Four levels of business to data requirements

Absrtact: The importance of data has been more and more well-known and accepted by more and more companies and individuals, even the momentum of too much overkill. The concept of large data flying, it seems that overnight everyone is talking about big data, see the face without large data to greet, it seems not

The importance of data has been well known and accepted by more and more companies and individuals, and even the momentum of overkill. The concept of large data is flying, it seems that one night everyone is talking about big data, see the face without large data to greet, as if it is not in the data circle mixed. Then, by the external marvellous data, in what ways can promote business Take-off? Or, in other words, what level of demand does the business have for the data? Where can the data help the business?

Combined with the author's years of work experience and the understanding of data and business, the demand for data in business has been summed up at four levels.

First floor: Know it

We can set up a data monitoring system to grasp what has happened, how much, to "know it".

Specifically, there are several aspects of the approach to data. The first is "Viewing the heavens", observing the overall trend of the industry, the impact of the policy environment, and then "know the ground" to understand the performance of competitors, and finally "introspection", how they do their own data performance. In terms of the cycle of data viewing, "Viewing the Heavens" can be quarterly or longer cycles; "By week or month, special time point, special event case;" Introspection "Class of data to get the most comprehensive, need to see every day, special someone to see, someone to study."

On this level, share two views of the data:

1. The data is scattered and the data needs to be framed.

How to look at the data is very fastidious. Fragmented data is hard to value, putting data into an effective framework for overall value. The so-called effective framework contains at least twofold functions:

(1) A lot of data, different people on the needs of the data, such as CEOs, middle managers, the bottom of the staff concerned about the data is usually not the same, effective framework for different people to take.

(2) An effective framework can quickly locate the problem. For example, trading volume indicators are concerned about, if one day the transaction volume index dropped 20%, then, the business is likely to be a problem, but the problem in the end is where? If only a few highly abstract indicators, such as conversion rate, turnover number, customer unit price, etc., is not positioned to the problem. Good frame can support us to drill down, from the category, flow channels, etc. to find the problem, the board can hit the specific person in charge. This is what we usually say, look at the data to the ground.

2. Data, there is a comparison to the truth.

I have 120 Jin, you say is heavy or light? A lone figure is hard to explain. To determine the growth rate of an index, we need to choose the right comparison object, reference system, that is, the baseline. This baseline can be a predetermined target, either an industry average or a historical contemporaneous data.

Second floor: Know why

Through the data to see the problem, this step is not enough. The data is only the appearance, is used to discover, describes the problem, the practice solves the problem to be more important. Data combined with business, find the real reason behind the data representation, solve it. The problem-solving process involves data, data processing, and may involve methods such as data models or tools, which are relatively high in technical content and are not expanded here.

There are also two points to share in the second floor:

1. The data is objective, but the interpretation of the data may be highly subjective.

The data itself is objective, but the consumer data is the person with subjective initiative. People tend to take the subjective factor into consideration when interpreting data: the same data in a seems to be a good conclusion, and from B it seems likely to produce diametrically opposed results. This is not to say that the situation is not good, the truth more debate more clearly. But if it is not through the data to find the problem, but first to the problem qualitative, and then selectively use the data to prove their views, this practice is not taken. But in fact, this is something that happens to us very often.

2. Understand the business can really understand the data.

Cheping's blog, "Do not understand business, don't talk about data" on this point of view made a profound exposition, here do not start to speak. Only because of the importance of this point of view, the author deliberately take out to do the emphasis.

Third Floor: Discovery Opportunities

Using data can help businesses find opportunities. For example: Taobao has middle-aged and elderly clothing market segments, there are large size women's market, these markets can be through the surrounding environment of perception, to know that there are some middle-aged and elderly people around us or fat mm on Taobao has not been satisfied with the demand. So is there any other channel to find more market segments?

Data Can!

Through the user search keyword and the actual transaction data comparison, found that a lot of demand is not well satisfied, reflecting the demand exuberant, but insufficient supply. If found such a segment of the market, published to the industry small second, published to the sellers, is not it can help everyone better to serve consumers? This example is what we are doing now in the "potential niche market discovery" project.

In this case, not to brag about how powerful the data is, but to tell you: The data is there, some people are blind to, but some people can dig out "baby". What is the difference? Business sense. Just mentioned the search data, the transaction data many people can see, but no one in the previous two data linked to see, this is reflected in the business sense.

Layer Fourth: Establishing a data operation system

I understand that the data operation contains a twofold meaning: data as indirect productivity and direct productivity.

1. Data as indirect productivity.

Indirect productivity refers to the data workers to pass the data value through the operation to the consumer, that is, the commonly called decision support, data workers output statements, analysis reports for all levels of business decision-makers for reference. I call it decision Support 1.0 mode. However, as business development and business personnel increase understanding of the importance of data, the demand for data is springing up, and it is clear that reliance on a few analysts alone is not enough. Teach people to fish than to teach people to fishing, so that the operation, product students can conduct data analysis, is my brain Decision Support 2.0 mode.

Decision Support 2.0 mode has three key words: product, ability, willingness.

It is not realistic and necessary for the operation and PD to grasp the SQL such language, and to master the analytical work such as SAS and SPSS. Providing a low threshold and a good user experience is the basis of the 2.0 model of decision support. Here the product, not only the operation function set, but also need to carry analysis ideas and practical cases.

However, the threshold of data analysis has always existed. This puts forward new basic competency requirements for operations and PD, namely, Basic mathematical ability, logical thinking ability and learning ability.

The last will, perhaps the most critical, only the heart has a strong drive, want to do this thing, it is possible to do well.

2. Data as direct productivity.

The so-called direct productivity, is refers to the data worker to the data value directly through the foreground product effect to the consumer. The fashionable point of speaking, called the data to become present. With the advent of the big data age, the management of the company has paid more and more attention to this. The Big Data age presents great opportunities, but it can also be a catastrophe. If you can't use the data to produce value, it's a disaster-the more data you produce, the more space you can save and the resources you'll waste.

Now a better understanding of an application is related to the recommendation, you buy a product, you recommend a most likely to buy goods. Personalization is the new wave of data as a direct productivity, and the tide is getting closer. Data workers, prepare for the welcome.