High-level operations need data Sense

Data operation can be regarded as a position or a kind of thinking.

In reality, only large companies are equipped with a data operation or BI (business intelligence). For many small companies, the work of data operation is permeated in every operation post, which is a kind of thinking required by every operation, meaning that data is used to guide business decisions and drive business growth.

For example, after looking through many job requirements, I found that the position of operation manager or above, or the position related to user growth, and the salary above 8K, all have certain requirements on data operation ability.

I. Data acquisition

The premise of data analysis is data acquisition. The trouble of many small and micro individuals is that they do not know how to obtain data. Even what data indicators, dimensions do not know, let alone use these data to fry a table of dishes.

Find data wherever your business is built. For example, your business is very simple, is wechat group + public number + like, then you see the public number, like the background data, and then add a small U butler to statistics within the group data can be; If your business in micro blog + Taobao, then look at micro blog background plus Taobao sellers can order business staff on the line.

If you pay for knowledge, the background data of Ximalaya and Qiancao can also do some analysis.

If you are an independently developed website or APP, you can subscribe to third-party data platforms such as Growing IO and Youmeng to collect and analyze data.

In addition, it is easy to ignore the big data in the industry, and how to use some open big data for me. I'll focus on that.

Most of the big data in the Internet are in the hands of BAT, and some of them are open to external use.

Baidu has big data on search, and its index is free and open to the public. It's on the web. Baidu Index is a data sharing platform based on Baidu's massive netizen behavior data. Here, you can research keyword search trends, insight into changes in Internet users' needs, monitor media public opinion trends, and locate the characteristics of digital consumers.

Take an example of an application. For example, I am a parent-child travel company, and I want to do autumn travel activities for children and families. But the whole of September and October is the time frame for autumn travel, so when is the best time to do it? It is a very inefficient method to research with kindergartens and mothers. Then I can search the word "autumn outing" on Baidu index, and I will find that the results of the past three years all tell us that the third week of October is the peak, and the third week of September is the second peak. Then our campaign can be promoted at the beginning of September as a one-month series with two peaks, or we can only do the campaign in October.

Tencent, on the other hand, has big social data, including the wechat index, which is free and open to the public. The app can be found on wechat. The popularity of the wechat index comes from a comprehensive analysis of wechat searches, articles on official accounts and publicly forwarded articles on moments of friends. For example, if a movie is released at the same time, I can search to see which one attracts more attention on wechat. If a new media editor or content operator wants to write film reviews and follow hot spots, I can refer to it.

Now look at Ali. It has big data on transactions and payments, among which the Alibaba Index is free and open to the public, from which media, market researchers and others can obtain the data analysis of geographical and industry indexes based on the e-commerce data of Alibaba. Two functions of regional index and industry index have been launched:

L Regional index: you can see the development of transactions, trade flows, commodity profiles and people characteristics between the two regions. For example, when I come back to Sichuan from Zhejiang, I might want to know about the economic exchanges between the two places, what things are sold from Zhejiang to Sichuan, and what things are sold from Sichuan to Zhejiang.

L Industry index: interpret transaction development, regional development, commodity profile and crowd characteristics from the perspective of industry. For example, I can see the hot search terms and the rise and fall of the wedding dress/Qipao/dress industry. When I searched, it was during the annual meeting, so evening dress and dress skirt grew fast. You can also see where people are buying more and where people are selling more.

There may be some business opportunities to help with decision making.

In addition to the three commonly used and free big data products of BAT, the following are commonly used:

Check the APP list: ASO100https://aso100.com/rank

View the public ranking: new list https://www.newrank.cn

Small program ranking: Aladdin

Short Video Ranking: Cass data

Second, data analysis

The next step is to do data analysis. Here we introduce six common analysis methods of data-driven growth, which are transformation funnel, DuPont analysis, quadrant analysis, user stratification, A/B test, and thermal map.

The first is the transformation funnel. Most of the commercial realization processes can be summarized as the funnel, such as the registration transformation funnel and the single funnel of e-commerce, which are widely used in activity operation and are the most commonly used analysis methods.

The second is DuPont analysis, which is used more for e-commerce operation and category operation. Mainly through disassembling the influencing factors, find out the crux of the problem for analysis, such as the website transaction volume decline, is the flow or conversion rate or customer unit price problem?

The third is quadrant analysis, which is mainly used for channel operation. Based ON TWO IMPORTANT ATTRIBUTES (indicators) of the PRODUCT as the basis for analysis, as the horizontal axis and vertical axis, the things to be analyzed are projected into four quadrants.

The fourth is user stratification, listen to the name you can guess the user operation with more, the user in line with a certain characteristics of stratification, to do targeted marketing.

The fifth is A/B test, which is often used to test design schemes, copywriting, product functions, etc., and is often used in product operation. According to the variables to be tested, two schemes A/B are given, and the users are divided into experimental group and control group. Different schemes are seen at the same time, and the effects are compared.

The sixth is heat map, mainly used for content operation and product operation. You can get a general idea of the user's product access preference, such as the more the user clicks, the darker the color.

For the specific usage and scenario of each data analysis method, I have made a table below. If you are interested, you can click to view it.

Loading in...

Third, cultivate data thinking

I'm going to talk about four ways to foster data sense that are more pervasive and understandable.

Just like we often talk about web sense, or business sense, data is a sense. This sense can be cultivated.

Here are four simple ways to cultivate your data sense in your daily work:

① Fun number guessing. For example, the team guess the peak trading before the e-commerce promotion, the product after the launch of the daily live and transaction volume, which is the most accurate guess has a reward, and he said that according to what logic to speculate, often there will be a fantastic idea oh, team cohesion will also be strengthened!

② Make useful gadgets out of data. Dull and difficult to digest data into the scene, daily, a little bit of infiltration, it is easier to remember. Remove sensitive, confidential data and often used conclusions, such as user portraits and user behavior data, and make them into posters, calendars, or brochures that operators can flip through at any time. Because sometimes he has to get historical data for analysis because of a small knowledge point, he may give up because of laziness and turn to pat his head. We can make it easier for them to get the knowledge. The company I worked for before, Feizhu, made a booklet, including user characteristics and query and purchase behaviors of hotel, vacation and ticket users, which was almost ruined by me. It was very useful to help me make some decisions when doing activities.

[images]

③ Always use data to talk. Limit yourself at all times. In meetings, emails, and reports, avoid blurring numbers if you can use them. This includes words like "a lot," "a lot," "a lot," "a lot," "a lot," and "a little." Learn to turn some qualitative descriptions into data, such as "frequently participate in internal sharing" into "participate in internal sharing twice a month", "become more fashionable brand" into "increase the customer price by X%", "recruit X fashion experts", "get X fashion media coverage" and so on.

④ When you see any data, don't jump to conclusions or believe it at first sight. Ask yourself what's behind it and whether there are other ways to interpret it. An absolute value is meaningless, going to see a year-on-year, sequential Numbers, compared to the market is also very important, although you are up 20%, for example, the market rose by 30%, but industry you hide behind it, will give a person a kind of the appearance of your business is thriving, but outperform the market is the real progress, or to advance is to go back. In addition, there are seasonal factors and abnormal points in some industries, which need to be taken into consideration. For example, Golden Week tourism and Double 11. If you take them as ordinary data and sum them up for average, it will be unreasonable and may bury some problems.

Four,

In a word, we should deliberately train ourselves in life, speak with data, do not blindly believe the conclusion of data, ask why, and cultivate data thinking. I wish you a senior operation soon.