Understand one article: four types of data index system

Many students asked ：“ Is there any universal 、 General index system sorting methods ？” Common online index system sharing , Most of them are Internet AARRR A class , In reality, the situation is very complicated . Of course, there are common methods , Based on business logic , Sort out the index system . In essence , The data index system includes 4 Large type , For four different business logic . Today, let's systematically introduce .

Evaluation type

“ How about the quality of this product ?”

“ How effective is this activity ?”

“ Is this customer group worth doing ？

This kind of data indicator , Evaluate a thing well / Bad index system , It is the evaluation index system . Be careful , Data index system is not necessarily used for evaluation , For example, how about the product quality , You can simply look at the sales volume . But just looking at one indicator will cause problems , For example, good sales , But the profit is low ; Good sales , But the reputation is poor .

Evaluation index system , The main thing is to solve ： Single standard evaluation has deviation problems . Therefore, when building such an index system , The point is to consider ：

1、 Reflect the quality from multiple indicators

2、 Try not to overlap each indicator

3、 Distinguish the importance of each indicator

4、 Comparability of various indicators

Evaluation index system , The relationship between indicators is often juxtaposed （ Here's the picture ）：

Difficulties of evaluation index system , It mainly considers the comprehensiveness of the problem and the feasibility of index data collection . Special attention here ： Data index collectability . Many students mention evaluation , Will casually say ：NPS、 User satisfaction 、 Indicators such as user intention . The problem is ：

1、 How to collect these indicators ？

2、 Using the questionnaire , How many real users can be covered ？

3、 Use other indicators , Can we reasonably evaluate the situation ？

4、 Intention this kind of prediction index , How to predict ？

If you don't answer these data collection questions clearly , Even if the indicators sound good , Can't land , You will also find challenges .

Process type

“ How about our sales ”

“ How is our production progress ”

“ How about our research and development ”

This kind of data indicator , Indicates the progress of a process 、 Complete the result , It is the process index system . Here is the most familiar transaction process （ Funnel analysis ） Model . actually , As long as meet

1、 There is a definite end

2、 There are several steps

3、 Every step involves rolling resources

Can use similar funnel analysis , Building data index system .

It's not just the transaction process or user retention , production 、 Research and development 、 Purchasing can consider the problem like this . But in these processes, there is no attenuation at every step , Therefore, there is no logic of funnel . In these processes , Consider the delivery time 、 quality 、 cost （ Here's the picture ）.

The process index system is the simplest of the four categories . Because of the process 、 The finish line is clear . When the goal is clear , The index system is easy to sort out . Process index system , What I fear most is data collection , Especially process data collection . quite a lot toB The analysis of the industry cannot continue , It's because there are too many missing process indicators .

Inclusive type

Inclusive index system , Generally, a large indicator is disassembled downward into several sub indicators / Analyze the sum of dimensions . We are familiar with DuPont analysis is this logic , Break down a big indicator .

Inclusive index system , Generally used to diagnose problems . Because sub indicators + Classification dimension , Be able to specify the problem to a certain business department and action , So we can better find the source of the problem , Find solutions .

But notice ： Diagnostic ability of inclusive index system , It's based on ：“ The main index itself can explain the problem ” On this basis . For example, when using DuPont analysis , By default ：“ profits ” Is the main problem . If profit doesn't tell the whole story , But also consider the customer experience 、 Market share and so on , You can't expect an inclusive index system to solve all problems , Need to answer every question , Establish an index system separately , Answer .

Influence type

It is called influence type , Because operation actions are usually superimposed on the normal process , Produce additional effects . Normally, there is a sales rhythm , Doing an activity will additionally stimulate sales ; Normally, there is a user retention curve , Be a member reward , Extra incentives for users to stay longer . Influential data index system , Is to ： superposition 、 Two additional points are clear .

here , The indicators to be considered are more complex

1、 A set of data indicators is needed , Reflect the normal pace of business

2、 A set of data indicators is needed , Describe the superimposed action itself

3、 Standards are needed , Judge the performance of the business itself

4、 Standards are needed , Judge the additional effect

therefore , The influential index system will be particularly complex . For example, an index system for commodity promotion activities , It needs to be considered

1、 Normal sales indicators （ Quantity of goods sold 、 Number of buyers 、 amount of money 、; profits ）

2、 Operation of the activity （ The target number of people for the activity 、 Number of qualified people 、 Number of awards received , The activity itself is invested ）

3、 Normal sales trend VS Activity increment

There may be several algorithms for activity increment . such as

1、 In terms of time , You can distinguish between activities / No activity time

2、 From the perspective of commodities , You can distinguish between activities / No active products

3、 From the crowd , You can distinguish between activities / Inactive people

But each method is carefully considered , There are places where it doesn't work . Because even if there is no activity , Commodities also have their own sales trends （ Natural growth rate ）, It's hard to find 100% Similar goods and people . And product promotion , There may also be overdraft effect （ Fans of products stock up in advance while they are cheap , Lead to subsequent reduction ）

So when constructing the impact index system , Set reasonable criteria , Very troublesome , There are often quarrels . Of course, in principle , What meets the business needs is a good standard , It's not necessary to consider every effect in detail . But remind the students , Be extra cautious when working in this part , Casually say a few indicators , It will bring trouble to the follow-up work .

Comprehensively use the four kinds of index system

The above four categories are distinguished , It can help us sort out our ideas , Deal with problems at work . Because in work and interview , People who ask questions rarely take the initiative to distinguish ：“ What is the scene ？ What logic is business ？” But in general ：“ How do I evaluate this XX？ I want to see what indicators ？”

here , For data analysts , You have to stay awake all the time

A simple example

“ Analyze our products ”

“ Analyze the revision plan of our product ”

“ Analyze the revision effect of our products ”

“ Analyze our product revision feedback activities ”

There are four completely different scenes , The supporting index system required is completely different

meanwhile , What product is it , Also have to look carefully

1、 Trading products ： The main process is trading , Facilitating the transaction is the ultimate goal

2、 Content products ： User behavior is scattered , There are many ways to realize , Have to observe from multiple angles

3、 Tool products ： The function is fixed , But the breadth of user use 、 There are differences in depth

These are the elements that need to be considered when building the index system . And that's why , I just recommend students , Don't try to “ Read and recite ” Some data indicators , But the ability to sort out the data index system , In the future, we can respond to changes with the same . Limited by space , It's not going to unfold one by one , interested , This is a complete collection 60 I'm looking , Our next article will share the index system of tool products , Please wait .

Common data index system sorting methods , I have a systematic column in knowledge planet , common 18 speak , Interested students can join the knowledge planet to view . There are many other dry goods in the knowledge planet .

Knowledge of the planet Four highlights

What to watch for a ： Data analysis ability column

from 0 Start , The system improves the analysis ability and data thinking . Include 7 The main content , So far 《 Nine methods of data analysis 》、《 Sort out the data index system 》 Two video courses .

What to watch for two ： Business analysis column

In different analysis scenarios , Data form 、 Analysis focus , Analysis ideas will be different . So far 《 Business analysis - Introduction 》《 Business analysis - Advanced 》 Two video courses .

Third point ： Data analysis and job counseling

Job hunting is not exactly equal to learning , Job seeking , How to choose the working direction , Use your strengths , Pack your resume , It's important to prepare for the interview . So far 《 Job search guide 》《 Business interview guide 》《 Technical interview guide 》 Three courses . And provide Mr. Chen to revise his resume one-on-one 、 Job interview guidance service , No limit of times .

What to watch for four ： Question answering and report sharing

Ask specific questions on the planet , Mr. Chen will answer in great detail , Share the reports of various industries from time to time , Read it . At present, it has accumulated 300+ Your answers and 150+ Information .

The contents of the column are constantly updated , Looking forward to your joining ！

existing 1000 Small partners join first

Scan the code

   

Click on “ Read the original ” Learn more about

Original selection ：

• 5 Pictures , Understand the data analysis system

• User churn , How to analyze ？

• 5 A question , Clarify the success of user portrait