1. Starting point of construction
Meeting business needs is the starting point for building a data analysis system, as well as the ultimate goal and the highest requirement. It should be noted that there is no uniform standard for "business requirements". Different departments and people with different identities have different needs. Broadly speaking, it can be divided into three levels:
1. Strategic level: senior management who can determine the overall direction of the company
2. Tactical level: determine the management of a specific functional work (sales, operations, products, after-sales...)
3. Combat level: front-line departments with no decision-making power but only executive power (salesman/customer service/auditor/warehouse manager...)
These three types of people require completely different data types, data timeliness, and data application directions. Therefore, the requirements need to be met separately (as shown in the figure below).
2. Data Analysis Serving Strategy
Throughout the system, business analysis directly serves strategic decision-making. When the top management makes decisions, it focuses more on macro issues, such as the achievement of overall goals, changes in the external environment, and the effects of internal initiatives. Instead of getting bogged down in trivial business details.
Therefore, when doing business analysis, it is necessary to:
1. Transform business objectives into quantifiable indicators
2. Monitor the progress of the goal and find out the problems in the process
3. Perceive changes in the external environment and give early warning of potential macro problems
4. Quantitatively evaluate the effect of various business activities on the target
5. Assess the benefits of various business activities and put forward directional guidance
Note: Accounting for operating results is very complicated and cumbersome. Many operational initiatives span weeks, months, and involve many departments and jobs. Some basic research and development, production line updates, and infrastructure investment span several years. Therefore, the frequency of business analysis is generally not very high, and is generally carried out on a monthly basis.
Decisions made at the level of business analysis are often directional, such as:
1. Stick to the original plan or make adjustments?
2. Sales/operations/products/marketing... who will be the main force and who will be the assistant?
3. Additional investment or replacement method?
These decisions directly impact tactical-level design. As for the specific design, it depends on tactical-level analysis to support it.
3. Data analysis for tactics
Analysis at the tactical level is specific to each function. for example:
Sales department: sales performance analysis, sales channels, sales methods, sales force analysis
Operation department: activity method analysis, promotion method analysis, platform operation analysis
Product department: product usage analysis, new version features, new version analysis
The specific content of these tactical-level analysis is often varied, but the core idea is the same:
1. Strategy formulation: From among many tactics, choose one that can achieve the goal
2. Monitor the progress: monitor the progress of the tactical landing, find problems, and adjust the tactical design
3. Review effect: whether the review achieves the goal, accumulates experience, and solves problems
There are too many specific details, so I won't give examples one by one. Interested students can read the operation and product analysis methods shared before. In fact, most students who do data analysis are most often exposed to this level of analysis. The final output is also a daily monitoring report + a special analysis report.
4. Data Analysis Serving Combat
Strictly speaking, what is needed at the combat level is not data analysis, but data. The front-line work is so busy that no one has time to sit down and listen to the report carefully. Seeing the data is enough to take action. for example
First-line sales: see today's performance target, today's performance has been completed, and the list of customers to be followed up
Front-line customer service: See the traffic volume to be allocated, the number of queued calls, the number of complaints, and the results of complaints
Front-line warehouse management: see the number of goods in storage, the number of goods in transit, the number of goods that are expected to reach, and the number of goods that are expected to leave the warehouse
With the data, the front line can already take action. Hurry up and finish the unfinished tasks
It would be even better if some auxiliary tools can be added to the basic list. For example, for sales, not only has a list of customers to be followed up, but also an estimated natural consumption (labeled by the predictive model), which can help sales focus on people who should be actively followed up. For example, if you give more: customers can participate in activities/customers can forward posters, it will give sales an additional tool to impress customers. These tools are much more useful than long-winded analysis reports (as shown below)
Quite a few companies' data analysis at the combat level only stays at the stage of excel daily and ppt, lack of tool design and development, resulting in problems such as data analysis not being implemented, and unable to assist the front line.
Seeing this, some students must be curious: Teacher, my company is not that big, and the data is not so much, how can I make it more systematic? There is a way.
5. Small and medium-sized enterprises, how to go from 0 to 1
Start-up companies certainly don't have the energy to develop such a large data system. For start-ups, the key is to find a profitable MVP as soon as possible, and then continue to expand investment and enhance income capacity. Therefore, for start-up companies, they generally focus on sales data/promotion data/channel data, and do a good job of tactical analysis.
For enterprises with a certain scale, the most important thing is not to engage in various analysis reports (generally all should be available) nor to engage in complex analysis reports. It is to strengthen the infrastructure construction and make up for the shortcomings left by the rapid development in the initial stage. for example:
1. Commodity coding system, commodity grading and classification labels
2. Event coding system, event material coding system, coupon system
3. The financial system and the business system are connected, and the financial data corresponds to the business data
These may not only design the database design, it is possible that the old transaction system, logistics system, and cost control system need to be upgraded, and the business process must also be standardized, so it is a very huge project. But if you don’t pass this hurdle and continue to live on the old basis, you will find that the larger the scale, the more chaotic the internal system, the more complex the data, the more inconsistent the old and new data, and the more difficult it is to go forward.
In 2023, Mr. Chen has experienced several digitalization proposals for medium-sized enterprises with a turnover of 3 billion to 10 billion. Without exception, there are problems of weak infrastructure + great achievements. Often the most basic product data, activity data, and channel data are not well established. Instead, they are eager to get CDP, algorithm in APP/H5, and full-link burying. The result is naturally: building skyscrapers in the muddy land... All kinds of entanglements and pains are not a problem.
6. Behind the problem
All of the above problems will be obvious to anyone who is exposed to them. But why no one solve it?
-
It may be that the business department is arrogant and strong, and does not want to let the data participate, only for the excel table
-
It may be that the boss of the technical department wants to be promoted, and the infrastructure is not conspicuous enough, so new things must be introduced
-
It may be that the boss of the company has no knowledge at all, eats industry dividends to make a fortune, and lacks basic knowledge
These are all likely to stop the data in the original stage. Then they hope that a data analyst with great powers can solve all the problems, and they will eagerly hold your hand and say: "Our company has a lot of data, and it's all there. We need a master to analyze it." ..."
So if the students who do the analysis encounter:
-
Do one piece in the east, one piece in the west
-
Only write sql to organize excel
-
Disliked by business without depth