Four Layers of metric analysis report information: data, information, analysis, and measures

Author: Chen Yong

Original article: http://blog.csdn.net/cheny_com

 

Is this often the case: When a leader receives a report, it is filled with various exquisite reports (assuming we no longer talk about reports composed of texts). However, the entire report is still on the cloud, leaders are overwhelmed after reading the report, and the report is final.

This is because the report producer ignores the report's end purpose: the leader or other readers want to take measures after seeing the report to prevent the occurrence of bad things or accelerate the occurrence of good things.

See the following example:

 

Defect Trend

Data:

Defect trend chart

Information:

The number of defects found in V3.0 is much higher than that in V1.0 and V2.0. We need to take action!

This requires that at least 1.0 ~ 3.0.

Analysis:

There may be many reasons for the number of defects, such:

V3.0 code is much larger than 1.0/2.0 (this requires that the number of lines of code be drawn simultaneously in the trend chart)

Because of the tight schedule, V3.0 does not perform automated testing (this requires that automatic test coverage be drawn synchronously in the trend chart)

The V3.0 team has joined many new people (this requires that the number of people or mobility be drawn simultaneously in the trend chart)

......

Measures:

For each reason, you can find one or more measures to improve the performance. For example:

...... Much bigger = change to defect density. Maybe there are no more defects.

...... No automated testing = guaranteed coverage

...... Many newcomers = training, norms, mentor system, primary programmer system, code review, etc.

......

Specifically, the author's habit is to produce multiple reports side by side. The first report shows the information, and the second report helps analyze the cause. If no exception information is displayed in the first report, you can ignore the second table. Otherwise, find the cause from the second report. Because there may be many reasons, we should first use the nose to "smell" the potential cause, and try to use only one valid reason table instead of listing all the cause analysis tables.

Another question: most of the measures are intuitive. Why can't I know what to do after analysis? Can't I do it all? This involves the "Don't waste principle" in evolution ". Why does the human eardrum muscle degrade (the correct question is: why are people who can understand the ears eliminated )? This is because all creatures that waste their behaviors in evolution are eventually eliminated. According to the famous saying, "If we can improve by 1% every day, we will be able to improve by twice in three months" (1.01 to the power of 70 equals 2 ), the problem with us is that "those who waste more than 1% things every day will spend three months ...... Almost collapsed"

 

The following two additional questions are provided for non-software professionals. I personally participate in and get the answer. Of course, the answer for each enterprise is different.

L Baidu keyword promotion performance report

L long-term and short-term sales analysis reports

 

Please indicate the source of the reference.

 

Click to download the free agile development textbook: Martian agile development manual