- SAP Managed Tags:
Image credit: flickr.com/applesnpearsau/
This morning I wanted to write this article but I didn’t know where to start, my mind was completely blank. Turning around in the living room, I said to my 12 years old daughter: “I need to write an article but I don’t know how to write it”.
Daughter: “What is it about?”
Dad: “Well, business intelligence and predictive analytics.”
Daughter: “Huh ?!? What are you talking about?”
Dad: “They are two ways to look at and use information” .
Daughter: “And what do you want to tell about them?”
Dad: “Hmm, that they are not the same thing.
Daughter: “OK, then just write that.”
Dad: “Actually it is a bit more complex than that: they are not the same thing but they somehow look alike and they can be used together.”
Daughter: “And what happens when you use them together?”
Dad: “You can do much more than if you were using them separately.”
Daughter: “Ok dad, so it is like the phone headset you have on your desk: if you just have headphones you can listen, if you just had a microphone you can talk, but when you put them together you can communicate, that’s more than listening and talking, right?”
Dad: “Right!”
Daughter: “Good, now you know what to write. When you are finished, can you prepare lunch?”
Dad: “Wait a second! Help me out and you’ll have the best spaghetti you ever ate.”
Daughter: “It’s a deal!”, sitting down on the desk, next to the laptop.
Dad: “You don’t know anything about business intelligence and predictive analytics, ask me some questions.”
Daughter: “Well, start by telling me what they are, and please don’t make it difficult. And keep it short.”
Dad: “Ok, business intelligence (BI) is the art of looking at your data to understand how your business is doing.”
Daughter: “So you can make more money.”
Dad: “Well, you can make more money, save people lives, build an airplane: it depends on what your business is. You look at the information you have gathered and try to see if there are opportunities or if you are facing some risks.”
Daughter: “Such as?”
Dad: “Hmm, if you don’t sell a product anymore there could be a risk; if you see that a medicament is working well on some kind of patients, there could be an opportunity, this kind of things…”
Daughter: “And how do you see that?”
Dad: “You take your data and visualize it to show you how things are going. You present the information in a way that you can quickly understand and see if there is something interesting.”
Daughter: “OK and the other thing you talked about?”
Dad: “Right. So I was talking about ‘predictive analytics’. The idea there is that you have the data, a lot of data about the past and the present. You might also have exotic information you don’t even know why you stored it (tweets, weather, …), and you want to find if there are hidden rules which can help you take better decisions in the future.”
Daughter: “Hidden rules? What do you mean?“
Dad: “Well, they are hidden only because they are too difficult to see for a human being. A rule could be something like 85% of times customers purchases a new lamp, they also purchase a lightbulb”
Daughter: “It doesn’t sound so ‘advanced’ to me!”
Dad: “True, but imagine you have thousands of customers and thousands of different products. You might have millions of possible combinations! Some of those rules might be simple to see but most of them might be very complex and, well, hidden. If a computer does some math on the data, it will find all the possible rules for you. The computer becomes sort of ‘intelligent’ because it can use mathematical algorithms and other stuff such as probability models, machine learning methods, big data analysis…”
Daughter: “Ok, ok. So you got the rules, what do you do with them?”
Dad: “Great question, thanks for asking that!”
Daughter: “Dad, I am starving, let’s get it done!”
Dad: “Sorry, I think that there are two main things you can do. First of all you can apply the rules to new data and by doing that you can take decisions quickly. I call it prescriptive analytics.”
Daughter: “Simple and short, please! Give me an example.”
Dad: “You know who your most profitable customers are. You feed the list to the computer, which applies some predictive analytics to it and gives you back the rules which explain what is the typical profile of a best customer. You then apply the rules to potential new customers, and you immediately know who might be a good one for you. You can then just contact these potential customers and do not bother with the others. Practically the predictive analytics tell you what is best to do next.”
Daughter: “I got it. What was the second way to use the rules?”
Dad: “You just look at them! If you understand what the rules are saying, you can better understand your business and take strategic decisions. This is called exploratory analytics.“
Daughter: “So I look at the rules and they tell me what the best customers look like? “
Dad: “Exactly! A rule might say: ‘your best customers typically live in rural areas. This is good to know because you might then want to change the way you do your business by making sure that the deliveries of your product are rapid. Or you might take a completely opposite decision and decide to do more for your customers living in big cities to increase the business there. The nice thing is that predictive analytics can also tell you if it is better to satisfy the rural customers or to invest in urban customers!”
Daughter: “Don’t get to excited dad! It is just maths. Now you told me what business intelligence and predictive analytics are. So far so good. But you were telling me that they are not the same thing, why?”
Dad: “You were listening, that’s good! They are not the same thing for various reasons. First of all business intelligence information has to be used by human beings: humans need to read and analyze the information carefully to understand it. This means that we need to make it as simple as possible. We have only two eyes and a bit more than 1Kg of brain so we need to simplify the information by aggregating it, filtering it, reducing it to the minimum size that still carries some useful insight and is comprehensible by us, human beings.”
Daughter: “You lost me, but it is ok, let’s move to predictive analytics.”
Dad: “With predictive analytics you do exactly the opposite: you get as much data as possible, you get every detail and never aggregate anything because you might not know in advance for what purpose the data will be used.”
Daughter: “And I guess that the more details you have the bigger chance you have to find a relevant rule. So this means that you might actually create even more data than the one you started with?”
Dad: “Yes!”
Daughter: “So the first difference is that in business intelligence you simplify your data and in predictive analytics you make it more complex by adding details. What’s next?”
Dad: “Apart from the complexity of the data there is also the complexity of the task: with business intelligence we can answer some questions about what happened in the past. When it comes to understand why things happened or what might be happening in the future, that becomes too much for a human being. At least for me. So we use computer science, mathematics, machine learning to understand that for us.”
Daughter: “Got it. Now you also said that business intelligence and predictive analytics were similar, why so?”
Dad: “Because both deal with using data to improve your business. Make more money or save people’s life if you prefer. They might not be using the same technology or the same processes. People doing business intelligence or predictive analytics might be different or have different skills. But both of them have a unique objective.”
Daughter: “Hmm if we think of the headphones and microphone, they both are used to transport sound, but they do it in different ways.”
Dad: “Yes, more or less.”
Daughter: “It is almost 12.30 PM, now the final question: why do they do much more if they are used together?”
Dad: “Because they are very complementary. Cheese is great. Pears are great. But cheese and pears is divine food. Well, at least your Italian grandparents said so!”
Daughter: “I don’t like cheese.”
Dad: “Anyway, with business intelligence you look at your business today and you see if there are problems or opportunities and that’s great but how do you know what to do?”
Daughter: “I guess you are going to tell me.”
Dad: “You use predictive analytics to get a recommendation on the best possible action to take! Think of this: in your car you have a dashboard telling you how much gas you have left. An indicator becomes red if you are low on gas. This is BI. You know that you quickly need to refill, you’ve identified the problem. At this point you can ask your GPS to drive you to the nearest gas station. This is predictive analytics. With business intelligence and predictive analytics you understand where you are, where you want to go and you are told how to get there!”
Daughter: “Dad calm down. Now let me sum up so that you get the points for your article and then let’s have lunch, ok?”
Dad: “OK I’ll shut up.”
Daughter: “Business intelligence is targeting human beings. It requires to keep the information simple and to simplify it if it is complex. With it you can understand the status of your business today and see if there are places where you can improve it. In general you get to a point where you want to change something in the future but you have to decide yourself what you want to do.”
Daughter: “Predictive analytics tell you how to improve your business in the places you identified with business intelligence. But predictive analytics is more complex. You need more details in your data and you have to do complex calculations. The complexity is dealt with by ‘intelligent’ machines which do this work for us because they use advanced mathematics. At the end, predictive analytics can tell you exactly what you have to do to in the future.”
Dad: “…”
Daughter: “The two things are complementary, they share the common goal of improving your business (or saving lives, I prefer) starting from the information you have. But at the same time they are different because of the complexity and of the skills you need to use them.”
Dad: “Perfect!”
Daughter: “Have you got your article now?”
Dad: “Yes!”
Daughter: “A deal is a deal!?”
Dad: “Put the dishes on the table while I prepare the tomato sauce for the spaghetti!”
