1, and master the idea of building an analysis framework.
2. Learn the writing skills of excellent cases.
Make sure that the readers of the report can get some valuable opinions and suggestions for their work through reading for 3-5 minutes.
Explain your point of view in the most concise and unambiguous way, and don't let the boss bother to understand and guess what you want to express.
No matter how many working papers you have prepared, only write the most relevant and valuable parts in your report.
Each PPT page or paragraph only expresses one point of view, and too much content will only drown out useful information.
Writing a business analysis report should have "empathy" and think from the audience's point of view!
Don't talk nonsense: the more detailed the analysis, the better. The simpler the conclusion, the better. Don't show off your skills.
Don't forget the beginning and the end: where the indicators start and end, don't go astray.
Don't cover everything: time is the biggest cost, so pay attention to the big and let go of the small.
Don't forget the conclusion: there must be a conclusion, even if it is normal or not done.
Don't believe all the data: the data is good, but sometimes it means that you should dare to make decisions.
The framework of the analysis report is particularly important. Before writing an analysis report, you must first write an analysis framework. What should be written in each report? How many large pieces are there? From which dimensions should each chunk be analyzed? What angles should be designed? What should I evaluate specifically? Everything should be made clear in your framework. Moreover, in the analysis report, it is best to have a first-class title and a second-class title.
The analysis report is actually telling a story and telling the background of this matter; However, a weekly newspaper, for example, can be reported every week for a long time, but the data range of analysis should be introduced first, such as what range of data you analyze in the report, what kind of users you choose, what is the caliber of each indicator, and so on.
In terms of analysis methods, if you don't use special analysis methods, if you only use a simple analysis without using complex analysis in your daily work, you don't need to introduce your whole method; But if you use it, be sure to tell us your analysis method and what you found through the data.
After that, be sure to show your data, then select the key parts of the whole data for analysis, and then calculate the income, but this is not what every report will discuss.
After analysis, have you come to a conclusion? So how did you come to this conclusion? The derivation of this conclusion is also concerned by the boss (for example, you said that there is a problem with the transaction transformation now, so how did you find the problem of the transaction transformation? What is the reason? All this must be made clear, otherwise no one will believe it. If you just say that I analyzed it anyway, it is the problem of that page, then what is wrong with this page, and how do you analyze it? How is it derived? You must be able to tell everything.
Remember that writing an analysis report is always like writing a proof. Your subtitle is the problem you want to prove. Your data and some detailed explanations are all explaining the content of your secondary title. After the final analysis and proof, you must have strategic suggestions. This is particularly critical.
Various methods are usually used when writing analysis reports, but these analysis methods are not very difficult, because our daily work is business-oriented.
Business-oriented means not emphasizing fixed analysis methods, but emphasizing analysis methods suitable for business; We will adopt any business that suits us. It is better to use classification in this scene, so we use classification; If it looks good, we will analyze it; If the relevance is appropriate, we will analyze the relevance and so on.
Don't deliberately write it in the analysis report in order to choose a difficult method. Daily work is not writing papers. The boss won't ask you to complicate a simple matter. What the boss always wants is to simplify complicated things and make others understand the best method, so he must pay more attention when choosing the method.
Everything will emphasize methodology, and behind these analytical methods is the concept of methodology. If you follow certain analytical methods in some things, your logic and system will be better.
Classification: equivalent to dimension drilling. For example, to analyze the level of GMV, it is necessary to drill down into the classification, such as city classification, user classification, category classification and so on. And RFM model, which is actually a classification analysis method.
Distribution: age distribution, city distribution, purchase time, user distribution, education distribution, grade distribution, etc. It is not enough to look at distribution alone. We must also think about whether this distribution trend has changed. If so, the content of this distribution trend should also be one of the contents of the analysis.
Comparison: Data are meaningful only through comparison. It is difficult to know whether a numerical value is good or bad. Other numerical values are needed for comparison to make these numbers more conceptual and meaningful, such as comparison, ring comparison and competition.
Correlation: analyze whether a certain factor causes the change of an index and whether a certain change is caused by an index. At this time, we need to do some correlation analysis, and also look at the correlation between the two indicators (whether there are famous schools and grades, income and price sensitivity, age and price sensitivity, etc.). For example, wake up a sleeping user, and then lead to a slight increase in the repurchase frequency of old customers. At this time, you should look at the fluctuation of users' purchase frequency, and we should show it to sleeping users.
These analysis methods are meaningful only if they are business-oriented.
Classification, distribution, comparison and correlation are big analysis methods, which will also be used in every analysis report. There may be some subclasses or some overlapping analysis methods under these categories.
Dimension subdivision: for example, by category and by city, all are dimension subdivision;
ABtest: control univariate analysis, blank group and experimental group;
Queue analysis: retention analysis, such as the next day, the third day, the seventh day, the thirtieth day, 180 day, etc.
Structural analysis method, DuPont analysis method, weight decomposition method and analytic hierarchy process: these mainly analyze drilling according to the idea of decomposition and subdivision;
Sales volume method:
Permeability method:
Quantile analysis: it is also a kind of classification, which can be compared with box chart;
Matrix analysis: a particularly commonly used analysis method, somewhat similar to RFM model. The RFM model is divided into eight categories with three dimensions, and four categories if only two dimensions are taken.
For example, take the down payment of real estate and the annual growth rate of real estate as two indicators, and then take the city as the dimension (that is, there must be three data), so that each city has the values of these two indicators, so that the positions of these cities can be drawn in this quadrant, and then four quadrants of these cities can be divided by the intersection of the average values of these two indicators. At this time, it is equivalent to drawing a scatter diagram, so that the down payment and annual growth rate of the property can be drawn.
Matrix analysis is also an analysis method that is often used in daily work. Especially, according to the performance of cities on certain promotion dates, such as turnover and growth rate, the capital cities and municipalities directly under the Central Government of China are analyzed in four quadrants, or the GPD and growth rate of each city are analyzed.
This is a reference for the overall typesetting form of the report. The first-level title is the table of contents, and the second-level title is your core conclusion. The data chart is placed in the middle and the strategy is placed in front of the chart. Data interpretation can be placed on the right side, or under the secondary heading or chart, with some special remarks (such as some special references or indicators, etc.). ) the lowest.