So the last part of the lecture I want to talk about today are some really cool applications of Advanced Management and Marketing Science by Leading Firms. And these are firms that'll bring all five aspects together. They're using better data, better exploratory methods, better predictive methods. They're using better optimization to actually make business decisions that influence the products and services they sell. So let me start with the first one. Many of you may know Kohl's department store, it's a very large department store national chain department store and kind of the low to medium end department store. And they're doing what's called, and by the way remember, from my earlier slide this isn't Google, Amazon ir Facebook, this is a brick and mortar retailer, Kohl's, doing the following. They're doing what's called smartphone targeting. So for example, they have data on your geo spacial location when you walk in the store and you might say, well how do they have that. Well if you have your wi-fi turned on on your cell phone, the minute you walk into the store and you pick up the wi-fi network in that store, it knows your geospatial location. So they're using their wi-fi network to know your geospatial location. Everyone's cell phone has what's called a fixed IP address, so they can know it's you. They can now link that to what you've done if you've gone onto their website, Kohls.com. Let's say, they know that Eric Bradlow was standing in front of the shoe aisle. They can now actually send me a real time discount for shoes, whether through text, or through, it can even be a phone call. They can actually send me a real time discount, because I'm standing in front of the shoe aisle. Let's now review this in terms of my five forces if you'd like. Why can Kohl's do this? They have data that it's me through my phone and the wi-fi network, they may have link this hopefully to my behavior online and possibly through their data network in online and offline in the store. They know physically where I'm standing, and now the action they're taking is to send me a targeted, or contextual discount, given my physical location. This is extremely valuable data. There's the old adage in marketing, it's not just selling the person the right product. It's the right product at the right place at the right time. Well Kohl's is taking advantage of this and saying, what's the better time to send Eric Bradlow a men's shoe discount when he's standing right in front of the men's shoe aisle? It's not valuable 30 minutes before I get to the store. It's not as valuable 30 minutes after. It's valuable right when I'm standing there. So this is an example of a company that's recognizing they can collect better data. They can use that data for decision making and they're going to operationlize against it. Here's another one, Netflix. I think so many of us have become huge fans of the content that Netflix is creating. As a matter of fact, for those of you that don't know, by 2020 there will be more video content consumed on Netflix than any other provider. That includes YouTube, that includes any broadcast like nbc.com. Think about it, Netflix. The US population will be watching more content on Netflix than any other place in the world. That's amazing. Now most people think, well, Netflix is getting lucky in creating content. Not so fast, that's not true. So let me tell you what Netflix is doing. Netflix is doing what's called meta tagging data. Meaning, of course, when you log on to Netflix, they know what you watch. This is the ultimate in customer analytics. They can measure customer by customer what it is you're watching. But, here's what there also doing. Every show you watch gets what's called meta tags, or if you'd like, attributes or descriptors. So, they know if Eric Bradlow watched a show, a police show that takes place in the 1970s in a warm weather city. Imagine having that corpus of data from millions and millions of customers. Now, rather than saying, what show could we create? Now imagine the director saying they're saying, I see what the data is telling me, people really like police shows that take place in warm weather cities, you know, in the 1970s. Hey, let's create a police show from warm weather in the 1970s. And so what companies like Netflix are doing is, they're using data mining and customer analytic methods to create content. And actually, this reminds me of a project I worked on ten years ago was how to optimally design ads using attributes of which ads were successful. And I remember getting a rude awakening, this is maybe an idea before its time, I remember going to all the different ad agencies and saying you know what I know you use art to create ads, I've got a scientific way for you to do it. I know how much music you should have in an ad, I know whether you should have dogs in an ad, I know whether you should have kids in an ad. I thought they would embrace me like, Eric Bradlow, you're the Messiah. Well, they weren't ready for science and art. They were only ready for art. I'm hoping things that what Netflix are doing can bring more science to the problem of creation of content. Another example, American Express. Obviously, one of the big problems American Express faces today, is Churn modelling. They want to know who's going to give up their American Express card, why? Well, one of the drivers that you heard about in the other lectures of this marketing content was the idea of customer lifetime value. I don't think I need to respeak but I repeat briefly. Churn is a big part of customer lifetime value. If someone churns from American Express, American Express makes no revenue from them after they've churned. So you say well, what does American Express need analytics for? Don't they have, you apply for an American Express card, you fill out a lot of data. What's the problem? Well, there's no problem, except what American Express has found is that your social network data Is a very strong predictor of whether or not you're going to churn. So what American Express is doing, which is what many firms are doing today, they're scraping. This is all legal, they're legally scraping data from the World Wide Web, of let's say Eric Bradlow. Like what did I say on Facebook, and how many friends do I have, and what photos of my posting on Flickr and all this other kind of stuff, and they're taking that data and they're adding that as extra variables and predicting whether I'm going to churn. For example, if I posted on Facebook today aw man, I'm broke and I just lost my job. That's probably pretty predictive of whether I'm going to churn or maybe not pay back my American Express card. Probably pretty valuable for American Express to know that. So this brings to bear a lot of issues. Better data, the company can manage and collect that data, they can quantify that data, meaning they're using what are called natural language processing techniques to take the textual data from the stuff I posted on the web and they're turning that into numerical data that they can feed into a numerical term model. That's remarkable to me, how far this has come. And again, not only are companies doing this but you should know this as a customer when you post stuff on the web, you should know that companies are scraping this. As a matter of fact, it's not just American's Express is doing this, the next person that's thinking about hiring you is scraping information from you about your social media usage, and they're deciding whether you meet their standards or not, based on this. So this is very valuable data that's now part of the analytics arena. The next example is an healthcare. Well, I'm not an expert in healthcare, I am an expert in analytics. And the number one problem in healthcare today, the two big problems in healthcare today is number one when it sees in the right patient adherence. So how do you get patients to take the medications that are prescribed to them? That's one big challenge. The second is what's called predictive analytics, and that's what I'm showing you in the bottom before. If you think about the way medicine works today and why it costs so much to the government and to the planet is because we wait until you get sick and then we treat you in the optimal way. But imagine I could predict from your medical records, which is the picture I'm showing you on the left. Imagine I could predict the illnesses you're going to get 20 years before you get them. Now, instead of trying to solve the problem by treating you optimally, I try to prevent you from getting this disease optimally. So imagine I knew that Eric Brodlow was going to have high blood pressure 20 years from now. Why don't I start taking the drug right now that prevents me from getting high blood pressure. And now I've turned a reactive problem into a proactive problem, and this is the future of analytics in health care today. It's trying to predict people's diseases far into the future, but if you think about what you've been talking about, what you've watched me talk about and my colleagues talk about, this is analytics. This is marketing analytics. I'm going to target individual customers based on their patient records, based on what drugs they've taken, based on their family history, based on their consumption of goods and services. I'm going to target them optimally, for the right way to make them adhere to drugs, and with the right protocol to prevent the future illnesses. This isn't backward looking, it's forward looking. This is maybe one of my favourite ones which is Google free taxi. If you haven't heard, Google has been working for the last five years on driverless cars. Now you might say, wow, that's kind of interesting. That would be kind of cool. Yeah, but here's an opportunity that they're thinking about to monetize it. So, what you see in the left is just an example of purchase history data. What you see in the right, is a formula for computing the customer lifetime value. Imagine Google doing the following. I'm just picking Bloomingdale's as an example. Imagine I go onto Bloomingdales.com and I start saying, I like this and I like this, and I start thinking about all kinds of stuff I want to buy. Now Bloomingdales can measure this and now they can say, hey wait a second, it's Bradlow, Bradlow's back he's a very valuable customer. Imagine Google partners with Bloomingdales and says, wait a second. I understand Bradlow buying online is valuable, but image we sent a driverless car over to Eric Bradlow's house to pick him up and drive him to Bloomingdale's for free. So imagine Google partnering with retailers to actually offer a free ride to a store based on your customer lifetime value. This may be, if you like, the Mount Rushmore of all analytics. It's Google, who kind of knows what you're doing, partnering with a brick and mortar or an online retailer to send an actual car to your home based on customer lifetime value. That has all the aspects of data capture, kind of analysis, prediction, and action. Which is physically sending the car to your house. If you think, by the way, that this is fantasy, and this is never coming, it's coming. In the next ten years, I guarantee you that not only will there be driverless cars on the road, but there will be driverless cars in the road that are tied to analytics. Starbucks. Starbucks is actually a much more brilliant firm around analytics than people give them credit for. And if you see here, I've given you these two pictures which seem like an oxymoron like they don't go together like customer loyalty equals no deal. What do you mean customer loyalty equals no deal? Give the best deals to my better customers, right? No that's not right. The people you should give the best deals to are the people that giving them the deal changes their behaviour. So let's imagine Eric Bradlow stops at eight o'clock in the morning at Starbucks every single day. Do you think if they give me a deal, I'm going to stop at Starbucks? I mean how can I stop there more than every single day? Maybe they could give me a deal to get larger wallet share. Maybe instead of getting my muffin at work and my coffee at Starbucks, I get both at Starbucks. But you don't give deals to your best customer. You give deals to your customers for which ROI is the highest. And Starbucks recognizes this through their loyalty program. They want to give deals to the people right on the brink of being loyal. They want to turn disloyal or infrequent customers into frequent ones, they're not trying to turn frequent customers into frequent customers, because frequent customers are already frequent customers. So they've taken analytics to the next level where they recognize you don't just treat your highest revenue customers the best, you treat your customers the best for which the ROI of that expenditure is the highest. And by the way, that's typically not the customers who buy the most from you. People get this wrong all the time. High value customers are already loyal, they already buy a lot and you probably get a lot of their dollar wallet share already. It's that interior customer, it's that midling customer that you want to tip towards a heavy buyer and not towards a light buyer, those are the ones you should target and Starbucks knows this. The next one, call centers. How many of you, a lot of you may not want to admit it, and unfortunately you can't see me and I can't see you, but a lot of you may not want to admit this, but how many of you have ever called a company on the phone because you're having a bad experience with a product, and started screaming into the phone? Or started screaming at the person on the other line? I think every one of us has to be honest, and say we've done this. Now you might say, what does this have to do with analytics? Well, has everything to do with analytics, and how marketing and sales are done today. Imagine Eric Bradlow calls on the phone. Let's imagine I call Comcast, my local provider, and I'm not saying anything. I mean, I like Comcast. They're a good cable provider, but, lots of people call up that they're upset with Comcast. I'm not picking on Comcast. They know they have a, if you like, a customer satisfaction issue at Comcast. Let's imagine I call up at Comcast and I start screaming into the phone. My cable service is out again, the picture's blurry. [SOUND] I start screaming at them. Well, two things. The first one's going to be not surprising to you. First of all Comcast knows it's me that's calling because I'm calling from my home. They can see I have ten TVs in my house, and by the way, I'm not exaggerating. I'm a TV family, I like TV. I have a lot of Comcast boxes which means I'm a very valuable customer to Comcast, so they can see that, notice already, look at the data. My phone is now link to my account. They can see it's me. They don't ask me it's me, they know it's me when I call and the person on and answering the phone knows my lifetime value. They can see how much I spent over the last, my average monthly purchase and they can see that I've been a costumer for 15 years. Second thing is, they can listen to the intonation software now, can listen to the anger in my voice. And decide and put a script up there for the person, hey, wait, Bradlow, angry, churn, not good gotta do something about this. And so here's an example where a firm has taken really natural language and intonation software, and merged it with CRM systems and database management, and they're merging these together. They could do one of two things. One is, the script could be for the person, wait, wait, calm down, Mr. Bradlow.. We understand you're angry, etc, etc. Or they could patch me through to a different person who's better at handling angry people. And that's even a greater possibility of using analytics. So here's an example of a firm, that's doing better customer relationship management call center routing, if you like, through the use of analytics. Amazon: Ship before you buy! This is kind of, again, one of my favorite ones. You see a picture here, again, of someone's purchase history and you can see a picture of a drone. No, I don't think Amazon is really going to have drones flying above my house ready to drop a package off. But what Amazon can do, and I mentioned this briefly in the introduction and a little bit in the summary, is what Amazon can do is they can predict what Eric Bradlow is going to buy in advance of me buying it, they can ship it to a local retailer near my house and so if I order it, I'm going to get it that day. Imagine the possibility and the customer lifetime value you can build by saying, if you order by noon, the product will be at your home today by five o'clock, think about what you be willing to pay for that. Now again, what allows Amazon to do this? Number one, they've got really good data, because they're tracking data the individual customer level. Two, they've got a recommendation engine and the ability to do predictive modeling. So they can predict what Eric Bradlow's going to buy in the future. Not what Eric Bradlow bought in the past. What Eric Bradlow's going to buy in the future. Number three, they have extensive distribution, so they can ship stuff not to my home, but to local retailers near my home. Number four, they have an ability to act and charge for this. So are they going to do it for Eric Bradlow for free? No, that'll be Amazon Prime Prime. Amazon Prime Prime, won't be free shipping, it might be free shipping, but here's an extra price you pay if you want it that day. And so imagine the power that analytics has brought to these kinds of companies. So just to wrap up and to talk to you about the ultimate takeaways here, what I've tried to talk to you about is, technology meets management and marketing science. So what you should be focused on if you're sitting here listening to this lecture, the first thing you should be thinking of is, technology's wonderful. It's cool. But this isn't the thing about your personal home use and devices. How can I use technology, whether it's eye tracking data, GPS data, web data, purchase data, survey data? How can I use this data to better understand my customer? The second, it's never the golden age of data. So if you're still making decisions today with store-level data or aggregate data, you're probably giving up a lot of money because there's data at a more granular level, that's going to allow you to make the same decision, but at the level of the individual customer. Third, if you think this is nice to know stuff, I'm not in the nice to know business. If I had a dollar for every study that was done that's nice to know, that's great. But I'm into problems, and the way I think of analytics is as towards action. And this is real monetization. Do you think Kohl's, Google, Amazon are doing this because this is nice to know? They're doing it because they see a real opportunity to make money. So I want to thank you for your attention. Again, go onto my website. Go onto the customer analytics website. Don't ever stop learning about analytics and how it can be applied to your particular business.
