Master Data Management Motivation

One of the main problems that we face during the process of data analyses is the lack of a common view inside a corporation and its market. We have some boundaries to cross involving operational perspective of data and the others perspectives: tactical and strategical. Here is a very good and simple explanation of these perspectives.

Imagine that you have to balance these three perspectives and deal with the misunderstood of all the business concepts that surrounds your company since in this equation you have a very complex "X": the human mind. Ralph Kimball told a long time ago that one of the most annoying thing to a CEO is when two employees in a meeting talk about a concept, for example, the churn rate and both have different numbers, this is one of the oldest Business Intelligence´s problem, and for a lot of companies it isn´t solved yet.

Ok, now it is becoming discouraging but nothing is so bad that can´t became even worse. Do you know how much data you have available in your company (I know that the information you have isn´t enough and you need even more)? Considering that you are in a global market, how much data do we have over the internet? And how about over the world? I know it could be more than we need but it is a good input to this problem.

Digging through some answer about the last question I found a very good research conducted some years ago by IDC sponsored by EMC with this subject. According to this research we hit 601 exabytes  of data this year and we are growing 35-50% per year. Now put all this issues together and try to organize this! Do you think it is impossible?

This opportunity opens the way to the Master Data Management. All this information that must be loaded, integrated, managed and shared among the company, creating values, prediction trends and showing where we are in terms of our business keys, this is the Master Data Management´s promise but it isn´t that easy.

I will explore this subject much more in the future, but here there is one definition about it:

"If (either deliberate or inadvertent) replication of these “key business information objects” impedes the achievement of organizational business objectives, then creating a unified view of “master data” will reduce barriers to success. We can define master data as the key business information objects used in the different applications across the organization, along with their associated metadata, attributes, definitions, roles, connections and taxonomies. Typical examples of master data include Customers, Suppliers, Parts, Products, Employees and Locations."

I extracted this definition from a very good guide that you should read for detailed information. If want some more quickly overview try this video:

As you see in this video that are some areas to focus: people, process, tools and business, and I will do in the next posts!

What Side of the Hustle?

Some time ago I watched one of those tradicional movies about people who instead of gamble in a cassino use a systematic way, or if you prefer cheating, to beat and win a lot of money. The movie that I am talking about is the 21 and in this movie some guys from MIT with their teacher created a system of counting cards and starting to act following some roles where some of them loose some money but other members of this group  wins a lot.

Ok, this is an old fashion Vegas movie script I confess but if you go further there is a valuable talk involving a man that´s probably the cassino owner and the chief of the security (interpreted by Laurence Fishburne) called Cole Willians. Cole is an old school professional that knows how to count cards to win in blackjack and while he watches the monitors he identify this group cheating, he starts a conversation with the cassino owner (the talk that happens in minute 2):

"Let's just say I like being on your side of the hustle". That is the point. This cassino assumes that if someone who is analysing some data is cheating but in the other side it is also analysing a lot of data about his customers to create a "sistematic way of" rise his profits.

One example of this that I extract from Super Crunchers is about the Harrah´s cassinos. They knows exactly how much money you can loose in their cassino in one night and still come back. They call this amount of money of the Pain Point and when a custumer is almost hitting this point they send you a "luck ambassador" and offers something "for free", e.g. a dinner, to force you stop playing in the right time and everybody have a happy "end".

This is just one example of a company that is thinking with numbers instead of having only the "old school" skills and no matter in what side of the market you are, you should also do this. The movie also talk about old companies that did not upgraded its way of work and because of that are loosing a lot of customers, this is really happening and the main dangerous trap is that if you aren´t counting your cards probably when you discover it, that your customers had gone, it will be late.

Calling your Army

One of the main discussions that appears when we are talking about the power of data analyses is: What is more important to have success in bussiness today, specialists or super crunchers? And my answer is that you need both of them!

There are a lot of skills and tools that are necessary to develop a good strategy in a company, one of them is of course having a database to be used as a place where you can centralize all kinds of information. This database can be a database marketing, a data warehouse or some data marts but it can also begin with some simple relational database. The fact is that you have to build an infrastructure where you can storage all the information you have because one day it will be necessary to learn and predict what you can do and what you cannot do.

Ok, that´s the easy part, but the question remains the same, who or what is more important today? Until now we didn´t talk about people, so let´s start. There is a brazilian psychologist that works with this subject, called Waldez Ludwig, who says that if you want to become a market leader you must focus on your employees because considering the competitions that you have today in market it is not enough to have the best product or services anymore, as we can see a lot of famous companies closing their doors or asking for governamental help. 

I found a very good presentation of Andrew Zolli about this subject:

Everything is so accessable and open today that one of the last choices you have is to invest in innovation, and innovation is equal to people because machines still don´t create nothing by their own "hands", so people are the answer, but is it enough?

No, it isn´t enough. When you go to a war you need to carry all kind of weapons you can because a good strategy is to attack in different ways depending on where you are fighting and who are your enemies. Data analyses is one more weapon or in my opinion your bullets, because no matter what weapon or strategy you choose it is barely impossible to achieve your goals without bullets.

First I have said that people are the key, and then I said that data analyses are fundamental to win, but what if we confront one to another? Once again I will use some examples from the Super Crunchers book where that are a lot of examples where mining some data with other techniques had win the battle but this kind of approach even if it correct is open, so once it became more popular everybody will have probably the same bullets and again people can be the key.

Following Dan Pink in A Whole New Mind the right brainers will be who will make all the difference in the future, this kind of people is who can think out of the box, without a norrow focus and they are harder to find, in the other hand, or better, in the other half of the brain, the left brainers who can do the manual and less thinking jobs is easy to be outsourced and cheaper.

Putting all pieces together just like every super cruncher loves to do the best approach to try this is to invest in the infrastructure to do the data analyses, data mining, predictions, etc. maybe even outsourcing some non-critical components of this, but just to support who will work above this game. Pay attention that without this infrastructure your are loosing for 1x0 but if you have this, the game is 1x1, so to win you must make the difference, and the difference are people, at least for now.

Crunching True

A common question that is in a lot of minds today is how can someone obtain an advantage doing  data analyses, you can call this data mining or in other way, maybe in a deeper one, just like Ian Ayres says becoming a super crunchers. Once I´m intent to reveal things that aren´t so public yet  my first post will be about this question, who knows with some answers.

That are many examples of companies that uses information to achieve market advantage and in some cases a unique kind of service to grow its market share. I can mention companies that are already doing this like Amazon does with the session "Frequently Bought Together" or even "Customers Who Bought This Item Also Bought" or companies that has the potential of doing this like Google (you can see more examples in Super Crunchers).

Today I won´t discuss about the beliefs and faiths in this subject (that are a lot of pros and cons) but I will talk about how to start doing this, massive data analyses in your bussiness. The first thing is: It´s not something that you can do overnight. Selecting data digging trough hidden truths is something that you must first design what you planning to discover with this, all the companies that have successfull implementing this didn´t start its business with this subject in mind, they had its main business, its products, its services and saw an opportunity of start analysing all the data they captured during the time.

Once you already have you established business it is easier to start, and as a second advice you should always think in how can you get more data during the lifecycle of your product/service, never is enough when you think of data, get as much as you can. It is also important to identifiy any gaps that you have to fulfill your statiscal models, the data and truths around it are always changing so you must change too.

My final suggestion here is to give some steps backward and try the see the whole scene, try to imagine what data or information usefull are missing, what services you can build to aggregate more data, qualify them, something like a master data management and don´t get yourself frustrate, first look can be deceiving and nothing is so hard that you can´t give it a try, just like I am doing right now.

As you can wonder this subject is far from finish, but that is all for today folks!