
Today, digitization is so deeply rooted in people’s minds, we don’t need to reiterate the importance of data analytics for business decision-making or explain why data analytics can predict the future. The more data we have, the better we can predict the colour of the dice God throws. But the question is, how many enterprises can do it?
Last year, Yonyou conducted a research in mainland China, more than 400 enterprises participated in the survey. The results showed that almost all enterprises claim to know the importance of data for decision-making, but still more than 60% of enterprises in the research said that their decision-making is mainly based on experience and intuition.
Look At Data, Not Intuition
While only 8% of the results showed that they rely on intuition to make decisions, the reality may be more than that. In the context of employee performance, various analyses suggest that complimenting an employee improves employee performance more than criticism. This theory does not hold for a certain pilot instructor. The instructor shared his personal practice with Yonyou: when he criticized those pilots who had poor flying performance, the performance of the pilot would be better next time. And when he showed the good pilots who flew, there is a high probability that their performance was not as good as this time. So the instructor suggested that criticism is more effective than praise for improving student performance.
In this scenario, it is not that there is a problem with the instructor’s experience perception, but that the instructor ignores a very important model in mathematical models, mean regression. That is, when we enter the data of all pilots’ flight performance into the system and conduct data analysis, we will find that most of the pilots’ flight performance will fluctuate around a range. Therefore, those pilots who performed poorly this time have a high probability of performing better than this time, while those who performed well this time have a high probability of not performing so well next time. This is not the result of the instructor’s criticism or praise.
When we counted the number of times the pilots were criticized and praised, and cross-analysed it with the overall flight performance data, we found that the overall performance of the pilots who often received praise was higher than those who were often criticized.
Workforce Reporting ≠ Workforce Analysis
We should analyse the differences between workforce reporting and workforce analysis from the five dimensions of boundary, format, data, purpose, and value.
Traditional workforce reports require secondary processing before they can be presented to leaders, while YonBIP’s intelligent workforce analytics can help human resources departments directly analyse and refine data, truly liberating human resource practitioners from data collation.

The Workforce Analytics "Pyramid" Model
After long-term customer service and research in the field of data analysis, we have sorted out the workforce analysis “Pyramid” model and divided the entire workforce analysis into four stages, namely statistical analysis, structural analysis, correlation analysis and predictive analysis.
In the two stages of structural analysis and statistical analysis, the current situation of data is still focused on statistics and presentation, and multi-dimensional cross-analysis is carried out on some simple data, such as employee age and employee level, such as resignation and attendance data, etc. When it comes to correlation analysis, we only start to enter the advanced stage of workforce analysis. For example, the correlation analysis and regression analysis of the reasons for leaving can help us to conclude that among the younger generation of employees, “salary” and “stress” are not the most important factors affecting employee turnover, while “vision” and “career development” are the most important factors. Yes; even “stress” has a negative correlation index with employee turnover.

What Data Analysis Lacks Is Not The Goal, But The Path
It’s not that we don’t know what kind of data analysis we want, it’s that we don’t know how to get there.
Step 1: Build A House With Pipes
To build human resources digital insight, the core is to accumulate data of the whole process and establish a complete digital system. In the past, the human resource system lacked an integrated platform and application, so enterprises had no choice but to patch things up and turn their business online. Now YonBIP platform can completely solve the business needs of customers, laying a solid foundation for building a house through a pipeline.
Step 2: Install The Instrument And Measure The Performance
Through the display of past and current data, it helps enterprises to recognize the current situation and strategic positioning. At the same time, in the field of human resource analysis, effective data must be more important than comprehensive data. In the data displayed on the dashboard, it is not that the more indicators the better, but the less is more, and only the three most valuable indicators are presented to the leaders.
Step 3: Identify Cause And Effect Across Scenarios
At this stage, the human resources system is more focused on connecting with the business system. When the enterprise strategy is decoded layer by layer and finally settled on talent management and human resources planning, the human resources department needs to formulate clear implementation steps and strategies for these strategic goals. Ultimately help enterprises achieve business goals.
Step 4: Dig Deeper Into Intellectual Judgment
We need to carry out in-depth modelling and drilling of data, and change workforce data analysis from “accurate expression” to “find trends”.
In enterprise management, we have achieved accurate portraits of employees. By superimposing and analysing the portrait data of a large number of high-performing employees, we can draw the most suitable group portraits for the enterprise, as well as the entire organization portrait, so as to help us make decisions in the process of internal and external talent selection.

Step 5: Evaluate The Value And Determine The Action
On the premise that all technologies and data are already available, through the digital twin business platform, it can help businesses to perform virtual calculations on various scenarios, so as to find the best problem-solving ideas. Find the path that maximizes the value of the organization, and execute it with determination.
Just as in the movie “Avengers: Infinity War”, Doctor Strange predicts the future will fight Thanos 14 million times only won once, our ultimate goal, is to predict all variables in virtual reality through the digital twin platform, to help organizations find the best way to win.