Introduction:

In this era of big data, data is not just idle information that a company has forgotten in a corner at a certain stage, but a valuable resource that is related to the “survival” of the company. “How to do a good job in statistics” is a question we must consider. Here, we will make some simple discussions around this topic.

Why do we need to do statistical work

Sun Tzu said: “He who knows both the enemy and himself will not be in danger in a hundred battles; he who knows himself but not the enemy will win only one battle and lose only one; he who knows neither the enemy nor himself will be in danger in every battle.”

Market competition is becoming increasingly fierce, and “knowing yourself and knowing your enemy” is often related to the life and death of an enterprise. So how can we “know ourselves” and “know our enemy”

Statistics are an important means. Internally, statistics can reflect the operation status of an enterprise within a specific period of time, allowing the enterprise to “know itself”; externally, statistics can reflect the situation of competitors and the market, allowing the enterprise to “know the enemy”.

Statistical data is the data support for corporate decision-making. Only by doing a good job in statistics can the company act in an orderly and appropriate manner and give full play to its own advantages.

How to do statistical work well

Since statistical work is so important, how do we do it well

Here, we can use American Shewhart’s “PDCA cycle” working method to reasonably decompose statistical work.

“PDCA Cycle”

Everyone should be familiar with the “PDCA cycle” working method, which guides us to make plans, implement plans, check the implementation results of each task, and then include the successful ones into the standards, and leave the unsuccessful ones to be solved in the next cycle, so as to achieve a step-by-step improvement.

According to the “PDCA cycle”, we can divide statistical work into:

1. Plan: Design data statistics tables and open up data collection channels;

2. Execution: data collection and aggregation;

3. Inspection: data collation;

4. Processing: Data analysis and corresponding processing based on the analysis results.

Next, we will discuss the statistical work of enterprises in detail based on this idea.

1. Plan

Statistical work requires a clear statistical goal. Statistics without a goal are meaningless.

The work at this stage is to compile a set of scientific, reasonable and efficient statistical tables through analysis of statistical content and objectives, and to reasonably set up information collection nodes and open up information collection channels.

The steps for analyzing statistical goals are roughly as follows:

01. Determine statistical indicators

02. Break down the metrics into parameters that can be collected directly or indirectly

03. Determine the collection parameters based on actual business situations

04. Design statistical tables, classify and arrange basic information, fill-in information, calculated information, etc., and arrange information reasonably

05. Refine the statistical table, fill in the requirements, set up error-proofing alarms, etc. to improve the accuracy of the report

Statistical tables should be clear, concise, reasonable, and easy to understand, with fixed, sourced information and individual indicators, and reasonable citations without the need for repeated statistics.

Designing a good statistical table does not mean that the design work is complete. It is also necessary to establish a reasonable and efficient collection channel. Therefore, after the statistical table is designed, it is necessary to design a reasonable collection channel and conduct systematic training for the person in charge of each node.

2. Execution

The issuance of statistical tables does not mean the end of the work. On the contrary, it is just the beginning of statistical work. The execution of this stage determines the quality of statistics and even the success or failure.

How to ensure the high quality of collected data We need to do several things:

01. Assign responsibilities to individuals

02. Proper training

03. Follow up in place and correct deviations in time

04. Incorporate performance

3. Inspection

Statistical reports are summarized and organized into electronic documents, summarizing multiple data into one statistical table.

Check the completeness, rationality and accuracy of the data. At this stage, careful attitude is more important than ability, and requires carefulness and patience.

4. Processing

The so-called statistical collation is to scientifically process and collate the collected raw data according to the requirements of the design and research tasks, so as to make it organized and systematized, and transform the large amount of raw data reflecting the overall units into basic indicators reflecting the overall units. The work here includes: classification, reorganization, summary, calculation, processing, etc.

It can be roughly divided into several steps:

01. Design and finishing plan

02. Review and revise the survey data

03. Conduct scientific statistical grouping

04. Statistical summary

05. Compiled and organized data tables and data presentation

Analyze the sorted data, use scientific methods and effective tools to conduct analysis, and formulate improvement measures for the problems behind the data.

Most statistical work is a long-term systematic process of discovering problems, solving problems, and optimizing improvements through statistics. We must make good use of it!