Discussion on "Energy" and "inability" of large data

Now, we are in the age of data explosion, global data volume is increasing at an alarming rate every 18 months, the world is high-speed digital. Big data is also a topic that is being talked about in all walks of life, and some data analysts have even threatened that if you can capture all the data in real time and accurately, and have enough efficient algorithms and storage equipment, large data can analyze and solve all problems. To think that such a statement is too absolute, the reality is not so. Big data is not omnipotent.

In this paper, I try to analyze the http://www.aliyun.com/zixun/aggregation/13989.html "> Telecom Industry Large data can do and can not do, but for the business level of" energy "and" can not. " will be left to the subsequent articles.

Big data at the customer level "can Do" things:

1, improve customer portrait, insight into customer characteristics: With more comprehensive customer data, can be more close to the customer's real situation. Large data because of its powerful digital memory function, to a certain extent can do more than customers themselves have to know more customers, with "mind-reading" function, this easy to understand;

2, the discovery of behavioral patterns of DNA, predicting the action of the customer: French mathematician Poisson said: Once we admit that human behavior is random, it suddenly can be predicted. Albert Laslo Barabasi, author of the outbreak, concludes that human behavior 93% can be predicted according to Poisson distribution. The core function of large data is correlation prediction, for example, to identify the behavior pattern DNA of the net customers before they leave the net, and to infer all the net customers ' net-rate at a certain time. Similar also has the customer to change the machine time, the preference model forecast and so on.

3, identify customer demand preferences, to carry out personalized services: or around the customer, large data can find customers interest preferences, channel preferences, in the rule engine real-time triggering action, the corresponding contact can instantly capture the opportunity to trigger the completion of the corresponding action, personalized precision services and marketing, to "timely occasion", "Right to the customer", this is to improve marketing efficiency, customer perception is certainly helpful, of course, it should also pay attention to let customers more comfortable to receive contact service, do not let customers feel that we are in the use of their privacy in doing things, which is exquisite skills.

Big data at the customer level "Can't Do" thing:

Large data can indeed record the customer's various attribute characteristics, behavioral trajectory, this data does reflect the customer's operation and use behavior, but the thought is not exactly what it is, the customer's behavior can not fully reflect its true intentions.

1, large data can not "calculate" the customer's creativity and imagination: Big Data comes from reality, but many human ideas do not come from reality, creative thinking and imagination is often unrestrained, beyond the reality, so the "big Data Age" author Kotto Maire Schoenberg bluntly: creativity and imagination, with large data is "counted" not out.

2, large data in a timely and intelligent can not replace the customer's thinking: Large data may help customers make some decisions, but ultimately choose the customer which plan, what action to make, the final decision or in the hands of customers. The human thought process, the inner real idea is the big data cannot be completely calculated. Human thinking, decision-making embedded in the time series and social background, but the data can not read these backgrounds, but also can not read the background of some of the unspoken rules, and therefore can not understand the emergence of human thinking process. Even a common novel, data analysis can not explain the thread of thought, the obvious big data can not replace human thinking.

3, large data can not predict beyond the human cognitive range of things: the core function of large data is the prediction, but large data can not predict without warning, beyond the human cognitive limit of things, such things are often called "Black Swan". Large data is based on historical data to predict the future, but when the history is not mastered, the big data is also helpless, moreover, the big data in the collection, the processing process inevitably is merged into the data analyst's value and the tendency, this will let the data often is not the original objective, will affect the final analysis result, but the real "Black Swan" Hidden in the invisible, it is difficult to be found; In addition, Nasim Taleb, author of the famous Idea, Black Swan: How to deal with the unknowable future, pointed out that as we have more and more data, there are more and more statistically significant correlations that can be found, and many of these relationships are meaningless. It is possible to lead people astray in the real solution of problems.

4, large data can not describe the feelings of customers: large data Another limitation is that it is difficult to show and describe the customer's feelings. Large data in dealing with human emotions, social relations, and related issues, the performance is often unsatisfactory. The big data can only tell us what the customer is doing, not what the customer is thinking about when they do it, what the background is, or what mood swings the customer is doing. Therefore, large data is often not directly to the customer's mental space, understand the value of what customers have.