This paper is organized by the online course "Human Resources Big Data Application Practice" taught by Wang Aimin.
So when I got to this place, I made fun of you. In the early years, people often said that human resources people didn't work, not that we didn't work. Because we are not familiar with strategy and business, we are limited to do what everyone thinks. Why? Because human resources were not particularly valued in the early years, they were in a state of being unable to keep up. Speaking of this, everyone may feel more consistent, that is, the boss did not pay attention to HR, such as some important business meetings, operational meetings, and executive meetings. Our human resources leaders have no opportunity and power to sit there and contact, attend meetings, listen and understand together. How do we communicate it to other human resources directors? So every time I consult, I am particularly willing to discuss this issue with the CEO of the company. First of all, we should do this project well. Your human resources really play a role, you must start with the boss. As the boss and CEO of the company, what is your understanding of human resources? If your understanding of human resources is not high enough, it is really hard to say whether the project can really help your organization after implementation. So when I worked as a human resources consultant for the chairman of a listed company in recent years, in the first few months, my focus was to share my thoughts with the chairman and the general manager of his branch, so that they could realize the importance of human resources.
After doing this, human resources professionals can truly integrate into the senior management, truly understand the business, and truly play their due role for our company and our organization. So I think the premise of business understanding is strategy, and the basis of business understanding is strategy, which is what I have been trying to emphasize and analyze for everyone. I think there are some technical things behind this, and we just need to get a general understanding. We are on this basis. A model evaluation application with clear data collection and data preparation can be said to be from data understanding to data modeling. I think it can be called a black box, just like an airplane black box. Here we can do it by technicians, which we understand as HR. On this basis, after guiding these technicians in front, the final application must of course be applied by our HR executives, including our CEO. I'll tell you more about this later. This is the business logic we want to pursue. This is particularly important. I have always stressed that I especially want to share it with you.
On the basis of the previous business logic, we also have the logical diagram of the indicator system. What does this logic mean? We use logic to drive, which is actually such a logic drive in evidence-based thinking. Logic-driven analysis method can make human resource managers more skillfully understand the relationship between human resource analysis and actual operation, and make our HR use more in line with the ideological logic of enterprise operation management ideological framework. In other words, what is our company's strategy? What is our business? On the basis of our HR understanding, in fact, you are at the same level as the boss of the company, so that you can understand what the boss is saying and how the enterprise should develop, so that we can better serve the enterprise and greatly improve the reliability of our analysis results. Let's take a look at the logical diagram of this indicator system. We only consider people's problems on the basis of some strategic business. What's the problem with people? Here we have to consider the business strategy of our company and the life cycle of talents, so I won't go into details about the image, service, innovation, quality and efficiency that we are all familiar with. Then this is based on the talent information collection, as well as some of our key indicators, such as finance, operation, ability, customers, etc., to select, evaluate, develop, reward and punish talents, which is an idea of our BIC. Look at this table specifically. Consider the selection, evaluation, talent development, talent incentive and talent retention from the perspective and logic of talent development cycle. Where do we analyze latitude? Optimization, predictive analysis and correlation, we are actually bottom-up, we describe and benchmark, correlation analysis, then prediction, and finally optimization. What should we do from the perspective of talent selection? From the basic data of correlation and cross analysis, how do we consider the optimization of prediction? From the perspective of talent evaluation, we also follow the logic that we describe basic data, we look for correlation analysis, we make predictions, how to optimize our management and so on. , including personnel training. How do we follow the logic of data analysis? This is such logical thinking.
As mentioned earlier, the business logic of our index system construction. One is business logic, the most important thing is from the understanding of strategy and business, and the second is from the perspective of the growth law of talents. Then on the basis of this business logic, we can establish our index system. In fact, the architecture of this system is divided into four layers: system layer, data layer, analysis layer and display layer. The system layer is that we have some data around us and an ERP system. We can write it in, in and out, choose words, some data in the intranet, some data in the business system, that is, the data around us should be selected at the system level, and how to do it at the data level. The data layer is a purely technical operation, so it doesn't matter if our traditional HR doesn't understand technology. I'm just saying it's okay to give it to our technicians as a black box. Behind the data layer is the analysis layer, which is based on the analysis of the previous information, knowledge and data. Then there are business scenario analysis, operational management indicators, organizational efficiency indicators, cultural vitality indicators, operational analysis indicators and so on. According to these models, we analyze and model according to these data, and the output is early warning forecast and user portrait, which provides decision support suggestions for our executives. This is a platform for human resources big data. This is basically an upgraded version of Baidu's big data platform. From this place, we can intuitively see that if any enterprise is willing to build a big data platform, it will basically follow a data platform such as system, data, analysis and presentation.
Then let's take a look at these platforms separately. From the most basic point of view, the foundation is the construction of lifting adjustment of entrances and exits and the selection of reserved pipelines, that is, from the perspective of the three pillars of HR, it is SSC or SDC.
In fact, I have always advocated building the three pillars of our HR, SDC or SSC, into a platform for big data analysis, which must have the greatest value to our HR, our business and our strategy. The key of our system is that we need HR to play a basic part of your ability. How can we pour some data into this platform? The data we are looking for may be different from the human resources information system, as well as the resignation management system, salary and benefits, which may be in our human resources information system itself, and more, such as performance development and probation management. This is where our grass-roots HR should play a role. For example, some of our commissioners and supervisors can build integrated data according to the logic of these data and some requirements of our HRD or HRVP or even COE. The second layer is the data layer, which is actually pure technology. We call it data analysis, including data cleaning, processing, extraction, preservation, labeling and so on. The data processed by this layer includes both structured data and a large number of unstructured data. For example, we participate in the main process construction of data, which are more professional contents. Data preparation and working data analysis involve data mining, modeling, verification and avoiding data traps, so we play the role of prediction, which is a data analysis logic. This is the data layer.
Let's look at the analysis layer of data. What does it contain? Theme analysis, formulation and analysis, index system and model construction. What is the most important thing in this? For example, what happened in the past, diagnose the problems, find out the reasons, learn from some past data through analysis and modeling, and find out the correlation, instead of verifying it through assumptions as in the past, let the data play the role of prediction and early warning through timely observation, timely analysis and timely adjustment of today's data, which is to provide convincing basis for possible future decisions. This layer is the process from data information to knowledge. Why did I tell you those concepts first, just to use them in this place? From data information to knowledge, data analysis needs the combination of mathematical theory, industry experience and calculation tools.