First of all, everyone must understand that not all projects need to find a large hall with tens of thousands of people, holding banners, and the chairman and general manager taking turns to come on stage and sound the gong to clear the way. As long as it meets "specific output at a specific time and under specific conditions", it is a project.

Therefore, the key to doing a project is not to have a name, but to have specific outputs. With the output of specific products, KPI/OKR documents can be easily communicated; leaders will be more satisfied with you; you will have more capital when you are promoted and evaluated; and you will have more things to write on your resume when changing jobs. This is what we are striving for. The so-called "excellent" projects refer to outputs that are more convincing than "I ran a data".

So, where to start?

When doing a project, the most important thing is of course HE Tuber to clarify the goal; to clarify the goal, the first step is of course to figure out who to serve. This is the biggest difference between newbies and veterans of data analysis.

Often newbies who are new to the industry are full of “templates, models, and formulas.” I thought that as long as I made a copy of the template, the work would be completed. Rookies who are new to the industry like to speak in general terms: business. But business is not a lonely, independent person. Behind the word business, there is a very specific and complex meaning (as shown in the figure below).

Concrete analysis of specific problems is the most basic requirement for data analysis and the first step to do a good job in the project. Because these five major elements and their specific forms determine the extent to which our data analysis can be achieved, what it should be like, and how it should be made to meet the needs. The specific relationship is shown in the figure below:

It is important to clarify the specific issues. In the past, we often talked about traditional companies and Internet companies. Today, with the development of channel integration, the boundaries between the two are actually becoming increasingly blurred. If there is no specific analysis, a lot of jokes will often occur.

for example:

I used to be a toC Internet company, but now I want to focus on toB, and I have no idea how to deal with customers;

It is called an Internet product, and its service targets are entity owners, and sales are still using the most primitive outbound phone calls;

It's called the Internet industry, and it still operates with physical products, and the profit from sales and purchases is not bad at all;

It’s called New Retail, but the data collection is a mess, even compared to traditional chain stores;

It is called a traditional enterprise, but it is engaged in digital transformation and is engaged in distribution and fission;

The above complex scenarios cannot be solved by just shouting "I am the Internet AARRR thinking". The ending of the hope set template is death. Moreover, after several years of experience, many operations, product managers, and planners have learned basic data analysis concepts. At this time, they are still holding ppt templates full of empty slogans such as "SOWT, PEST, 5w2h", and data analysts are waiting to be laid off. Bar. Specific issues and specific analysis cannot be overemphasized.

Moreover, understanding the situation clearly is very important for seizing the opportunity in the next step. If you wait for business to come to your door to do everything, then it will be no different from a dog holding a Frisbee (the business proposes a hypothesis, and the data verifies a hypothesis, just like a dog holding a Frisbee). Only by judging the situation yourself can you proactively discover opportunities.

2. Find the opportunity to exert force

The biggest enemy of data analysis projects is: daily work. Therefore, not everything is suitable for setting up a project. Timing is very important.

Often we have to choose the following opportunities in the business department to start:

Want to innovate

Want to improve the present

New job makes eyes dark

Feeling at a loss when encountering problems

Three axes have no effect after cutting

In these fighter moments, throw out systematic solutions and solve the problem independently in one go (as shown in the figure below):

After finding the opportunity to make efforts, negotiate with the specific business parties and prepare to start work. Before starting work, you must confirm the project requirements, specifically the project iron triangle:

There are three points to note here:

1. Numbers, models, and reports themselves are not outputs.

The business goes from not knowing the situation to knowing it, from having no way to having a way, from not knowing how to choose to knowing how to choose, from not having preparation to level one, two, and three contingency plans. This is the output. So don’t just talk about numbers without being separated from the problem. Draw conclusions from numbers.

If time is tight, try to draw conclusions as quickly as possible; if time is long, output the paper step by step. Companies are not schools that leave you half a year to write down your thesis.

If the data quality is poor, the manpower is insufficient, and the analysis experience is lacking, just take it step by step and do not expect to solve all the problems at once.

These three points are crucial to project results. In the past, too many data analysts were addicted to tossing "scientific methods", neglected project management, and ignored time-investment. As a result, the pie was big and the pie was small. In the end, It ended in disgrace.

Here we also need to pay attention to the working method. Confirming needs does not mean asking the business directly: "What do you want to analyze?" This kind of question is too passive and goes back to the old way of holding a Frisbee. And often the business will give you an answer that leaves you confused.

for example:

Please help me think of a way (I want to ask you to implement the plan)

It must be the opponent/weather/luck... (attempting to pass the blame)

I want to analyze users’ mental resources (no data at all)

As long as there are artificial intelligence users, they will pay (the method is unrealistic)

Therefore, the reliable approach is to sort out the requirements step by step, guide them to problems that can be solved by data analysis, get to the root of the problem, and solve the problem objectively (as shown below). There are many details to talk about the specific guidance method. We will explain it in detail later based on specific cases.

After completing the requirements, the follow-up is formal work. The specific content is related to the analysis topic and will not be discussed here. If the early stage is sorted out clearly, the middle process will naturally go smoothly. I only emphasize one point here: remember to hold back on big tricks when doing data analysis. The longer you hold it in, the higher people's expectations of you will be, and the greater the disappointment in the end.

Therefore, as long as the project salary exceeds one week, there must be a weekly report to inform everyone of the progress; if the time exceeds one month, there must be a monthly summary and share the intermediate process with everyone.

Especially for projects that use algorithms, when they hear about the algorithm business department, they often think that they are magic soldiers descending from the earth, and they will be invincible wherever they go. Therefore, there are many examples of algorithm projects dying due to excessive business expectations. During the process, the specific algorithm process does not need to be reported to the business, but the difficulties encountered and the expected output conclusions must be communicated frequently, and business expectations should be appropriately controlled to avoid discovering that the goods are not correct at the last moment, and ultimately being ruined.

I won’t go into details here. Teacher Chen has updated a series of data analysis reports. You can follow the public account and refer to it in the menu bar. In short, when reporting, you must consider the identity and purpose of the target audience, and make a personalized report based on the project goals. Only in this way can good results be achieved (as shown below).

Based on the audience's thinking, even the same data and the same conclusion can be expressed in different forms, finally catching the audience's attention, making everyone interested, and completing the project perfectly.

Throughout the entire process, we can see that the process of doing a good job is the process of applying data methods to corporate practice. The data itself involves statistics, mathematics, programming, database and other professional knowledge, but a considerable part of it (such as data warehouse, ETL) is to ensure the normal operation of the data itself; a considerable part (such as semantic judgment, image recognition) is used for industrial applications , without considering business understanding and cooperation; quite a few (such as statistics) are suitable for scientific experiments, agriculture, forestry, animal husbandry and fish research.

A lot of business is not a scientific problem, but a practical problem. How O2O platforms manage merchants, how new media platforms develop local customers, how live e-commerce merchants choose products, and other issues require a combination of data knowledge and practical work to output conclusions.

Not to mention, everyone’s workplace is mixed with office politics, and everyone is asking how to get ahead and avoid taking the blame. That’s why we have today’s discussion and various ways to promote projects.