How can a good data analyst not believe in data?

For the attitude of data analysis, there are a few complaints to vent, purely personal feel in these years of work experience.

After the interview feelings

I have been interviewing this weekend, a total of more than 30 people, only half of them can go to my level, no matter how many years of work or a bit of work experience, whether it is a prestigious university or an ordinary school, they are right I am somewhat disappointed with the attitude of the data.

I asked them, if I was the CEO of Jingdong Mall, Monday morning you want to show me the last three data, what data would you choose?

Almost all people do not answer 1 second, such as traffic, conversion rate, trading volume and so on.

I went on to ask, did you hear my question, I said to the CEO. Most people will then gasped, perhaps CEO will not pay attention to such details of the data, it should be XXX data.

Then I asked again, I said last week's situation, you noticed the "week" this time? Then most people suddenly realized that yes, ah, weeks and months should be different, it should be XXX data.

Finally, I asked, this is to Jingdong CEO to see, not to the CEO to see customers. Did you think about it?

At the end, let them score for themselves. Everyone scored very low for themselves. The highest score was also six or seven points (out of 10).

I asked where the 4 points ah, they summed up three reasons, one did not expect to see the data is the CEO, and second, do not understand the Jingdong, the third is not sensitive enough to the data.

I do not want to take this issue to stumped them, their answers are not right or wrong points. In the middle of the question I gave them three opportunities to remind them that they can ask me questions, but also pay attention to the elements of the question. I did not consider the answer they gave, just wanted to look at their attitude and thinking about the data. Is not very respect for the data, do not know how to think of the data and business together, which in my opinion is an excellent data analyst with the basic conditions. It's hard to be a good data analyst because you have no passion or belief in the data.

Dissatisfaction interview reminds me Some time ago someone accused IriTie microblogging conversion rate too B2C fly this matter, some people say that the conversion rate should not be such a calculation, do not know how many experts in this pit It's falling down I think this thing is quite interesting, indeed, as this person said, do not know how many people fall into the pit do not know: we take the conversion rate to talk about things, but no one asked what kind of conversion rate, What is the numerator, what is the denominator? You say a conversion rate, I say a conversion rate. Market department said a conversion rate, website operations say a conversion rate. But what do you mean conversion rate in the end, very small people to dig the root problem. And what is the decision we make with conversion rate, or what is the real meaning of a conversion that we talk about barely?

The weirder is that today's online retailers do not seem to be unprofessional not to mention data, but few have actually respected the data and the data as a science.

Of course, this is not the fault of a data analyst, but the overall environment makes it difficult for our data analysts to do this. Based on my work experience over the past few years, I think it is actually very simple to destroy a data analyst's respect for the data. Killer can be seen everywhere. At the same time, data can be almost red-hot only when the data is really wonderful and sweet. Faith.

Destroy three common reasons for analyzing data attitude

First of all, the environment does not respect the data, especially the boss's attitude. If data analysts simply give a report, the numbers are a little more and less, and everyone laughs and does not go after it, it is hard for data analysts to treat the data in a serious manner.

For example, several domestic data analysis institutions are basically anxious to expand their industries and compete for occupations. Therefore, Ereli's data has only recently been often said to people not to rely on the accuracy of the data they put out. Spectrum. "

Data analysis, today is not allowed to do, tomorrow again is useless. Such as iResearch, if the data is not stable, rushing to do a lot of industries, this is not reliable practice, mean maybe smashed his own brand of day.

Someone and I mentioned FACEBOOK data analysts why so cattle, because they do not think data analysis is a painful thing, a dozen people in a house data analysis as a very happy thing to do, data analysis for them Is in the pursuit of science.

Second, a good data analyst needs a bit of talent, but also requires an expert call, but the circle of e-commerce, the real understanding of data analysis, no more than 10 people, so most people difficult to get the Scriptures. This is the same as faith. It is much more difficult to get in without a master.

I reviewed myself from Microsoft to eBay, then from Dunhuang to Alipay, and made a big leap in data analysis thanks to a lot of inspiration from the two teachers. One is Amazon's chief scientist Weasconang, once, I told Wesconkang, KPI report shows Dunhuang network takes 4 seconds, he immediately called me to do technical colleagues (he wanted to hear first-class students reaction), ask this 4 seconds How to measure the bell, the Americans open with 4 seconds, or the British open with 4 seconds, what is the use of Browser and so on. Is this 4 seconds related to business value (such as trading volume)? I was touched at that very moment. Even such a very basic data analysis was based on a confirmatory mentality. What impressed me even more was that when he was asked to be a consultant for a long time in Dunhuang Network, according to his work experience, it was easy for me to fool around for a long time. However, Weasconkang was very strict and first spent half as an ordinary person An hour to buy things in Dunhuang network (resolutely to pay the real), personal experience Dunhuang network user experience, and then do not look at the data, but many can learn more about Dunhuang network business problems. His questions are more professional than many investment analysts. Now many data analysts, including myself at the time, only speak of the data and speak no questions about the company's business model.

Weixikang told me the data is an attitude that makes me understand that people who do data are to devote one's whole heart and mind to one another. There is a lot of ways to go in the middle of data. And data and business are closely related and can not be confined to the dead ends of data .

Another is Xie Jinhong, a professor at Tsinghua University. One summer happened to attend his class and take a pile of data for him. While giving me a glimpse of his thinking, he could quickly find them in a pile of data The relationship between. Later, I used to go to Tsinghua University to find someone who often talked to him. He taught me how to read network data and view network data in a coordinated way. It can be said that he enlightened me to use the "relationship" thinking to see the data. Upon hearing Dunhuang Dun Dun went to see a lot of data, found a new world.

Third, data analysts can not afford to land the slightest bit of data and can not talk about business. If you do not know business, and simply look at the data, not only hard to have creative thinking, and it makes no sense (once talked about this issue, do not know business, do not talk about the data: http://blog.sina.com.cn/ s / blog_5025e3880100kwn1.html).

For the average data analyst, most people do not have the system thinking, but also can only see part of the data, can not understand the operating data of the entire company, so that it is difficult for data analysts to form a comprehensive way of thinking .

In my own work experience, for example, why I was in Dunhuang data analysis capabilities will be by leaps and bounds, but also because I can only see some of the data in the first two companies, but after I went to Dunhuang, what I love to see, thanks After I was inspired by the professor, I was quick to link the marketing data, market data, financial data, product data, sellers and buyers data, etc., which greatly changed the way I use the data.

Sounds easy to do

Compared to the profound method of data analysis, it seems that the reason is very simple. But in reality, as most people saw when they interviewed, most people did not realize that they were reminded.

Simple things to say, but really hard to do.

Sometimes, including myself, can we really make sure that every source of data is reliable, and sometimes not necessarily. For companies this attitude is even more important, today's independent B2C says how expensive marketing, but their marketing is really careful enough? Data analysts have not asked from the portal traffic is divided into several Channels, what is the conversion rate per channel? Is the conversion algorithm and data collection accurate? The user who can not convert from there and then go? Use this conversion rate to decide what is the point?

Just put a data out, not a qualified data analyst.

This is what I want to say a few words.