding Augmented Reality, the “invisible innovation” of the smartphone and other topics. Most surprising are the colossal price declines in Machine Learning made possible by new generations of GPUs and FPGAs.  Duncan predicts that corporate use of AI and machine learning technologies will double this year and again in 2019.  We also discuss the implications of today’s “ad-allergic” media consumers and the expected fallout from recent controversies around data ownership. This is an episode you will not want to miss!



Deloitte TMT Global Predictions 2018

The Long Way to a Small, Angry Planet by Becky Chambers

A Closed and Common Orbit by Becky Chambers



We'll notify you weekly about new podcast episodes, upcoming guests, and news. You can subscribe to the podcast and if you'd like to be considered to appear on the podcast contact us.


View Transcript

Hello everyone, and welcome back to another episode of the Momenta Edge Podcast, this is Ed Maguire, Insights Partner. Today I have a distinct pleasure of welcoming Duncan Stewart who is the Director of research for Telecom Media and Technology at Deloitte Canada Duncan, it’s great to have you. 

Thanks for having me on today. 


We go back a little way; Duncan and I were actually counterparts or competitors going back a little bit! 

20 years now at it I guess, something like that. 


Well, both of us did our time as cell-site analysts and have found our ways to doing other more meaningful work! So, I guess it’s all in a process, but anyway, it’s great to have you on Duncan.  

One of the things that Duncan is responsible for is, Deloitte’s annual technology medium, and TMT predictions. This is a document that essentially outlines vision for some of the most important trends over the next year and decade, frankly. It’s widely read. Duncan travels around the world and is regarded as a leading thinker and researcher in the space. So, it’s great to have you on the podcast Duncan. 

Thanks Ed, and I’m glad you didn’t use the ‘F’ word. One of the things we try to avoid is to be called ‘Futurists’, we don’t like that term. Futurists tend to talk about, ’20 years in the future, this is gonna be big’, but there’s seldom a precise timeframe, and there’s seldom precise numbers. 

One of the things that Deloitte predictions has done, and continues to do, and will always do is, that we tend to be extremely precise. For example, we don’t say, ‘Oh, we think sales of tablets are going to decline in 2017’, we say, ‘We think sales of tablets are going to decline, and at end of year the number of tablets sold with be 165.0 million’, which allows our clients to make assumptions and do things with that. And of course, when IDC comes out and says the actual number sold was 164 million, we’re pretty pleased, because that means we were accurate within a single percentage point. So, that’s what we strive for. 

Across the 2017 predictions our accuracy rate, and anybody can score these at home, our accuracy rate across our predictions last year was 90 percent, 9-0, so that’s a pretty good track record. 


That is, I think that’s even more accurate than Ray Kurzweil, he claimed about an 86 percent prediction. 

And no offence to my buddy Ray, who I’ve never met, I’ve gone over his scoring and I respectfully disagree with some of his self-assessed scores, but anyway. 


And he is much more of a long-term futurist. So, what I think would be helpful for the listeners would be just to talk about your background, and what in particular had shaped the way that you look at, and examine the technology landscape? 

Although I’m not a futurist, I’m sort of a futurist! So, the way it came about is, our shared history as onsite analyst, I was a Fund Manager at buy side for a number of years, sell side for not that many, I think only seven in total, but that idea of not just saying, ‘Here’s what I think about the future of this technology’, but that discipline of being, ‘Here’s how many units Apple-is-going-to-sell-in-this-quarter’, that kind of precision from the sell side is not normally applied to the business of predicting the longer-term future of technology. So, what we’ve done at Deloitte is we’ve created a kind of hybrid model, we’ve got the attempted precision and timeliness of a sell side research report without any investment implications, obviously. We’re not talking about buying and selling stocks, but I am taking that ability to build a multi-thousand-line model to try to get to some fundamental truth, one, two, or three years out. That’s really the inspiration for what Deloitte predictions does, and how I look at the world of technology. 

The only other comment I’d sort of add there is, we looked a lot at the data; you’re probably familiar with the world of medicine, there’s a thing out there called evidence-based medicine which is we don’t use this drug, we don’t use this procedure or surgery without looking at actually does it work? And the idea of evidence-based medicine is a lot of the things that surgeons used to do, like ripping out appendixes willy-nilly, that sort of stuff was not supported by the data. So, with predictions what we try to do is to look at the data; do consumers actually use this device, and that kind of thing. 


Talk a little bit about the origins of the predictions. You guys really are putting a stake in the ground, you do have some skin in the game as it were, in terms of at least committing yourself to numbers. But, what led to the effort to really be able to provide this type of data, and how do your constituents, I’ll say it consists as a broadly read publication, how do people use the data? 

Well the origins were 2001, a guy at Deloitte named Paul Lee, Paul is my colleague, partner, and co-author of the predictions, Paul got the idea of writing this back in 2001, was supported by a great guy at Deloitte, now passed-on sadly, Igal Brightman, they came up with the idea of doing this many, many years ago, 2001. We’ve published it every year since then, it has become over time I think, according to the data, our most widely read, Deloitte’s most widely read Tech, Media and Telecom document. 

How do clients use it? There’s two really important parts to that question Ed, so it’s a great question; the first part is, who are the clients who use it? You might think Tech Media Telecom obviously, and that is true, but what we’re discovering is that more and more, people outside of TMT are looking to Tech, Media and Telecom as their growth levers. If you’re a bank these days, it’s not about how many physical bricks and mortar branches do you have on the corner in Chicago, or New York, it’s what’s your FinTech ability, what’s your back-office doing, how’s your website, how’s your mobile app? Over, and over technology is the differentiator in nearly every industry, whether its financial services or retail, it doesn’t matter. Around the world people from every industry are dying to know what comes next in Tech Media and Telecom? It is to quote Phil Asmundson who’s our former technology telecom leader, Phil once said that tech media and telecom isn’t just an industry, it’s a service line, it’s a service that we can offer to everybody. So, people around the world are looking to the predictions in all industries. 

How do you use it? We seldom tell people exactly what to do, instead we say, ‘Here’s what’s going on, here’s the data’, we try to be real myth-busters about it and talk about things that either other people are talking about, or going out there and saying, ‘This thing that everybody says is going to be big, no it isn’t, and here’s why’. That sort of data then goes into their strategic process, so this is the sort of stuff that CEOs, CFOs, CIOs and CMOs, they’re gathering around, and they take the data. We go around the world and we talk to them, and we give them the data, and then we have conversations with them about what their response to that information should be. So, it’s a critical part of the strategy process for multiple companies, multiple industries, and multiple geographies.  


That’s great. Let’s dive into some of the work that you’ve done for 2018, and what was noticeable this year was the fact that you highlighted augmented reality, which has been an area I think where we’ve see VR be incredibly hyped a couple of years ago, and a few unnamed pundits that I’m friendly with had got I would say a little bit overly excitable about virtual reality, but now augmented as augmented reality is starting to emerge. First of all, if you could highlight some of the forecasts, and also what has changed in the market that’s really shaped the mainstreaming, and the emergence of augmented reality as such a powerful shaping force? 

If I recall Ed, you’re a bit of a baseball fan, right? 


I am, yes. 

So, the batter was up at the plate, and he’s owing too, because it’s not just VR, it’s AR. So, three or four years ago we had the AR thing, augmented reality glasses that people would wear, they could record video and they could have information displayed on a lens only an inch or two away from their eye. A lot of people thought that was going to be huge, and Deloitte wrote a report to say, ‘No it isn’t, people aren’t going to wear these’. Then two years ago, everybody was saying virtual reality, once again the head-mounted devices, especially the high-end head-mounted devices, people were looking for these units to be selling literally billions, tens of billions of dollars-worth of head-mounted displays. We looked at our own research and we conclusively said, ‘There’s no way, there is just no way… some gamers will buy it, longer-term the enterprise market might be interesting, but VR headsets we predicted two years ago would be a massively disappointing failure. And of course, we were entirely correct. 

When we’re talking about VR and AR, those are two strikes, the glasses weren’t a big success, and the head-mounted displays for VR weren’t a big success; they’re out there, but they’re small numbers. Where we think it gets interesting this year is, on the smartphone itself, were not asking people to tie a computer to their face; over and over consumers have proven that they do not want that solution, and no matter how much Silicon Valley and venture capitalists believe in this as a growth market, consumers themselves at the end of the day will not wear these devices, so you can give them to them, they’ll play with it once or twice for 12 minutes, stick it in the draw and never look at it again. So, conclusively up until 2018, that’s been very clear. 

What we think is more interesting is, the use of augmented reality on the smartphone, everybody’s got a smartphone, and unlike head-mounted devices people love their smartphones and carry them with them all the time. So, the ability of creating and viewing augmented reality content on smartphones, especially with the various improvements in smartphones, both software and processers enabling a very high-quality AR experience. If you look back at even a few years, AR was there, you could do it on the phone, but the lighting was bad, the resolution was low, motion was jerky, and objects tended to float in mid-air and not cast shadows. 

In the last year or so we’ve seen tremendous progress on that, and augmented reality on the mobile device is much higher quality experience. So, the Deloitte prediction for this year is, that over a billion people worldwide will be creating AR content. Now, I’ve got to be honest Ed, most of it is the selfies, the cat ears and the dog ears, the goat ears, and gosh knows what else. 

Super-useful stuff yeah! 

Well, no-no-no, and Ed you’re of my vintage and I’m going to put back on you, the first device that the average American ever owned which had an integrated circuit on it, played digital ping-pong. Frequently new technologies began with the triviality and inconsequential, and then that’s the thin end of the wedge, and that’s their entry point to the broader consumer market. So, I do think that AR this year, there’s a lot of people using it and playing with it, and as you say, it’s not exactly world-changing, but it’s the start, we’re in the first inning and right now a lot of people are getting used to the technology and figuring out what it can and can’t do. So, we think AR on the handset is a real thing, and a growing thing. 


It’s interesting because if you go back a couple of summers and Pokémon Go seemed to be that trigger, that realization for a lot of folks. When you look at the used cases of AR in industrial contexts, I’ve just had a conversation with a company, Augmate with the CEO, and they’re very much focused on using AR and headsets for field service. Do you expect to see AR used cases being much more broadly adopted on the smartphone for business-use cases? 

Let’s split that question into two, so first of all, augmented reality on headsets, everybody goes, ‘Ah yeah, this will be great, you can use it in field repair; send a lineman up the utility pole with a pair of AR goggles and he can see which is the live wire’, and stuff like that, there is some of that. I gave a speech down to the utility industry in the States down in DC a few weeks ago and we had this conversation. One of the problems with those AR goggles in the field force, up on the pole; it rains out there like a lot! These devices don’t tend to be rugged, sometimes they’re down in a place where it’s a harsh environment, a potential risk of explosion, you can find space with dangerous gasses, you can’t have a device there that might potentially produce a spark. So, they’re not that rugged, they’re not that safe. 

One of the critical issues is, dropping a drop-down menu in front of a guy working on a 20-kilovolt line and obscuring something that he/she needs to see, is potentially a very serious hazard. So, I’ve spoken to the folks in the utility industry and they’re like, ‘Yeah, that will happen one day, that’s certainly going to be a thing, and we’re interested in it, we’re following it but it’s not imminent’. Instead those AR headsets are certainly being used in medical, and the other one that certainly jumps to mind is manufacturing, working inside a control factory you’re not working on anything dangerous, but you’re running a bunch of fiberoptic cables or wires through a wiring harness; having AR goggles drop down showing you which wire to feed where, that’s a real use case that does seem to be used. 

Will people do that on their smartphones in the enterprise world? There will be some of that, but even there, a little bit cautious because frequently you don’t want to hold your phone up, and hold it over something, because then you’re down to one working hand. So, we’re going to see a mixture. 

But what I do like Ed, and I posted this on LinkedIn, Facebook, and Twitter two weeks ago, maybe one week ago, Google did a really superb example of Augmented Reality and it wasn’t on a smartphone, and it wasn’t on a headset, it was on a microscope. What they did is, they took a conventional microscope in a pathology lab looking at slides of potentially cancerous tissue, and they combined machine learning and Augmented Reality so that the pathologist, when he/she peers down the lens through the microscope they’re looking at the tissue, but there’s a little green circle highlighting a bunch of tissue that to the machine learning algorithm doesn’t look quite right. This to me was just a brilliant example, because it makes the process faster, there would be a real ROI in detecting cancer or saying it’s not cancerous, faster, and it fits within the existing workflow. 

Anybody who knows anything about the enterprise market, if you’ve got a rip and replace strategy where people have to throw out everything they’ve spent 100 years using before, that’s a tougher sell than, you keep using exactly what you have, exactly the way you have, we’re just going to make it work slightly better. So, I thought that Google example of Augmented Eeality through the microscope and the path lab, I thought that was an absolute home run. 


That’s a great example. I think what’s becoming apparent is, that smartphones do continue to evolve, and as we’ve been looking at Connected Industry and the evolution of the Internet of IoT, it isn’t just the connected fixed devices, or static devices that are evolving. The smartphone itself as an extension of the human, if we think about the ability to track people or essentially provide information actuation to human actors in the field, is incredibly powerful. 

I wanted to move on a bit to this concept of invisible innovation that you have talked about. Could you explain a little bit what you mean by invisible innovation, and how is that going to play out in the industry? We’re talking about smartphones, and over the next several years a lot of these innovations are creating just incredibly powerful capabilities in the palm of our hands. 

This is sort of an automotive analogy, the outside of cars these days just doesn’t change much. I’m old enough to remember when I was in the seventies there were still cars from the sixties driving around, and man they looked cool. Some of them had wings, and some of them from the outside looked totally different. These days every single car looks more or less the same, because they’re all aerodynamic and they all do this, and they all do that, and we’ve sort of standardized, and there’s sort of a form. But inside, you could have a gas engine or electric battery, or bigger/smaller motor, better brakes, various levels of autonomy, all of the innovation in cars from now until the next 10 or 20 years, all of its going to be under the hood. It’s maybe going to cost more, it might save your life, but you wouldn’t be able to tell by just looking at the outside of the car. 

In the same way the smartphone externalities are kind of frozen, we can’t make them much better than six-inches and still have them fit in our pockets, or in the human hand. Similarly, the display, the high-end OLED displays on the top-end phones today, they’re great, I love them; super clear, super sharp, the brightness and all that kind of stuff, but there’s nothing out there right now that we’re working on that’s better than that. So, when I look at the ability to look at a phone and say, ‘Hey, this phone is two-year’s old and I can tell just by looking at it’, that’s not going to happen anymore. Going forward, phones will get better, but it will be on the sensors, on the cameras, on the processors inside, machine learning chips dedicated to doing machine learning on the edge device. These are the trends we see. 

It’s not bad news, but it does the idea that you don’t wear last year’s clothes because they look out of fashion kind of idea, with the smartphone we see the external aspect of the phone remaining unchanged for at least the next five years, possibly the next decade. 


So, what that will create is essentially incredible power, but really it ends up being I guess subsumed to the design. So, no you’re not anticipating any radical shift in design. 

Not only radical Ed, I’m not anticipating minor shifts in design, I mean this whole thing about its got a notch, it doesn’t have a notch, I mean for God’s sakes I’m rolling around here saying, certainly it’s different but it’s not a major-major design decision. 


And of course, that is incredibly powerful, and as you’ve discussed what’s evolving under the hood, what I thought was really interesting in the report was some of the work you’ve done about how AI, the evolution of AI and cognitive technologies have shaped the development of applications and features, can you talk a bit about the work that you’ve done around AI with respect to phones and the user experience. How is this technology going to change the way people interact with devices? 

There’s two or three topics all jumbled together in one there, and that’s not your fault Ed, that’s mine. 


Of course, that’s the classic multipart question as well! 

I’ve got to take a swing at all of them at the beginning. So, in our report this year, we have an entire section on smartphones, and we have an entire section on machine learning chips, and then we’ve an entire section on machine learning in general. They’re all connected, the big thing that’s going on, and I’m going to step back two years ago to 2016. In 2016, if you wanted to do machine learning, you did it in the cloud. Machine learning, either training or inference occurred in the cloud because machine learning was run inside data centers on giant racks of cards, and the chips on those cards cost 5,000 bucks and burned 300 watts a piece, and that’s where you did machine learning. The only place you could do machine learning in 2016 was at the core of the network in the data center and connected over the cloud. That was an architectural reality, there were no other options. 

Now, there’s a few things going on, in that data center at the core of the network there are new chips being used, not just the CPUs and the GPUs, the Graphics Processing Units that we used in 2016, they’re now being augmented with A6 applications, specific integrated circuits, and that’s not just the Google TPU, lots and lots of other companies like Facebook and so-forth are out there hiring chip designers to go build their own version of proprietary applications, specific integrated circuits to accelerate machine learning in their data centres. FPGAs Field-programmable gate arrays are being used to accelerate machine learning in the data centre, and that’s half the story. 

By the way, all of those chips used in their various combinations still with GPUs, the GPU doesn’t go away, all of those chips effectively work to drop the price per machine learning by orders of magnitude. I have seen estimates of the cost of machine learning on a per decision basis, is down a thousand times I’ve seen estimates that its down a million times, in the last two-years. We are talking somewhere between a three and six order of magnitude decrease in the cost of machine learning, in the core, at the data center level in a two-year period, which is raised to the moors. 


That’s pretty unbelievable, this is the first time I have heard that. When I look at exponential technologies, this is hyper-exponential almost. 

That’s the phrase I’d use, it’s a big one. So, the idea on that side is profound and transformative, but it can be simply boiled down to the fact that doing machine learning in 2016 was somewhat tricky and expensive, it is now both less tricky and less expensive, and when you make things cheaper and easier they are more widely used than they were before. That’s not a radical statement, but you and I have been around in technology long enough to know that is essentially true, more or less 100 percent of the time. So, that’s part one of the equation. 

Part two is in my view a much more interesting one, believe it or not, more interesting than exponential-exponential. If I wanted to do machine learning in 2016 on a turbine in a factory, I had to connect the turbine over a network, and I had to worry about latency, and I had to feed my sensor data from the turbine into the cloud, run my predictive failure algorithm, feed it back and shut down the turbine before it rips itself to pieces. In 2016, that is the only road, the only road to get there. In 2018, what we’re looking at is, there are now chips, there are now APIs and SDKs; so, we have the Apple devices, the bionic family, the A11 has machine learning on your smartphone if you have the 8 or the 10, the S9 from Samsung has machine learning on the device, the Wuawei has machine learning on the device, the Google Pixel 2 has machine learning on the device, I can keep going. So, it’s on our phones, that’s one place. 

It’s on our computers; two weeks ago, Microsoft came out with an API so that you can now use Windows to do machine learning on your PC, and there are chips being built and being designed that are at the extreme Edge, these are devices that would be in volume 15,20, 25c burning micro watts of power, very-very small. The point that’s going on here Ed, and this is what I think is the most transformative prediction that we made, is the idea of Edge machine learning, we are not just doing it in the core, the data centre, we will be able to do machine learning inference for less than a dollar on the Edge device, which means I don’t need to worry about the cloud, I don’t need to worry about the network, and I don’t need to worry about latency. This is a game changer, this is something that two-years ago we would not have thought possible, and now this is happening in real time. 


That’s a substance of development, more than substantive. I hate to use the term ‘game changing’, but it really is. We’ve been looking at this broader transition in computing paradigms from centralized to decentralized, to centralized to decentralized once again, as we went from mainframe to PC client server, to cloud mobile, to now this concept of the intelligent Edge, and what you’ve just articulated is this quantum leap in processing capability and declines… 

Ed-Ed-Ed, I’ve got to stop you, don’t say ‘Quantum leap’, because one of my topics for 2019 is Quantum machine learning! So, don’t you steal my thunder now. 


Absolutely, and I think that certainly has become most topical, but this really is amazing that you have orders of magnitude accelerating unconceivably by any prior standard. That’s amazing! 

And to your point, how does all this translate into a prediction for our clients? Deloitte’s have a real simple prediction on this, not consumer, not playing chess robots, not the Google Home speakers, or the Alexis, at the level of enterprise; the Deloitte prediction is, that enterprise use of machine learning will double in ’18 compared to the year before and will double again by 2020. In other words, true exponential 100 percent growth over 12 to 24 months, on an industry that is already tens of billions of dollars in size. So, that’s the takeaway, so when we tell companies this, they’re kind of going, ‘Well, so I need to start thinking about this, start planning our 2022-2023’, and I’m like ‘No’.  

I was in Israel a few weeks ago now, there was a company in Israel in their data center, financial services, they had been using GPU’s to do machine learning, and they’re looking at FPGAs. I said, ‘Looking? What do you mean by looking at FPGA’?’ and they said, ‘The FPGAs arrived this week, and we put them in, we loaded them up yesterday’. So, this is not some sort of, you’ve got to put this on your roadmap, this is actually happening around the world now. And when I say now, I don’t mean now ’ish, I mean now. 


That’s amazing. I think the implications down the road for every industry, and certainly employment, the refactoring of processes and tasks, it is fundamentally changing everything. 

It’s every industry, it’s ever part of every company, it’s not just the IT department, its marketing, its sales, its logistics, its finance, every part of every company is looking at machine learning. 


I’d like to touch a little bit on some of the media topics that you focus on. Our focus is much more on Connected Industry from an industrial perspective, but I do think from a broader point of view its instructive to see what is happening to industries in media for instance, music, and publishing. I’d be very interested to get a sense of what you see transpiring over the next several years, essentially, it’s an arms race between getting messaging out, and then blocking it. I mean you’re talking about ad blocking technologies, and you think, ‘Well, why is that relevant to an industrial firm? Well again, if you’re trying to get your message out, and certainly if the avenues from which your constituents or customers, consume information becomes limited, or becomes diffused in another way, how is this technology, this arms race between the end users and those who were seeking to harvest data, to persuade people to hear their message, how’s this changing things? 

Three big macro trends there, so one of them is ad blocking, we’ve invented a word for this, I think it’s kind of cute, we came up with a name for people who don’t just block one kind of ad, but instead are blocking multiple kinds; they’re blocking on their computer, they’re blocking on their smartphone, they’re using a PVR, DVR, TV skip on ads, and they’re listening to Spotify in the car rather than radio, because they don’t like radio commercials, that’s somebody who is blocking almost every kind of ad out there. The word we came up with for this, we call these people, ‘Adlergic’, the idea that people are allergic to advertising. We went out and we asked people in Canada and the US, and we found that the percentage of people who are blocking more than half of all the ads they see is quite small, less than 10 percent. 

That’s not so bad, a lot of people in the media space thought it was more like 50 or 60 percent, so only 10 percent? I can live with that! The problem is who? That 10 percent of the audience who is blocking more than half of the ads out there, they tend to be young, they tend to be higher income, they tend to be highly educated, and they tend to be employed; meaning that from a demographic perspective, although relatively few people are blocking most of your ads, the ones who are blocking most of your ads are usually amongst the most desirable from a demographic perspective. So, if I am thinking about this from the person who is buying the advertising, how do I reach young highly-educated, high income people? 

There’s a bunch of solutions out there, we talk about this in the report but, also there’s other things; you can move away from an ad model, you can have a subscription model, and we have a prediction around that, more and more people getting more and more subscriptions, it’s not a cord-cutting world, it’s a cord-stacking world, and the example I always use is, I walk into a room and say, ‘Who here cancelled cable?’ Somebody puts up their hand, great. I say, ‘Now, who here in the last couple of years has got a Netflix subscription?’ Hands go up. ‘Spotify subscription?’ Hands go up. ‘Gaming subscription?’ Hands go up. ‘Newspaper or magazine subscription, like The Times, The Economist’, and a bunch more hands go up. Some of us have managed to cut one of our subscriptions like cable, but we’ve added four others, so our point is that the subscription model isn’t dead, its shifting and its changing, and people are willing to have multiple subscriptions. 

The final thing we point out is live, and as a musician Ed I know you’re enjoy this one; it’s a big deal, the global market for live is $545 billion a year. Around the world people prioritize live entertainment, and whether you’re going to make them pay for that through a subscription, or stick some advertising on top of it because it’s a lot harder to ad block when you’re looking at live, we think that’s a big three of media trends; not that many people are ad blocking, but some are, and if you want to reach that audience, you want to think maybe about a subscription model instead, or you want to think about focusing on live content, because live is the stuff that isn’t going over the top, it’s the stuff that’s being consumed frequently with advertising. 


It’s encouraging I guess, certainly for content creators. I do have a question, whether you’ve given any thought to whether we may be seeing a data backlash in this, with the revelations around Cambridge Analytica, Facebook testimony, of course in Europe GDPR is about to come into place in May, but these business models that have been based on harvesting people’s data, I guess the old joke was that if you’re not paying, then you are the product. How do you see this dynamic shaping the way that not just media companies, but all companies that are going online with their messages over the next couple of years? 

This is not a Deloitte prediction, we have Deloitte predictions and we spend months researching and writing them, and we have not done one on this. This is more Duncan as a guy who travels around the world, talks to people and watches things. Certainly, there’s a lot of people out there saying they are quitting social networks, quitting other online things, in an effort to protect their personal identifiable information. A lot of people say that, and when I look at the actual numbers it’s not unlike the percentage of smokers who promised to quit on January 1st, a lot of people have intentions around this, but the follow-through is weak. A lot of people actually never log-off the social network, and of those who do, a lot of them are then coming back within a month or two. So, from a consumer perspective I actually believe this is not a big change, there’s a lot of talk and a lot of noise, sound and fury, but signifying nothing. 

On the other hand, the regulators I think is where it actually gets very different. We have GDPR in the EU, I believe the ability of regulators either in the EU or globally to enforce laws around privacy and sharing of data, I think that’s going to be the thing that could be a game-changer. Does Deloitte have an official prediction on this, no. Does Duncan have an official prediction? No, but at a very high level, if I’m not worried about the consumer reaction I instead would focus on what regulators do, I think that’s the thing that’s going to move the needle one way or the other. 


It makes me think of the old saying people used to say about cigarettes, ‘Quitting is easy, I’ve done it thousands of times!’ 

Well, the other thing is Ed, you and I have been covering the tech space for decades now, and when you look at ‘People are thinking this about Windows’, that wasn’t a big deal. On the other hand, here’s the anti-trust stuff, here’s what’s going on with Microsoft, here’s what went on with Intel, here’s what… going way back, take a look at the break-up of AT&T. When the regulators get involved things can actually change a lot. By the way, I’m not advocating that regulators should or should not, and I’m not suggesting what their actions should be, just as an observer of technology over three decades now, when the regulators get involved, sometimes big changes do occur. 


They do. I was having dinner last night with a friend of mine who is a first amendment lawyer, he’s a free speech expert, and a lot of his clients are very concerned about the data privacy issues that are emerging now are spurring in particular European regulators to be much-much more aggressive. It’s in his view that we can expect a lot more active, maybe potentially intrusive role from regulators all over.  

Great, it’s been fantastic talking to you Duncan, your insights on so many different topics. It’s a lot to pack into a short conversation. I will post the links to your work in the show notes. The one thing I’d like to ask is if you can provide a recommendation for a good book or resource that you would recommend to your friends or colleagues, do tech related, could be anything. 

Well that’s a wide-open question and it actually is right over the plate for me. I actually think that one of the things that helps me to write predictions more than anything else is reading Science Fiction. I love Science Fiction, I read a lot of other stuff too, but I read a lot of SF every year. Although I could recommend all kinds of stuff I’m sure, I just want to share one. 2018/17 has been… Im not American I’m Canadian, but there has been a lot of partisanship and a lot of negativity out there in the world and I don’t always love all that. So, there’s a couple books written by a woman named Becky Chambers and I’m also a big supporter of women in tech and I think that includes female writers in Science Fiction. So, Becky Chambers out there she’s written a couple books called The Long Way To a Small, Angry Planet and another one called A Closed and Common Orbit. Theyre both excellent Science Fiction and nominated for multiple awards, it’s a finalist for the Hugo. The thing is with Becky’s book is it describes a universe where there’s still conflict and so forth but where people get along. It’s a happy, happy Science Fiction universe. There’s still things that need to be solved, but it’s one of those books you read and you laugh and at the end of it you feel kind of good and positive. I think in 2018 we need books like that. So definitely recommending Becky Chambers two books, one from 2014 one from 2016 and shes got a third one coming out soon. 

That’s terrific Duncan, great recommendations. Its always nice to highlight these types of books because people are always looking for particularly something that will leave a smile on their face and get us away from some of the dystopian visions that seem to predominate. So, as always, Duncan its a pleasure to talk to you. We have been speaking with Duncan Stewart who heads up research for Deloitte’s TMT effort out of Canada. This is Ed Maguire, Insights Partner at Momenta Partners and this has been another episode of our Edge Podcast. Thank you for listening.