What is a data indicator system?
After looking at Baidu Encyclopedia, there is no entry for data indicators . It seems that this term is not used much at ordinary times. Then you can only be lazy indirectly, check the meaning of the two entries of indicators and data separately, and look at the combination.
Data: Data refers to the symbols that record and identify objective events, and are physical symbols or a combination of these physical symbols that record the nature, state, and relationship of objective things. It is a recognizable, abstract symbol.
Indicators: parameters for measuring goals, expected indexes, specifications, and standards that are expected to be achieved, generally expressed by data.
——Source "Baidu Encyclopedia"
Data is a summary of the results of things, and indicators are methods for measuring goals.
Combined, the data indicator is a target measurement method that can summarize the results.
In human terms, it is possible to quantify the result of a certain thing and form a numerical measurement method to measure the goal.
Data indicators are the embodiment of a quantitative way of thinking, which has at least two functions:
1 Can’t think of data indicators, indicating that there is no clear understanding of this matter (what the team has to do)
2 Think clearly about the data indicators, but can’t make them, which means that you lack control over the entire team and can’t promote the implementation
It is impossible to establish data indicators, and it is impossible to be data-driven at all, so the data indicators are actually a pointer to truly reflect the state of our team and the state of what we are doing .
The reason is that organizational execution and product health need to be quantified to some extent
Among them, the traction index corresponds to our business data index. When the traction index is unhealthy, it can warn whether the team's direction is deviated from the goal. The leader should consider adjusting the goal or correcting the team's direction.
Combined with data analysis, data indicators are to split and combine complex and abstract businesses, and find measurement methods that can intuitively and clearly measure these combinations, and can be quantified by numbers. At the same time they are independent of each other and can be exhausted.
But to fully measure a transaction or business, one data indicator is often not enough. Just like describing a person, just describing a single dimension such as height, weight, etc. cannot reflect the whole picture of a person, and a single data indicator cannot reflect the overall situation. At this time, an indicator system needs to be established - a series of logically related data Indicators to evaluate business conditions through multi-dimensional data indicators.
For general Internet industries or products, the data indicator system is the main way to systematically reveal the business level and user behavior.
Why establish an indicator system?
The essence of data indicators is to use data to speak and accurately signal the business.
1. Unify the standard to measure the quality of business
Traditional enterprises or small enterprises may not have the concept of a data indicator system, nor will they make great efforts to build a data indicator system, but they cannot completely break away from it. Data indicators are more or less involved, but they are not comprehensive enough, cannot be unified, Fragmented.
Generally speaking, the quality of a business depends mainly on financial indicators, such as revenue, gross profit margin, and net profit margin. For some innovative and exploratory businesses, you may pay attention to the number of users, GMV , conversion rate, etc. No matter what stage the business is in, we all need some data indicators to measure it.
Without indicators to systematically measure the business, we cannot control the business development, measure the quality of the business, and see whether the business development has reached the staged goal. Moreover, for some complex businesses, the measurement of a single data indicator may be one-sided. It is necessary to build a systematic indicator system to comprehensively measure the business development and promote the orderly growth of the business.
When the organization has a comprehensive and unified data indicator system, it can unify weights and measures, reduce conversion, translation (caliber interpretation), and other tasks, and reduce communication costs within the organization.
2. Guide product development and operation
The R&D and operation of products are actually very dependent on data support. Data indicators can not only help you see the results of business development, but also help you see the process of product R&D and operation, adjust strategies in a timely manner, and achieve goals without fail.
For Internet companies, the product R&D and operation departments are the core organizations that promote the company's development. Through a comprehensive data indicator system and data analysis, they can effectively focus on work goals and guide members' work. At the same time, a clear relationship between indicators at each level within the indicator system can be established, and the focus of work can also be clarified starting from the indicator system. Finally, be data-driven, find deficiencies, and improve performance.
3. Help build a data analysis system
The data indicator system is the first step in the data analysis system. The essence of data analysis is to find business problems and predict business results based on changes in data indicators. Data analysis work is meaningful only under the guidance of the data indicator system.
A complete data indicator system business can make data collection more purposeful and avoid omission or lack of indicator data during analysis. Although some data analysis software can handle missing data values, if there are no indicators, such missing values must not be handled by the software. In particular, the lack of key indicators will cause the credibility of the analysis results to decline.
The ultimate goal of the data analysis system is to help the organization build a set of operational information feedback mechanisms internally, which can continuously discover problems, warn of risks, and help decision makers to "make decisions before acting, knowing what to stop and gaining results . "
For example, when we measure the early stage operation of an official account, we can use a core indicator — the number of new users yesterday.
If the number of new users was 1,000 yesterday , I suddenly feel that the official account is operating well. But add the indicator of the number of new users the day before yesterday. If the number of new users the day before yesterday was 2,000 , then the number of new users has dropped by 50% . We added a comparison indicator, which made our understanding of the development of this business completely different. If we add more indicators, such as reading volume, open rate, etc., we will have more understanding.
The process of continuously increasing indicators above is the process of sorting out the business indicator system. A data indicator cannot measure business development, but an indicator system can clearly explain the problem.
For an organization, a good indicator system can be a ruler with a unified communication language , a Sinan with a unified direction , and a think tank that continuously discovers problems and warns of risks
What phase of construction?
The construction of the data indicator system is complementary to the development of the business. When the data indicator system is relatively complete, our business should also be relatively mature.
If the business has just started, it is difficult and impractical for us to build a complete data indicator system.
Even if there is barely, such a data indicator system is not rooted, because the business is constantly changing, and the operation method will also be constantly adjusted. Most of the data indicators need to be extracted and summarized from the business results and business operation process.
Only when the business is relatively mature and the operation mode is relatively stable can our data indicator system achieve initial results and operate effectively.
But it’s not that we shouldn’t invest when the business is immature. Except for some data indicators that may run through this business stage, we should explore and refine the data indicators that should be paid attention to in each stage of the business, and iterate continuously , which changes as the business changes.
For example, financial indicators such as revenue and profit margin should be paid attention to during the entire development stage of the business. In addition, in the early stage of business development, we may pay more attention to indicators such as the number of new users, conversion rate, and new acquisition costs. In the later stage of business development, we may pay more attention to indicators such as activity rate, retention rate, and operational efficiency.
The data indicator system is not a Rome built in a day . It requires continuous investment. There are different small goals at different stages of business development. When the business is stable, these small goals will converge into the final big goal.
Therefore, we should invest in the initial stage of the business, not only to help the business stage goals, but also to contribute to the final data indicator system.
resource requirements
The data indicator system seems to be a very professional thing, and it needs a very professional person to do it, but it is not entirely true.
The construction of the data indicator system does require some professional data personnel and some tools, but this is not the most important thing.
As mentioned above, the purpose of data indicators is to measure the quality of the business and help business development. Therefore, the most important thing in the construction of data indicators is to be familiar with the business enough to be able to go deep into the business. The knowledge and understanding of the business is even more responsible than the business. people.
It seems that the boss or the person in charge of the business should be the first person in charge of the construction of data indicators, which is indeed the case ...
In actual operation, the data indicator system is generally built at the request of the boss and the person in charge of the business, and can only be promoted with the authorization of the boss or the person in charge of the business.
Because the construction of the data indicator system involves product development, operation, sales, and even finance, human resources and other aspects, it requires strong coordination ability.
Therefore, the person in charge of the construction of the data indicator system should preferably be a senior data analyst, product manager or operation person, preferably a student who has been following the business development, which can greatly reduce the cost of being familiar with the business.
In addition, it is best to have a good relationship with the boss or the person in charge of the business, and have a stable communication and reporting channel, because they are the biggest beneficiaries of the data indicator system. In this way, we can communicate at any time, ensure the consistency of information and cognition, and at the same time increase our influence and coordinate the resources of all parties more conveniently.
Other human input also requires some data product managers (or data analysts) and data development students, who are mainly responsible for execution.
Data product managers or data analysts need to define the concept, caliber, etc. of data indicators, and organize them into a book, which is convenient for all parties to consult and unify their cognition. In the later stage, data indicators must be visualized and analyzed. Data development students need to clean the data according to the caliber of data indicators, establish a good data model, and facilitate data analysis students to use.
Of course, data cleaning may also require the cooperation of R&D, IT , operations, sales, finance, and human resources, because the data required for indicators not only comes from business systems, but may also come from various places such as sales systems, financial systems, and human resources systems.
In addition to human input, some data development tools and data analysis tools may also be required. These tools can be self-built or purchased. If you build yourself, you need to invest more manpower, but it may be more cost-effective for small and medium-sized enterprises or teams to purchase.
In general, to build a complete data indicator system that can be put into practical use, the investment should be very large
Organizational structure adaptation
As mentioned above, the data indicator system is only the first step in the construction of the entire data analysis system. After the data indicator system, there is still a lot of data analysis work. This is the stage of using the data indicator system to generate more value.
Therefore, our organizational structure is not only set up for the construction of the data indicator system, but may need to be set up for the collection and use of data by the entire company or team.
According to the previous experience in building a data center, this team needs to have the ability to share public data across business departments and be able to undertake the responsibility for building a data center, which includes the ability to build a data indicator system.
In order to be able to fairly and fairly measure the pros and cons of each business, it must be a department that is independent of the business team, and the head of this team should report directly to the boss or relevant executives.
In order to avoid being out of touch with the business, the organizational positioning of this team is to understand the business, be able to go deep into the business, and take root in the business. Within a team, there can be three sub-teams:
- The data analysis team, which is the core team for data indicator construction, is responsible for the planning of the data indicator system, the definition and maintenance of indicator caliber, and the output of analysis reports, etc.;
- The data platform team is responsible for building a platform that supports the data indicator system, including indicator systems, metadata centers, data maps, etc.;
- The data development team is responsible for cleaning data and data modeling, maintaining the public data layer, presenting the results of each data indicator, and meeting the customization needs of each data indicator.
Appropriate team composition and organizational positioning are essential tasks for building a data indicator system. It is best to be an independent department. At the same time, it must avoid being out of touch with the business, be able to go deep into the business, and be bound to the business goals.
What is the path?
The first difficulty in the construction of the data indicator system is the confusion of indicator management, such as the following:
- The same indicator name, different caliber;
- The same caliber, the index name is not the same;
- Indicator caliber description is not clear;
- Metric naming is difficult to understand;
- Indicator definitions and calculation logic are not clear;
The above problems may be forgiven before there is no dedicated team in charge of the data indicator system, but with a dedicated team, they should not appear.
Therefore, the first step in the construction of the data indicator system is to establish a good indicator management specification, iterate and update the indicator content according to business needs, it is best to establish an indicator management system, which can update and maintain our indicator content more conveniently. There are also some tricks to follow in indicator management, such as:
It can be oriented to subject domain management, splitting atomic indicators and derived indicators, formulating indicator naming standards, and classifying indicators for hierarchical management.
For hierarchical management of indicators, we generally divide indicators into four levels.
The first level is the North Star indicator, which is the most important and only indicator of the company. When other indicators conflict with it, it shall prevail;
The second level is the company-level indicator, which is an important indicator that the company pays attention to, and there can be more than one;
The third level is the department or product line indicators, which are generally the indicators concerned by the department or product line;
The fourth level is generally business process indicators, which reflect the indicators that need to be paid attention to in the business operation process.
The so-called data indicator system must be meaningful when it can be measured by data. Therefore, the second step in building a data indicator system is to establish a data model for each data indicator and provide data support.
The key to building a data model is data collection and cleaning, which is very dependent on the completeness of each company's informatization construction. For general operational data, the data warehouse team can handle it well.
If it involves sales system, financial system, and human resources system data, it will be more troublesome, especially the purchased systems of different manufacturers, which require a lot of cost to get through each system, otherwise it will take a lot of manpower to extract and split various data. This workload is huge, and it is error-prone and inefficient. The biggest headache is related human coordination.
Regardless of data collection and cleaning, data model building is actually a test of our data warehouse design capabilities and model development capabilities. Of course, there are some ready-made tools and platforms on the market, which can be done without strong technical skills.
But there are also some points that we need to pay attention to. For example, try to avoid decentralized and chimney-like data warehouse models. It is best to build them on a reusable and shareable platform. You can also evaluate the model with completeness, reusability and standardization. Whether the design is good or bad, these can improve the efficiency and quality of our development.
The last step is the presentation and data analysis of the indicator data. Only when the data with data indicators is fed back, the data indicators are meaningful. We generally establish a kanban system or reporting system for the data indicator system.
In the more advanced stage of use, the function of self-service data retrieval can be realized, allowing business personnel to independently obtain the data related to the indicators they need, breaking the solidified analysis thinking of reports or Kanban, and not relying on analysts for everything.
In order to be able to conduct more comprehensive data analysis, it is also necessary to realize full-dimensional drilling of data, because analysts generally can only rely on experience to judge which analyzable dimensions an indicator has.
If our indicator system can provide all the analyzable dimensions of an indicator, and can present the value of the indicator in each dimension according to the needs, and even drill down through different dimensions, it will be easier to find out the reasons for the fluctuation of the indicator. The reason, this is omni-dimensional drilling.
In this way, data-driven lean operations can be realized, and a data-driven business closed loop can be realized from target quantification, continuous tracking, abnormal diagnosis to decision feedback.
epilogue
The data index system comes from
To solve business problems, we must first figure out what problems exist in the business
Therefore, what business problems the data indicator system can solve is the most important thing. To be able to find the reasons that affect the business based on the appearance of data indicator changes and help solve this problem, the boss or business side will recognize the value of the data indicator system.
Similarly, the value of the data indicator system ultimately comes back to the business value. The data indicator system cannot directly generate business value. It needs to go deep into the business, extract valuable indicators, and establish a data evaluation system to feed back the business.
But generally speaking, the number of points does not understand the business more than the person in charge of the business, and it is easy to become a team that produces reports. How to go deep into the business, how to 1+1>2 requires more thinking, at least I have no answer now ...
Otherwise, in the event of layoffs, this kind of team that cannot explain its own value will be very dangerous