Be careful not to drown in a sea of data
The data value is not in the volume itself, but the insights that are generated from a clear vision of what needs to be analyzed
I see that many still refer to Big Data as basically a huge and varied amount of data in the company plunges to take fantastic insights. Herself datificação the company generates more and more data. In fact, the basic idea behind the name Big Data is the realization that virtually everything we do in our day to day will generate a digital footprint that could possibly be analyzed. And we read in the media, that data volumes in the range of petabytes is behind the initiatives of companies like Walmart, Amazon, Facebook, LinkedIn and others that are the benchmarks of Big Data. But the reality is yet another: unfortunately few companies have the capacity and expertise of these companies to develop similar actions.
But therein lies the distortion. Big Data does not necessarily imply store and analyze huge volumes. Also, start any of Big Data initiative to dive into an ocean of data, without well-defined questions, you can generate a lot of correlations that simply will not make any sense. For example, what relationship will exist when you discover that an increase in margarine consumption corresponds to an increase in the divorce rate? Nothing makes sense! A simple spurious relationship. Insights and more insights of this kind will make you drown in an ocean of data …
The positive aspect of Big Data is that even if the company does not have data on the scale of petabytes it can generate much value analyzing your data. Since it considers that the data value is not in the volume itself, but the insights that are generated from a clear vision of what needs to be analyzed.
Yes, the first step is to have a clear vision, a clear strategy of what to do with the data. Big Data will only be a competitive advantage tool if you have a well-defined strategy. If you have questions, you can look for answers analyzing their data. If you do not have the, what will you do with the correlations that appear?
A simple example. If you do not know what I want, that will analyze data? All throughout the enterprise? Those that are stored for more than five years or only the most recent? It will be worth accessing external data, coming from social media? Transform the audio Service Center Customers in text and analyze them? What benefits will I get after all this costly work? In summary, for the data to be really useful, it is essential to know what data is needed.
Let’s clarify the importance of starting with a vision, with a well-defined strategy. Suppose a retailer wants to increase its customer base. The questions that make sense are likely to be something like “Who are our current customers? What is your demographic and socioeconomic profile? Why are our customers? How many people in this profile are our customers? What would encourage others to be our customers? “And so on. This makes delimited the scope of data that will be needed.
With a well-defined strategy and the questions that will be useful if we analyze the data exist and are available.
Let’s imagine a situation where a chain of stores want to know if it is worth or not to keep certain filais in shopping malls. There may be need to generate information not collected yet, installing sensors in the windows of these shops to measure how many people are in front of them, how many stop to see the windows, how many come and how many actually buy a product. Crossing the data generated by the sensors with the sales system data can be concluded that certain shopping centers do not generate enough customer traffic to justify the maintenance of a store.
With well-defined strategy you can begin to identify areas for which the company must start their Big Data initiatives. What areas will bring higher return with less investment of resources? For example, in the relationship with customers, the company is really acting in the socioeconomic segment that intended? Current customers are satisfied with the company, its products, its after-sales? What is the dissatisfaction rate, measured in some sectors such as mobile telephony, churn rate, when the customer leaves the carrier?
Probably with time and maturation process, Big Data & Analytics will be ordinary activities. Working with facts and insightsdata-driven tend to be commonplace. Interestingly, as the company directs its decision-making processes to be more grounded in data, many myths and beliefs, crystallized by intuition can be demolished.
A very interesting case is Google, detailed in this article from Harvard Business Review, “How Google Sold Its Engineers on Management”. The article shows that Google’s HR culture, a software engineering company, did not value the management function. In the company’s culture, supervision activities were a “distraction” of the activities that really mattered, how to program and debug code. In 2002 came to eliminate managing, but it was an experience that was short lived. Today Google has management levels, but on a much smaller scale than the majority of companies with same number of employees. And, through the use of data is able to measure the performance of their managers and identify where and how to improve this performance.
It’s worth reading the article, it shows how big data can be applied to innovatively HR function. They went out of the judgment formed from the beginning of the company that managers would not be necessary and not impacting employee performance for clear and measurable vision, which proves statistically so, good managers cause great impact on the performance of a team. Were also able to mathematically identify the main features of what being a good manager. But again I emphasize, came to this answer because they knew what questions to ask!
The summary of the story is that when you start with a clear question, you will identify what data is needed to answer it. The answer may come from structured data within the company. Or also, internally unstructured data, generated by emails or audio captured in call centers. In practice, usually does not begin with unstructured external data generated by social media …
Acting consciously, not by fad, we will not go diving into an ocean of data, trying to mine anything, often finding spurious relationships that lead nowhere. The important thing is to get out of the misconception of “all data” to just “relevant data”.
Full article: http://cio.com.br/
About Administrador
