!– Twitter Card data –> <!– Open Graph data –> <!– Schema.org markup for Google+ –>
As with any technological tool used in business, big data runs the risk of being more of a fad failure than a functional success. The line between fad and function is not clear or distinct, and is not exclusive to big data success. The same issues that deflated the expectation of business intelligence (BI) threaten the successful implementation of big data.
Frankly, these terms get thrown around so much without precision that the meanings of the two have morphed together. For my own clarity, I look at BI as an inward focus, looking at the data that already exists in my own company data systems. Big data is data from outside my existing systems. This is how many CIOs I speak to look at the difference between the terms.
If you read my previous articles, (go here to start) you know how much I value the art of asking the right questions. You might even get the impression that I am anti-data in making business decisions. However, I do believe in using data to make decisions, with proper judgment and the discipline to ask the right questions.
Asking the right questions is the first step in powerful analysis. Analysis is the missing step in any failed attempt to harness business intelligence and big data. Analysis can happen without any data, based on the observations and impressions of the analyst. Truly gifted analysts look beyond the realm of numbers, applying what they see and read, gathering sentiment, watching for trends not just in numeric data, but in the opinions and feelings of people. That is why polling numbers are so important in politics, and the political analyst looks deeply into the wording of the question to understand the meaning of the polling result.
I believe there are four vital components that any successful analytical effort depends on. This framework goes beyond big data, or business intelligence. This framework fits any analysis project. I call this a framework because I think of it as a solid object, something of substance that assures you that you are working with answers of substance. Imagine the basic components of a building; there is a floor, a roof, and four walls. There are different ways to build the building — different materials, different shapes, and differing details. Yet the building can’t be a building without these six basic components.
The word aimless comes to mind. The definition of aimless is “lacking a definite plan, purpose, or pattern.” A journey can be aimless if you do not have a destination in mind. An analysis project can be aimless if there is no defined goal when the project starts. While there are fantastic stories of aimless research discovering something huge, I can’t seem to remember one that was actually true. Every discovery is the product of someone searching for something, so every successful analytical project starts out with a clear goal.
People who discover cool stuff are not stupid people. They may be awkward. They may be silly. But they are not stupid. Great analysts have an intense curiosity; they want to figure out how stuff works. They come from many different backgrounds, many different places. If you are paying attention, you can spot the analysts when they are kids; look for the kids who figure out the games the fastest, take things apart, or point to the most obscure details they see out the car window. If you want to create great analysis, you must have extremely curious people.
Some of you may think that this is painfully obvious. OK, so tell me this: if this is such a painfully obvious statement, why do so many people waste time looking at the wrong data to answer the right questions? If my question is, “Why do people stay away from my store?” should we look at our sales data? What do you think?
If you said “Yes,” then you are looking at the wrong data. The sales data will not give us any insight into why the customers do not come in to buy.
This is business, not a twisted view of science. There is an underlying idea that data analytics is like throwing spaghetti against the wall to see what sticks. No successful analysis works that way. Scientists build their research to ensure that results are repeatable, eventually building to a predictable outcome. Data analysis models must create results that support convincing outcomes.
In business, analysis models must measure success using the same metrics and optics that leaders use. Telling a CFO that the model indicates the change will improve vendor lead-time variability by 10 points is meaningless. However, saying the model predicts $150 million in potential inventory reduction is speaking in the language of business.
It is one thing to have a team of analytical savants able to tease meaning from data. Think of the power of an entire company armed with the tools to see the information and the ability to convert the knowledge into action.
An organization with the capability to see opportunities and exploit them is powerful. The transformation requires the core analysis team to develop relevant analytics and measurements and create simple tools that integrate analytics into front-line processes. The next step in the transformation is teaching execution managers and employees not only how to use the tools, but how to exploit opportunities to learn more through more localized analysis. Building analytical capability across the organization is building knowledge.
Would you buy an umbrella and then go walking in the rain carrying it in your hand unopened? No, you would not. Would you invest in an analysis, pay for systems and people to develop answers to your deepest questions, and then ignore the results? Making the commitment to change an organization’s behavior is perhaps the most important step in any analysis project, because committing to change is like putting the roof on the house. Without a roof, the house provides no shelter. Without commitment, the analysis provides no value.
Management Sciences pioneer Peter Drucker wrote, “Enterprises are paid to create wealth, not control costs.” To create that wealth, executives use four kinds of internal information:
These four sources of information all look at the internal functions of the enterprise, and only provide visibility of the current conditions of the business. They illuminate the tactical conditions.
The future results are outside the organization. A review of your current sales data will tell you about the customers you currently have, and nothing about the customers who do not do business with you. No matter how successful an enterprise is, they never have more than a small fraction of the market as their customers. The market is the non-customers. Developing outward knowledge through analysis is a strategic effort, because it fulfills the commitment to growth.
To become powerfully successful, an organization must master the analysis of both inward and outward data.