Detailed and comprehensive: Data analyst growth strategy

Data analysts often face multiple challenges in their careers: heavy data collection work, lack of leadership guidance, unfamiliarity with the business field, and frequent and high-quality reporting requirements. This article will explore these challenges in depth and provide a set of practical solutions to help analysts clarify the problem-solving steps, sort out the data indicator system, establish problem diagnosis logic, and measure the effectiveness of business behavior.

Being busy collecting data every day, lacking leadership guidance, and being unfamiliar with the business are the three major obstacles that hinder the growth of data analysts. If you add: being chased for reports + reports being criticized for not having depth, it's like a bolt from the blue. The most typical example is this star friend, who has all kinds of difficulties gathered together:

Because the problem is so typical, today I will specifically share some ideas to break the deadlock, and students who have similar difficulties can use it as a reference.

1. Clarify the steps of problem solving

There are fixed steps to solve business problems with data analysis:

Step 1: Sort out the data indicator system based on business processes

Step 2: Diagnose problems in business development based on the indicator system

Step 3: Analyze the effectiveness of business actions based on problem diagnosis

Therefore, when you arrive at a new environment, the first step is to understand the business process and sort out the indicator system. Like this student, the indicators he wrote down were scattered and it was difficult to tell the priority. This is the first step to solve the problem.

2. Sorting out the data indicator system

The reason why it is called an indicator "system" is that it is difficult to see the problem with a bunch of scattered indicators, and they need to be organized logically. There are three common organizations: parallel, total score, and funnel. This student mainly does toB sales, so it is suitable to use the total score method. First, clarify the relationship between GMV, average order value, frequency, gross profit, and net profit (as shown in the figure below).

Afterwards, you can combine the new customer visiting process/old customer ordering process to clarify the sales process indicators (as shown below).

The data indicator system also needs to be coordinated with classification dimensions. The sales indicator system generally uses people (customers), goods (commodities), and places (sales channels) as classification dimensions.

This will build the skeleton of the indicator system, which can be modified later according to the details of the business process, and diagnostic analysis can also be performed. This step is basic work, and you must first be familiar with the business scenario and clarify the relationship between indicators, otherwise it will be difficult to do the rest.

3. Establish problem diagnosis logic

With a data indicator system, it is easy to diagnose problems. Generally, the logic of diagnosing problems is constructed according to the principle of "from far to near, from big to small":

1. Whether it complies with the regular sales cycle (excluding false alarms)

2. Whether the decline continues (start from the source of the decline)

3. Is it the number of customers that has decreased, or the order amount that has decreased? (Drill down)

4. Are there fewer big customers or fewer individual customers? (Human dimension)

5. Is it because there is less seasonal fresh produce or less hard currency (the dimension of goods)?

6. Is it because there is less sales development or fewer independent orders (field dimension)?

In this step, there is no need to split questions 1, 2, and 3. You only need to look at the overall indicators, so you can get started quickly. Questions 4, 5, and 6 are more troublesome. First, you need to split the data and observe it many times. Second, the problems may be different at different time points, so you need to change the split dimensions in different months. Third, you need to accumulate business tags to facilitate the interpretation of data.

This process requires the accumulation of analytical experience and continuous attempts , so it may take a long time. With the increase of accumulation, a data analyst also appears to have more and more industry experience.

4. Measuring the effectiveness of business actions

"Selecting categories that merchants frequently purchase and raising prices" is an action taken by a business to improve performance. This kind of analysis of specific business actions requires first proposing analytical assumptions, which makes it easy to draw conclusions.

For example, raising prices can improve performance. In the best case scenario, the products selected are high-premium, high-quality non-standard products, such as small blue dragon fruits and cherries. It is difficult for customers to compare prices horizontally. In addition, the quality of the products is good, so they are not so sensitive to prices, so they can accept the price increase.

In the worst case scenario, the products selected are ordinary standard products with little quality differentiation, such as rice, flour, oil and eggs. Customers will run away as soon as they see “Oh, you are so much more expensive than the market price!”

When doing analysis, the above assumptions can be expressed as follows:

After the data comes out, just observe the results directly.

1. After the campaign was launched, did the overall performance increase (success) or remain the same or decrease (failure)?

2. Which categories of products stimulated by the activity have high price sensitivity (failure) and which have low price sensitivity (success)

3. Did the customers who participated in the activity reduce their overall consumption (failure) or switch to purchasing products without price increase (failure)?

It is relatively simple to evaluate the effect of an activity. As a data analyst, as long as you can understand the activity goals and the business logic of how the activity achieves the goals, you can quickly output the evaluation results. This is a skill that data analysts should prioritize.

However, answering the business department: "What can I do to improve performance" requires a long time to settle down. Because there is more than one way to improve performance, you need to be knowledgeable in practice and understand more business methods before making suggestions. This requires data analysts to take more notes and sort out different business actions around the same goal, so that they can compare the results later.

Of course, if there is a leader to lead the way, the above is not a problem. But it happens that many students lack leadership guidance at work, and even don’t have anyone to discuss things with, which makes it very difficult.

Author: Down-to-earth Teacher Chen

WeChat public account: Down-to-earth Teacher Chen