Building a comprehensive data analysis system

In today's data-driven business environment, building a comprehensive data analysis system is crucial for enterprises. This article details how to build such a system from two dimensions, "vertical and horizontal", to overview business processes and span different departments to ensure that data analysis can fully cover and support enterprise decision-making.

Today, I will give you the most comprehensive data analysis system, covering all company-level scenarios. You can refer to it and see where your focus is. The whole system can be summarized as: one vertical and one horizontal.

Without further ado, let’s get to the point!

01 A vertical: Looking at work from the perspective of results

Ask a simple but key question: What is the use of data analysis?

Answer: From a business perspective, data analysis has six major uses

1. Goal setting: determine quantitative goals and break down and distribute goals
2. Trend forecast: predict normal trends and provide decision-making reference
3. Process monitoring: monitor business development and discover process problems
4. Results review: review performance and summarize results and experience
5. Cause analysis: Analyze the cause of the problem and explore solutions
6. Method testing: test optimization methods and select better practices

(It is recommended to read and memorize the above parts)

These six scenarios run through the entire business process and best reflect the value of data analysis. Therefore, when we think about what data analysis can do, we can first think about these six scenarios. How much of the current work meets the needs? What else can be done (as shown in the figure below)?

Note: Data is not irreplaceable! The business side can do these 6 steps even without data.

For example, the classic scenario is:

1. The goal setting was done by the finance department at the instruction of the boss.
2. Trend forecasts are made by leaders on the spur of the moment.
3. After reviewing the results, it was found that the business itself was flattering itself.
4. Method testing, it doesn't exist at all (I said it is, it is!)

Most likely, only process monitoring and cause analysis after the problem is discovered are left to data analysis. But in this case, the work is too passive. Without knowing the goal, the basic business trend, and the business logic behind the method, it is difficult to analyze the cause by just looking at a number. Therefore, many data departments have degenerated into being able to only monitor individual data or even only provide numbers.

This passive situation should be avoided as much as possible when making plans at the beginning of the year.

If you don't fight now, when will you fight?

At this point you can:

1. Lobbying to big bosses, instilling the concept of "full-process data management" and increasing work scenarios.
2. Combine successful digital cases in the industry to promote the value of data to everyone and expand work.
3. Observe the working style of each department and see which departments are easier to negotiate with and where you can find cooperation opportunities.

To do this, we must carefully study the division of labor among departments within the enterprise, which involves the concept of "one horizontal".

02 One horizontal line: Looking at opportunities from a departmental perspective

Ask a simple question: Do all departments attach equal importance to data?

Answer: Of course it’s different!

From the perspective of departmental responsibilities, departments can be divided into four categories:

1. Management

Typical examples include: President's Office, Strategic Development Department, and Finance Department. These departments communicate directly with the top management of the company. Many major development plans, annual KPI targets, and business tasks are formulated with the participation of these departments. These departments attach great importance to data!

Because all goals, tasks, and plans must be quantified. Their common problem is: they know the results, but not the reasons, especially in the finance department. They are very good at accounting, but they don't know much about the business details.

At this point, if you want to make friends with these departments, you can start by providing basic data. While providing data, you can actively help them sort out key business processes, clarify the conventional business baseline, and make up for their shortcomings in business understanding. This will give you more opportunities for cooperation. (As shown below).

2. Income

Typical examples include sales, placement, and growth departments. These departments bear the main revenue tasks and are the main sources of performance and profits. However, these departments generally do not pay attention to data. It is enough to see the task goals and completion rates. They prefer to see case analysis, operation guides, and specific practices.

In short, good things are what you can use.

At this point, if you want to make friends with these departments, you can start with tools. Don't prepare complicated reports, but provide data in different levels according to the content that the viewer is concerned about. The lower the level, the less data is provided, and it is best to only keep key KPIs. In terms of functions, connect with CRM and other tools, and directly provide actionable functions while providing data, so that it will be welcomed by the front line (as shown in the figure below).

3. Cost type

Typical examples include purchasing, R&D, and design. These departments basically consume costs and it is difficult to see direct results. If they are not done well, it will cause users to dislike them, product backlogs, insufficient inventory, and affect sales... Moreover, these departments are affected by sales and marketing, so it is difficult for them to remain independent.

At this time, we should treat them differently:

For departments such as procurement, production, and supply that are prone to hard losses, focus on rolling data forecasting and data monitoring. Collect influencing factors from upstream supply, downstream demand, and promotional activities in a timely manner, and combine supply progress and inventory conditions to forecast possible backlogs/out-of-stock problems (as shown below).

For departments such as R&D, design, and product that are prone to soft losses, it is best to focus on testing platforms and testing services. Use regular monitoring to find problems, and use good tests to verify the improvement effect (as shown below).

4. Hybrid

Typical examples are marketing and operations departments. These departments like to look at data, and their work results are superimposed on sales, which is difficult to observe. Therefore, it is particularly difficult to do analysis. Fortunately, their work is mostly project-based, so they can break through one by one.

Common projects include:

• Big promotion
• User Insights
• Brand Communication
• Community Operation

These have been shared in many previous articles, so I won’t repeat them here. If you want to have a deep insight and analyze well, the key is to cultivate good habits in the business department and lay a good data foundation. For example:

1. Improve and maintain tag libraries such as user tags, product tags, content tags, and channel tags.
2. Management of activity labels and classification information, recording of activity assessment standards, and prior reference group design.
3. Basic data collection on community operation, new media operation, short video sales, and live streaming sales

Once the basic work is done, a large amount of data can be analyzed afterwards.

If the basic work is not done well, there is no point in post-analysis...

03 Overall arrangement: phased and benchmark projects

Note: The above work has a sequence of priorities within the business department.

1. Generally, January, February and March are the months for annual planning.
2. In the first half of the year, we recruited new people, formed teams, and conducted small-scale pilot projects.
3. Major promotions and festival activities are concentrated in the second half of the year

After clarifying the actions of the business department, the overall arrangement of the data department is very clear:

1. At the beginning of the year, the focus was on basic work such as forecasting, goal setting, and system building.
2. In the first half of the year, priority will be given to strengthening basic capabilities, and tools such as basic data, testing platforms, and label libraries will be strengthened as much as possible.
3. In the second half of the year, we will focus on major projects, provide project support, monitor and review, and provide user insights.

The output goal is to launch/update a project every month. This makes it easier to write quarterly reports and save time for the annual summary next year (as shown below).

The above is the overall idea of ​​data department planning. Of course, the specific situation of each enterprise is different. Students can tailor it to their own specific situation.