Data-Driven World: Innovate or Perish (ep28)
Nemanja: This is Chair, a place where we discuss innovations.
You know the awful feeling when you’re running out of storage in your cell phone. Well, in the last two decades alone, we amassed more than 59 zettabytes of data. That’s, let me say this correctly, that’s 125 million years of F.R.I.E.N.D.S., which is also my favourite TV show. The problem at hand is how to properly and ethically use this data to stay ahead of the competition. So we will try to answer this question today with help of our guest Bojan Ciric. Bojan is a senior leader in Deloitte New York. And he’s a part of the analytics and cognitive team. In his spare time, he’s a passionate cook, who dreams of becoming a chef.
Bojan, welcome! A pleasure to have you here today.
Bojan: Thank you. Thank you for having me.
Nemanja: So we are going to talk about data today. And obviously, humans have been gathering data for a long time now. Since humanity started, how has this changed in the digital era that we’re living in right now?
Bojan: Thank you Nemanja, I think this is a great question and it’s a very interesting subject. We live in the data-driven world, and data has an impact on our everyday lives, but what does it mean? So let’s look at a few examples.
If you look at Facebook, Facebook is the most popular media company, but it has almost no content. If you look at another big company like Uber, it’s the world’s largest taxi provider, but they own no vehicle, right? If you look at Airbnb, which is the world’s largest accommodation provider, it owns no real estate. Finally, we look at Alibaba, which is one of the most valuable retailers and owns no inventory.
So if you look at all these companies they are technology and data companies. On the other hand, the data is rapidly changing. So you just mentioned one great example about the volume of data that every few years, we actually have double the capacity of data that we are processing and using and so on… But there are other aspects of data. Let’s say, the variety… At the beginning of the data journey, we were able to process, manage and use only structured data. But now we have semi-structured, we have unstructured data, we are processing images, we are processing videos, we are processing voice. You have the situation that when you are talking with the customer representative of some company, complaining about their products or something, we have the analysis of the real-time of this information. And the Customer representative gets valuable insight on how to manage and drive that conversation. That gives another aspect of data when we talk about velocity. That in real-time we process, we manage data and we provide valuable insight. So in general, we are making decisions every day. We make personal decisions, we make business decisions, and instrumental for these decisions is the data. Actually, the insights that we can get from this day. The better insight, the better decision will be. And in today’s competitive world, the companies who can get great insight from their data potential, in the shortest period of time, will be in the market game. It’s that simple.
N: So you mentioned some of the companies there and obviously if a company is not leveraging that data they are going to be destroyed by the market. Right? They are not going to succeed. I’m sure that based on your experience, you have great stories of data usage. Can you share with me some good examples of how some companies use data to improve business and after that basically kill all the competition with it.
B: Yes, for instance, I mentioned these companies. Obviously, everybody knows them. But there are some other areas. So for instance, one good example is the John Deere company. John Deere is one of the world’s largest providers of farm equipment. And they have been well known for decades for doing this, but now they have also become the data company, right? So because of the development of the Internet of Things, the devices that can basically take — and the sensors that can take and transmit the data over the internet, tremendously improve the agricultural business. Because now you can put a lot of Internet of Things, the different kinds of sensors in the fields. You can get real-time information about the weather, you can get an image of the field and based on that you can make the optimal decisions on how you will manage that field. Also, you can do the setup of the equipment to do the spraying or something like this. So this is something that tremendously improves the outcomes and the value of the specific business.
N: Yes, you gave me a great example, which is a bit counterintuitive, because we’re talking about agriculture. And whenever somebody is talking about data, it’s usually technological companies. I remember, I think, 10 or 15 episodes ago, I had one guy, here from one huge company here in Serbia and they talked about how they’re doing lots of things in agriculture. Sugar beet — that’s something that they do and they talked about “uber-azing” their fleet. It was very interesting to meet them. How traditional companies that don’t have any connections, obvious connections with data, they’re becoming data companies.
I will leave that on the side now. I want to ask you about the technological companies. From your perspective, from your experience, can you share with me how you see, for example, how Netflix is using the data? And how is Netflix leveraging the data that they have?
B: Yes. Basically, I think most of us actually heard about the TV show The House of Cards, a very popular show. And there is an interesting story. Basically, how Netflix actually committed to doing the two seasons even without knowing that the first season would be a success. Based on their large data potential, taking the different information about, let’s say, when we watch movies, we’re doing some ratings, we are picking some movies or shows we want to watch and based on that, they have created a segmentation of their customers. They have literally created our profile, what are our preferences, what we want to watch and so on. Based on that, they were able to close with some additional information about that specific show, let’s say, the House of Cards. Meaning who will be the director, who will be the cast and what will be the topic of the specific show. And then crossing this large information about the Netflix users with some characteristics of the specific show, with a big certainty they knew that that show will be successful.
N: And they nailed it actually. Yes, an interesting story with this. How it goes, usually you have a pilot, test the market, and focus group, but they did it without right?
B: And even they went further, right? They have created specific profiles of users, and then they have created the different trailers for that show. Which will actually target the specific user's group. So that was one additional thing. They secured the success of that show.
N: Okay, so Netflix, that’s a great example. But let’s use another one. There is lots of talk about Elon Musk and what he’s doing with SpaceX and Starlink, but Tesla is something that is almost basically financing a lot of his endeavours. So let’s talk about Tesla, how they are leveraging big data and analytics.
B: Yeah, Tesla definitely is a great company. Great example and one of the most innovative companies in the world. And what is an interesting and hot topic, for instance, is self-driving technology. The service — self-driving technology is completely based on the data and advanced analytics on top of the data. So I will give you an example.
Basically, if you want to break down that specific term into four different areas. One is perception. You need to know what objects are around you, the other vehicles and everything. The second is localization. You need to know where you are at this specific point in time. The third part is planning. So you need to find optimal roots to go from point A to point B. And finally, the fourth aspect is - control. You basically need to steer now between point A and point B, different angles of steering. So, these are the segments of self-driving, and each of these segments actually requires a huge amount of data. Also, the different kinds of analytics algorithms will analyze this data and make decisions during that self-driving process so you won’t have any crashes or something like that. This is the typical example where data is really instrumental data and different political techniques on top of that. It’s really instrumental to make some advanced features possible and to make this innovative.
N: Good thing that you mentioned self-driving. I think a couple of weeks ago there was a guy who was arrested in the States, who was driving and sleeping in the back while Tesla was self-driving. This is an example of ethics and how everything crashes with the data. One of the bad examples…
B: Yes, it’s a great example, but also gives some other aspects that we need to talk about. Basically, this is about data management, how to make sure that, let’s say if you have a huge amount of data if we make the decisions based on this data, how to make sure that this data is actually accurate, and that we can trust this data.
N: So, we mentioned a couple of good examples. How companies are leveraging data, and they’re getting better and getting more profitable, but can you share with me some examples of the companies that haven’t leveraged this data and maybe died in the process?
B: Yes, I will give you a few examples of different companies. The most recent example is, for instance, the financial services industry. This is my core industry. So, we have one global bank, which actually got a fine of $400 million, in the US. And the only reason that they were fined, basically, is because they did poor management of that data. So the bank is now working to fix their data, but that’s a huge liability. Right? Another aspect is if you look at, let’s say, we all know about the financial crisis 2008 when we had the double failure of data. First, the bank did poor management of data, they did a lot of house loans to the customers who were basically insolvent, so they did a poor analysis about to whom they need to give the loan. So, that’s one thing. The second failure is that regulators actually failed to control the banks, to control the data and how banks are operating. And then we came to the situation that the whole system was close to a crash. So basically, as a result of this, there are a set of regulations and scrutiny so now, basically, the banks are forced to manage their data properly. And that’s the only way that you can make the financial system stable. So the data management and the data comes into the very core focus on the entire business. A few other examples are, for instance, Yahoo, or Nokia. So Yahoo was a dominant search provider, but again, based on the market research and poor analytics, they decided to focus more on the media content rather than search. And then we get a Google who comes in and overpasses them, right? And there was a situation to acquire Google. They decided not to go. So this is another example where you are making a poor decision because of you…
N: Even though you have a lot of data to make the decision.
B: Yes. So, another example is Nokia who was a dominant mobile service provider, they even made a good strategy, they created their operating system — Symbian. But they failed to recognize that actually, the future is in smartphones and applications. And then, we have Apple who came in with an innovative approach. And then Nokia actually failed and now basically, get out from that.
N: And they had a dominant position on the market.
B: Yes, yes. And now they aren’t even in the market. So that’s all about the data. And because more decisions that you make, you make on top of the data.
N: There is an addition to the data story, and it’s the buzzword of the day when somebody is talking about data and big data. There is always AI connected to this. My question is a bit, I won’t say tricky, but I want to hear your opinion on this. How do you see AI in terms of stealing people’s jobs in the future? What do you think about this?
B: That’s a very interesting question. So definitely the artificial intelligence algorithms and methods are taking a lot in terms of managing the data in the future, about the extracting that insight from data to make the good decisions, right? On the other hand, yes, there is concern that with process automation, with data automation in everything, a lot of jobs will be lost in the end.
N: That your company won’t be here in the future.
B: But on the other hand, I really think that human presence or some aspect will still be required in that process. On the other side, I think the future will create a new type of jobs where basically humans can work collaboratively with artificial intelligence, and again, make that insight that will win the game in the competitive market.
N: Yeah, I saw one research recently that said that in 2035, 85% of the jobs that are going to be at that point, don’t exist at this point. So I’m sure that AI can influence that greatly.
B: Absolutely. But it won’t happen. Even if we go back in history, a lot of jobs that were actually 50 years ago, now don’t exist. So yeah, I think that is natural.
N: Just speeding up with all the data.
B: At this point, we can just try to predict what will be going in the let’s say near future, but we’ll see what time will bring to us.
N: So let’s say a bit more on the AI subject. And we are talking about how AI can, or should, or would take people’s jobs, but can AI mess up big time since we are moving at record speed now with all the data, and leveraging that data and using AI for the data? Can AI mess up?
B: Yes, absolutely. So basically, when we talk about AI, and if you go more specifically, a good example is machine learning. Where you create some algorithms which can learn from the previous experience in the data. And then to provide some predictions about what’s going on in the future and what will happen in the future. So when you get outcomes from these data processing and analytics from machine learning, this is the best I can do, the best prediction I can do based on the previous experience and the quality of the algorithms that I use. But then at the end, maybe we can connect to the previous question, where is the place of the human here. In the end, humans are making decisions.
So, based on this data, now the human needs to process that and to decide whether I will go with that decision, I will make the opposite decision. So that’s basically definitely something where humans still have a very important place in that decision process, especially for the big decisions.
N: Whenever a human decision is in question, we need to ask the question of ethics. How can we look at the data and leverage data from the perspective of ethics?
B: Yeah, that’s a great question. So, we talk about machine learning, algorithms and everything. But now, there is a question of whether I made some business decisions based on the outcomes of the machine learning algorithms, even if that decision is unethical. So that’s the dilemma that many business executives have. And then, there is something that we call, for instance, algorithmic biasing. So there are the companies, the big companies, they have the data policies, and one of the statements in their data policy is data ethics with a period preventing algorithm biasing. And what does that mean in practice? So, let’s say, if you send the application for a job, you have that machine, you use machine learning to process the resumes of the candidates.
And that first selection is typically made by machine, not by a human, but you need to make sure that the algorithm that is used to make the selection of the candidates does not take into consideration, let’s say the age, or the gender of the candidate, or the nationality. So you’re committing actually, that you will not put that thing into a decision making factor in the algorithms for that. So that’s another question. Another important question is also privacy. With all this volume of data that you mentioned, in the beginning, we provide a lot of private information to the different kinds of platforms like social networks, applications for different products, loans and everything. So that data privacy question is also important because you never know how somebody will use this data. For just good reasons. And that is another aspect, which is now heavily regulated. We have the GDPR In Europe, you have the California Consumer Protection Act in the US. If companies want to use your private data, they must get consent from you for that. So that’s another aspect where basically, society tried to regulate and control that data craziness.
N: I always like to finish with the future and what is the future bringing to us. Today's most of our talk is about data and how we are leveraging data, how we are manipulating data, right? What is this going to look like in the future?
B: So in the future… The human brain is still perfect, right? Then we are not yet there. When we talk about artificial intelligence, this is still artificial, right? This is good enough maybe for the time where we live now and do the business, but for the future that will not be sufficient. As I mentioned at the beginning, data is rapidly changing. At the beginning of the data journey, we just had the structural data, then we had the data warehouses, then we had the big data, right? A few years ago, there was a big boom around big data. But now, when we talk today, we also know that big data is not the solution for all the problems that we have. At this point, one of the hot topics is Knowledge Graph. And this specifically what I’m working at Deloitte now, call it an initiative, where you’re trying to manage organized processes and similarly use the data as you do in the human brain. This will open other endless possibilities not just to manage data, but to use them in an effective way to get this data insight and to try to simulate how we process the information in our brain.
N: Bojan, thank you so much for this conversation. I enjoyed it. And for you out there, next Thursday, next innovation talk and see you then. Subscribe in the meantime.
B: Thank you. It was a great pleasure.
Keep up with us:
