Detailed! 7 Types of Data Analysis Report Writing Guides

This article mainly shares 7 types of data analysis report writing methods, providing data analysts with a logical way of thinking about data analysis and avoiding meaningless presentation of data. It is recommended for people who have data analysis writing experience.

"Why is my boss not satisfied with my data analysis report?!" is a question that bothers many students. In particular, sometimes the boss will complain that "the report is too detailed and needs to be focused." Sometimes the boss will complain that "it is too rough and needs to be more detailed." What is the scale and how to strike the right balance?

The key point is: data analysis report is first of all a report, and it must have a clear report logic. As for the data, it just makes the report more credible and more real. So how many common logics are there in making reports? According to a person's familiarity with the content of the report, there are 7 types, which are introduced one by one as follows:

1. Introduction Report

An introductory report is suitable for the first report to someone who does not know the situation. An introductory report generally adopts a general-specific structure and is introduced from several angles. For example, it introduces the activities, some members, and product lines. At this time, the report framework is as shown in the figure below:

Attention! Many reports to big bosses will start with an introductory report, because big bosses may not understand the business details, so you need to explain it clearly from several angles. At this time, the indicators must not be too detailed, otherwise you will be criticized for "not understanding" or "too detailed".

2. Monitoring Report

Monitoring reports are suitable for: providing business development trends to people who have a certain degree of understanding. The length of monitoring reports depends on the needs of the leaders. Some leaders like to be detailed, so the monitoring content is more, while some leaders like to focus on the key points, so the monitoring content is less, and only key indicators are retained.

The key to monitoring reports is to explain the trend of indicators in conjunction with business actions. The most taboo in monitoring reports is to write them as a running account of numbers: this week, the year-on-year increase was x%, the month-on-month increase was y%, the time progress ratio was z%, and the KPI completion rate was xx%. People are confused and have no conclusions. When making monitoring reports, it is necessary to link the fluctuations of monitoring indicators with specific business actions to facilitate the readers to make judgments.

3. Exploratory Report

Exploratory reports are suitable for: giving people who have a certain degree of understanding the next direction of action. Note! What is given here is a "hint" and a "suggestion", not a conclusion. Therefore, exploratory reports must have strong logic and explain the logical relationship between the data listed and the conclusions given. The final tips/suggestions are also based on solid data.

There are many ways to use the logic of an exploratory report. The simplest one is: positive and negative example method.

• Positive example: After doing XX business, the performance is better.
• Reverse example: Businesses that do not engage in XX behavior perform worse
• Suggestion: Promote XX behavior to improve overall performance

Of course, the more examples you can give, the more complete the argument will be (as shown below)

4. Diagnostic Report

Diagnostic reports are suitable for: explaining the cause of the problem to people who already understand the problem. The simplest diagnostic method is structural analysis + indicator decomposition/funnel analysis, which points out that there is a problem in the XX link of the XX business. However, for a more in-depth diagnosis, more complex logic may be required, and even AB test/labeling group comparison may be combined to produce results. Therefore, diagnostic reports can be written in depth. If the current analysis is not in-depth, do not make a closing report, but rather a process report, listen to everyone's opinions, and then decide how to go deeper next.

Note! Diagnostic reports are only useful for people who know the problem well! If a person does not understand the background of the problem, he should first make an introductory/monitoring report. Otherwise, after giving the report, people will often criticize him: "I don't understand!" "Why do we have to worry about this problem!!!"

5. Test report

Test reports are suitable for explaining test results to people who already understand the problem. Test reports are also only useful for people who understand the problem! If a person does not understand the background of the problem, he should first make an introduction/monitoring report. Otherwise, he will often be criticized after giving the report: "What's the point of going through so much trouble!"

The structure of the test report itself is simple and clear:

• Problems to be solved
• Test plan ideas
• Test Results
• Post-test recommendations

The main reminder here is to not forget to explain what the problem is! Even if some tests are done because: "I have a good idea", "I saw my competitors do this", "I heard my boss say this", it is best to find a clear and quantifiable problem point/test goal. Otherwise, after the test is over, you will have to look at dozens of indicators to see the differences, and then you really can't write a report...

6. Predictive Report

Test reports are suitable for giving predictions about a problem to people who already know the problem. Pay special attention! When writing a prediction report, you must pay attention to the level and appetite of the audience in advance.

• Do the audience want to participate in the prediction process?
• What concerns do the audience have about the future?
• Do the audience have their own predictions?

This must be sorted out in advance.

Otherwise, when you are talking about forecast reports, you will often be stopped and scolded: "Your forecast does not meet business expectations!" "Your forecast does not consider important issues that the business cares about!" "Let me ask you, if XX happens, will your forecast still be accurate?!"

VII. Evaluation Report

Evaluation reports are suitable for giving a comprehensive evaluation of the problem to people who already know the problem well. Note! Like predictive reports, the success or failure of evaluation reports is also determined by the level and appetite of the audience.

• Does the audience have their own evaluation insights?
• Does the audience's opinion matter?
• Does the audience have a preconception of the evaluation results?

This must be sorted out in advance.

Especially when the evaluation involves departmental interests, such as the results of major projects, such as product performance. It is best to get a feel for the audience's attitude before writing the report. If it involves departmental interests, tell everyone the evaluation method in advance, and discuss which indicators and reference standards are used in advance. In this way, everyone will accept the result. The most worrying thing is that there is no unified method in advance, and everyone will raise or lower the results afterwards, or even temporarily modify the goals. It is better not to write this evaluation report. Writing it will only make the data analyst lose his integrity.

When making data reports, think clearly about:

• Who do I report to? (Supervisor/colleague)
• What am I going to talk about? (Choose one of the seven categories)
• Does he understand what I am saying? (If he doesn't understand, I will introduce it to him first)
• Will his opinion influence the results (especially the forecast/assessment report)?

In this way, the data report can hit the other party's pain points, answer their questions and reduce doubts.

Some reports can be very complex and long. The longer the report, the more you need to think clearly about what the central idea is. Avoid laying out data without any meaning.

Author: Down-to-earth Teacher Chen

Source: WeChat public account "Down-to-earth Teacher Chen (ID: gh_abf29df6ada8)"