Shulou(Shulou.com)06/01 Report--

This article mainly introduces what the operation must master data analysis thinking, has a certain reference value, interested friends can refer to, I hope you can learn a lot after reading this article, let the editor take you to know about it.

Data analysis is a fine operational work, we must establish a systematic thinking, do not blind analysis, rough analysis.

For operational data analysis, I believe many partners will have the following problems:

In the face of abnormal data, there is often the subjective conjecture of "what seems to have been done? what seems to have happened? so it may have caused an impact"?

In the face of data report, I don't know how to analyze it. Don't know what to analyze?

Data analysis as the most basic skill of operation, whether you really give full play to its value, qualified operation must be data-driven operation, not operation-driven data!

1. Thinking from a single dimension to a systematic one is a necessary change in data analysis. For data analysis, you need a systematic data framework!

When we consider problems, we will follow a train of thought, that is, from macro to micro, from global to local, data analysis is no exception. There is no need to say much about the position of data analysis in product operation here. Data analysis must be based on a detailed understanding of the product data system, and a data system needs to be established in mind when doing data analysis. the product data dimension system can be divided into three levels: macro data, meso data and micro data:

The above data analysis latitude does not include all the data latitudes of the products we operate. When doing data analysis, we need to combine our own product situation to do useful data screening. Of course, the operation must be the basic data requirement when putting forward the background demand, such as user profile data, PV number, UV number, UID number, startup times, retention rate, bounce rate, page access path and so on.

When many operators improve the data background requirements, they put forward a lot of data, and many of the data involve complex definitions and calculations, which will only increase the operation pressure of the background data and do not use much for operational analysis. on the contrary, it affects the efficiency of data viewing.

Operation data analysis can be based on the background basic data combined with Excel table export function, and with the help of third-party data platform for auxiliary analysis, which can not only reduce the background data development costs, but also greatly improve the efficiency of data analysis.

2. Data analysis needs to be goal-oriented, learn to split data dimensions step by step, and use structured thinking to do comprehensive and systematic analysis of operational data.

When doing the data analysis of product operation, we can follow the following ideas:

1. Determine the goal of data analysis

two。 Identify the key impact dimensions of data targets

3. Find out the relationship between different data latitudes and establish a data relation model.

4. Find the problem data and the cause

5. Optimize the influence dimension of the problem data.

For example, when we analyze the profit situation of Tmall's store, the store operation is most concerned about the turnover, but the most essential thing is the profit situation, which is analyzed according to the ideas mentioned above:

① data analysis goal: store profit analysis

② determines the key impact dimensions of the data target:

③ finds out the relationships between different latitude dimensions and establishes a data analysis model:

Profit = sales-cost = flow * conversion rate * customer unit price-(store fixed cost + operating cost + goods cost + personnel cost).

④ discovers problem data based on the data model:

To maximize store profits (L): l (max) = R (max)-C (min)

If the store is losing money, then it must be R < C, that is, the cost is greater than the revenue. Let's assume the following:

According to the above hypothetical thinking, we can draw a conclusion that when the cost is reasonable, the store has a loss, then it can be concluded that the sales are too low, the impact of low sales is due to the low flow conversion rate. Therefore, in view of this situation, what we need to do is to improve the conversion rate of the store.

⑤ optimizes the impact dimension of problem data: increase the conversion rate

We can improve the conversion rate through the following aspects:

Improve product packaging

Optimize the details page pictures and introduce copywriting

Optimize the payment path and experience for consumers to place orders

Improve customer service level and order promotion skills

Do a good job in user evaluation and management optimization

Implement corresponding promotion strategies, such as full reduction, full gift, discount, etc.

……

We continue to take product operation as an example. For example, we suddenly find that the DAU growth rate of the product becomes larger one day, and we sort it out accordingly according to the above analysis ideas:

3. Data analysis is more concerned about the correlation between multiple data dimensions than the causality generated by a single data! A data analysis model is established through the correlation of the data dimensions that affect the key indicators.

For example, we take the official account operation as an example, the key indicators of the official account operation are the number of fans and the number of articles read, and the number of fans and the number of articles read affect many latitudes. There is also a corresponding influence relationship between these latitudes, as follows:

When doing official account operation, you can try to sort out all the data that affect the amount of articles you read, and then filter out some relatively useful data dimensions, and then establish their relationship. In the course of actual operation, many partners of the operation only pay attention to how many articles have been pushed and how many fans have increased each week. In fact, they should also pay attention to some detailed data, such as the relationship between article title, content length, content type and the number of readings and reposts, and the influence of push time and frequency on the number of readers and the increase or decrease of fans. In addition, there are picture and text, plain text, the number of articles, official account single picture and text push, multi-picture and text push, the impact of headline push and non-headline push on reading volume, and so on, all of which need to be considered in the process of operation. and get into the habit of recording these data.

The most basic model in the process of community operation is the user's pyramid model. The establishment of this pyramid model is based on the user's activity and contribution value. The pyramid model divides the user into several levels. The higher the level, the greater the value of the user, the higher the contribution value.

Of course, the establishment of this user pyramid model must not be fixed, but will vary in hierarchy and the proportion of each level according to the specific community data, and the specific needs and operation methods of each level are different. For example, take the operation of a K12 education community as an example:

The improvement of the core data index of the number of community posts is related to the number of users in the whole community, the proportion of user levels, the transformation of user levels, user behavior at each level, user stickiness, and community content quality. there is a certain correlation between content display and push. Therefore, in the process of community operation, we should constantly promote the positive relationship between the impact dimensions and the number of community posts.

So how to establish the relationship between the number of community posts and other data dimensions? Brother Chao tried to do a simple carding, the corresponding data dimensions are not all included, this diagram still needs to be improved, here is just a carding idea, as follows:

4. In order to do operation, we must talk about data analysis and training into subconscious behavior, all behaviors and means in the process of operation can be digitized, and data-driven operation.

① cultivates systematic thinking of data analysis

There are generally two directions for data analysis, one is top-down, the other is bottom-up.

The top-down idea has been mentioned earlier, and the specific ideas are: establishing the goal of data analysis-- disassembling the target influence dimension-- establishing the correlation of each data dimension-- discovering the problem data and its causes-- problem data optimization. This idea is the establishment of a data analysis system or model for multi-user products, so as to ensure the comprehensiveness of data analysis.

The idea of bottom-up data analysis is mostly used to find data problems in existing data reports. the specific ideas are as follows: abnormal data discovery-influencing factors of abnormal data-the relationship between influencing factors and problem data-finding out the cause of abnormal data-finding the solution of abnormal data.

Sensitivity of ② culture data

There is no other way to cultivate data sensitivity, in addition to mastering the correct method of data analysis, is to look at the data every day, analyze the data every day, and speak with the data.

③ forms the habit of recording data

In the process of operation, there will be a lot of detailed data, which need to be recorded. When the number of recorded data is accumulated to a certain extent, the corresponding data rules can be found through the aggregated data, such as:

Records of UGC posts, hot posts and boutique posts for the community

Data logging for message center PUSH

Record the historical tweet data of the official account

You can even record the content of your daily work and the time spent on your work, so as to optimize your work efficiency.

The data must be rational and rigorous, so we need to treat it from a rational point of view. of course, different operating products and different data dimensions we need, we must learn to define the data. and to ensure its logic and rigor, to be able to withstand scrutiny.

Thank you for reading this article carefully. I hope the article "you must master the thinking of data analysis for operation" shared by the editor will be helpful to everyone. At the same time, I also hope that you will support and pay attention to the industry information channel. More related knowledge is waiting for you to learn!

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.