[MUSIC] Hello, everybody. Welcome back to the last and final session of this introductory course. Today the topic and focus is digital media, so without any further ado, let me basically try to motivate the pervasiveness and the power of this media. After a 3 year gap, basically Adele releases Hello on YouTube, so this is her hit single. By 24th of October 2015, and this is one day away, the video gets something like 27.7 million hits. 27 million hits within 24 hours. Fast forward a couple of months, by the 6th of December 2015 that number jumps to 602 million. To get you an idea of how big that number is, it is bigger than basically the entire population of Europe, in some sense. Over half of those views come from mobile. So this is kind of starting to get interesting now, right? Over half of those views came from mobile and sales search, too. So in some sense people, I want to say, all that is fine, you put it out there for free. Did it help sales? Actually, it did. 1.1 million downloads in the very first week. Where is all this going? Well, it's the mobile thing that is of great interest to me, all right? So in many places mobile is, there it is. So that, as you can see, was taken about a month ago, 1.5 billion hits by then, in 7 months. So you can already see mixes coming up. You can already see fans making their own. Well, there's going to be parodies and there's going to be fan fiction. Fan crowdsourced commentary, and so on, on anything like this. It's part of the nature of digital social media. All right, now in many places, mobile is and will continue to be the only way to access the web. And what does that imply? Well, by Feb 2015, the number of mobile devices on the planet exceeded 7.2 billion, which basically was more than the number of humans on the planet now, just to put that in perspective. This includes all the mobiles ever produced. A lot of them, right, the older ones were probably junked already. But still, that number was still an important milestone. All right, some more facts. For over half the internet population, mobile is the only way to go online. Something which we have eluded to earlier, as well. And the primary usage of mobile service? Any guesses? It's basically video watching. That is the number one activity on mobile online. 6 billion hours of YouTube watched a month, 3 hours per Internet user per month. So that's more than half of those hours are on mobile. And that in some sense makes things very interesting indeed. Right, an average human pulls out his or her mobile a whopping 150 times a day on the average, right? So that works out to about once every 10 minutes or 9.6 minutes. Now because of mobile, we are moving from in some sense going online to practically living online. Being online will be the default mode of sorts, okay? All right, let me in some sense proceed to a couple of conceptual preliminaries on this, on digital social technologies. Quick question, what are social technologies? Can you list some examples? Think about it, type what you feel is the answer to that. Nothing profound, something simple, yeah. We can basically collect and collate what's happened there. Technology that enables social interactions at scale is a social technology, right? So Facebook is a social network, yes. A bulletin board, online bulletin boards, Usenet forums back in the day were also in some sense, are also social technologies, all right? So everything from online bulletin boards to social networking sites have a social component to them and are digital in nature and hence we can call them social digital technologies. Digital, social, it doesn't matter. All right, which brings up the question, what is the social part of social technologies? What does it mean that something is social? If you think about it, the social part of social technology comes about because of this. Any event or any transaction using a social tech is not strictly private anymore. It is relatively public, which is why email is not a social technology. It goes from A to B. Relatively public and hence, because it is relatively public, it is amenable to social influence, right? In terms of information flows, in terms of expectation setting, in terms of building reputational capital, building social capital, you name it, all right? A good example would be online marketplaces where I can go and see what other people are doing, how other people value other items, a lot of these things. And they influence my own estimation of the value of things out there. Going wide along the line requires a social component to resonate very well. So for instance, a lot of companies want a lot of things to go viral, right? It doesn't always work out. Well, the social resonance has to be there in some sense. So that is so much with the preliminaries. Let me in some sense, well before I go on give you a quick around of what is going on in terms of digital advertising. Going online, what you see out there and this is a source from The Economist, you can basically see that among the television, newspapers, magazines, radios, advertising channels or outlets available out there. It is the internet which is rising at the fastest clip. Okay, there was a brief dip in 2009, but since then it's been up all the way and getting more social, too. So within that Internet advertising the proportion of dollars going into social technologies are just so much higher as well, and that's clearly seen in the second graph. Evolution of digital media, evolution of the early web. So let me, in some sense, take you to where everything is coming from, and maybe we will get some sense of where things are going to, right? So both of them coming together. Okay, in the mid-90s the Internet was perceived to be very different from traditional media. Okay, this is the early to mid 90s. The Internet had just come about and people knew that this was somehow different from traditional newspapers and magazines, in what way, all right? In terms of consumer profiles, the people who read the Internet were different from the people who read the newspapers. And because the consumer profiles were different, consumption opportunities were different. And because consumption opportunities were different and profiles were different, consumption choices were different. Consumption patterns were different. Consumption experiences were different, enabled partly by the medium itself. There was interactivity, there was some degree of personalization. And yes, because of interactivity there was also a social dimension coming into the picture slowly, right? So if I put a comment on a news article online, other people could see my comment and maybe respond to that. So there was a social dimension, too. All right, so let us in some sense look at the history of the web. Back in the year 1991, there was a grand total of 1 website. By the year 1992, that number jumped to 10, 900% increase. By the time we came to 1993, that number jumped further to 130, and so on. A 1000% increase on the average, which actually maintained itself throughout the golden decade of the 90s. By 1991, 99, we had 1 million websites. And by 2000, 17 million. Which brings up this question of this manic proliferation of websites, what implications did it have for media consumption, and then we will get into digital media. Of course, this was the beginning of digital media, what happened, what were the implications in traditional media? But what we do see is that per capita media consumption increased. The average person was consuming more media than before. There was this proliferation of content available. However, per medium the average consumption crashed. Maybe relevance trumps reach, all right? Previously there was not enough choice and people would go with whatever was available. Reach won. Now because there was choice available, relevance became more important. And we will see this reach, relevance trade off play out in the next few slides as well. All right, so after the decade of the 90s came the early 2000s. And in the early 2000s laptops emerged, which basically meant portability. Not true mobility though, portability. The genius of Steve Jobs was that while everyone saw wireless networks as designed for phone calls, wires basically, Jobs actually put media first. To him, that particular device, the then basic feature phone, was actually the harbinger of potentially of an awesome media device, right? So he basically said, the first iPhone, okay, the first smartphone, launched less than ten years ago, June 2007. He called it an insanely great media device. Voice was secondary there. Why? Well, we can see why and we will see it soon enough, right? And just when media disruption couldn't get worse, guess what happened? Social media exploded on to the scene. Implications, what happens now? Take a minute, think through it, write down. What happens because of the emergence of mobile and social media together? What do you think is going to happen to the media landscape in general? How is it affected? What is now possible that wasn't possible before? Take a couple of minutes, type down your answers and let's see how that goes. Okay, what social media in some sense brought forward was the fact that in people there appeared to be an innate desire to share. Of course, academics still debate about it. But it did seem that when given the opportunity, people were willing to tell the world about themselves. Media consumption and creation patterns change, especially when combined with true mobility. So let me in some sense, and true mobility that is brought about by the smart phone, all right? This insanely great truly mobile media device. Sample some quotes. YouTube really gave media and advertising industry the first inkling that privacy concerns aside, people really wanted to tell everyone, not just their friends and families, about themselves. Why was this such a big deal? Because for years before that marketers had spent years trying to figure out how to build profiles of users and their online habits. And just at that time, users decide to volunteer all this information. Here's an example, 300 million pics every single day on Facebook. Over 1 billion posts every day on Facebook, right? So mobile devices fueled this fire by basically making it easy to take pictures, sharing them, whenever, wherever, anywhere, anytime kind of setting that happened, right? Now let me come back to the reach versus relevance in some sense trade off, okay? How now can we get better targeting for more relevance and therefore more engagement and therefore better outcomes, all right? So that's basically where we will try to go. Setting the context, digital advertising. Consider this to be a three player game, there are three players here. One, you have marketers and advertisers who have to serve, they have something to sell. Two, you have publishers with inventory. Who is a publisher? Anyone. So New York Times has a web site, and people come to the website. The New York Times becomes a publisher. Facebook is a publisher, people come there, right? So they have inventory, they have space which they can use for advertising, perhaps. And then there is the third party, which is audience, which may pay attention if something really interesting, and more importantly, relevant comes along, which may does not guarantee they will. The marketers may say to publishers, I want to reach people in the market in tier one cities who are in the market for a Sudan. Or I want to reach hiking enthusiasts in. Or I want to target test takers in India, or South Korea, or Turkey. Doesn't matter. Basically marketers will come to publishers with, in some sense hopes like these, demands like these. This is the target segment that I am looking for. Which brings up the question, how might publishers offer such customized segments to marketers? Speculate a list in a few ways. Concomitantly, what type of publishers are more likely to succeed here? Write and submit your answers, take a short while. And do so for both these questions, before we proceed to the next slide. Which brings me to the next slide, Setting the Context in Digital Advertising. The reach versus relevance tradeoff in advertising. Now examples of marketers trying to elicit engagement. So there are reach based, right, for instance, creative means, which are intrusive ad formats. So you have pageskins, you have interstitials, you have takeovers, which will take over the screen unless you get them to go away, and so on. And then there are relevance based native advertising where the ad content cannot be stored distantly from the rest of the site content, right? Content marketing, wider marketing, personalization harder to pull off but the ROI is just so much higher in these cases, right? So what does relevance mean anyway? How might media in firms and publishers sell relevance? Relevance basically implies some sort of an alignment with customer's needs/preferences and interests, okay? Without this alignment, relevance does not even enter the picture. Now hope is a customer's demographic and web-surfing profile correlates with their preferences and interests, right? Now the web-surfing profile and demographics are easily observable variables. Preferences and needs, not so much, okay? If there is a correlation, then maybe you can get some relevance out of it, okay? The challenge then, right, would be to micro-segment a population for targeting, re-targeting, and so on. If the segment is fine enough, personalized ads create, basically it can be created and served to the right people at the right time. The definition of relevance, to the right person at the right time, the right ad in some sense, right? Figuring out the right person and the right time for the right ad, is the challenge in some sense going forward. Let me give you a simple example of a dynamic advertising campaign. At its simplest, dynamic ad campaigns require data signals which come from APIs and a set of event rules or what I'll call trigger conditions. So I'll give you a simple illustrative example that will make this point. Look at this, okay? Now this appeared on the screens of office workers in downtown London around Westminster, okay? It's lunch time 11:34. Okay, so basically it is using the computer's own system clock to give it that. It's not a regular ad, it's basically time sensitive. It will only come when the event rule says around lunch time launch. Two, 10% off for today's lunch. Basically put in your pin code and I will direct you to your nearest Starbucks. All of that, and you can get to print a coupon. So basically what it is doing is it is trying to build a greater relevance. For instance, one could imagine this happening. Eight groups of four or more get 15% instead off instead of 10% off. I have to figure out if you are a tennis buff, hey, there's a shot up of Venus William's match happening. Why don't you catch it at Starbucks? All of that is possible once we know in some sense there are these weighty data sources and I have some even truths and I know something about you. And I put it all together to craft a relevant message. Take another example, here are some brands. Audi, which is car brand, Lipitor, which is basically cholesterol medication on prescription in the US. You have L'Oreal, which is a cosmetics brand, Wells Fargo, which is a bank, yeah, and Starbucks, all right? Some of the data signals from some of the APIs that the ad makers could in some sense get access to, could be weather, profile of the consumer, sports, entertainment, geography, and so on. And you can in some sense fill this up in a lot of various ways. The Audi guys might want to know weather information. They definitely might want to know about profile information, sports information. The Lipitor guys are not so much concerned about whether our cholesterol is there, whether it's winter or summer. The need for it doesn't go away with the weather. Social may or may not actually cholesterol, Lipitor probably doesn't in some sense need social information, and so on. Yeah, sports, so does it help that you are a sports fan? Does it help sell in some sense some particular? Yes, Lipitor apparently yes. And now you can play sports without having to worry about cholesterol. So in some sense there is a relevance created and creative can put together an ad based on that basis, all right? Judgment is called for, there is in some sense there are no set rules here, different cultural context, different things. What events might trigger which ad copies to get served to whom? That is the question and so basically we have some idea and we want to build a rules engine Here now. So here's an example, take Nike, which is advertising shoes. Again, very loose trigger, but just to get you an idea. Data signal, weather and city, suppose it is only taking those two data signals, weather and city. Trigger condition says if the weather is rainy and the city is Seattle, you serve this message. And what is the message? Hey, Seattle, it's been pouring, time to hit the gym. There is some relevance because it really is pouring outside. And the image would be some animated gym scene and so on. Data signal, weather, city if the weather is rainy the city is Seattle you could, or the weather is warm or cold basically. And you could say hey Seattle it's freezing, warm up. Yeah, so in some sense it is giving you a message based on weather and city some combination, and you can have a different image for that. Weather is hot, city Seattle, hey Seattle, it's a sunny day. Why don't you go out for a run or something? So basically the point being the ad is now being crafted. There is some degree of personalization. There is this hope that it will be more relevant than a blind ad served blind, which is not going to get clicked in two days, DNH. So marketers write these event rules and creatives build them as agent copy and all of it can be put together, and you can in some sense create a large number of possibilities. So I can mix in image, I can mix in sound, I can mix in city. I can build on the fly dynamically, construct an ad that did not exist before by joining together components. Basically what is happening in digital media makes it possible. Mix and match, modularize it, mix and match it, dynamically create it, serve it on the fly. Which brings me to, in some sense, digital advertising, data signals, and brand. [MUSIC]