[concept] do you really understand the increasingly popular “data driven”?


With the continuous development of Internet and big data technology, more and more enterprises have realized the importance of data and know that they need to understand the business status through data statistics / analysis / mining to assist business decision-making. However, after years of shouting the slogan of data-driven, how many people really understand and understand what data-driven is and how to implement it?

Don’t worry. Before answering “what is data-driven”, let’s discuss a question. What does “Data-Driven” need? Before we understand the essence of the problem, we do things rashly, and the effect is often not very good. So, let me help you understand the concept of “data driven” more thoroughly. Only after you really understand the essence of things, can you really do things well. Whether enterprises need to drive business through data business or business data, their essence is the same.

What is “data driven”

What is data driven?

Data driven is to collect massive data by means of mobile Internet or other related software, organize the data to form information, then integrate and refine the relevant information, and form an automatic decision-making model through training and fitting on the basis of data.

In fact, this definition is relatively difficult to understand and inappropriate.

Now let’s answer three questions from a different thinking mode and perspective, and redefine what data-driven is

  1. What it can do;
  2. What it needs;
  3. How it does it.

The first question is “what can it do?” my understanding is to establish decision-making modeling through data statistics / analysis / mining and other technologies, so as to find out the current situation and problems of the enterprise in advance.

For example:

1. For an e-commerce product, such as Taobao, the sales of Gmv brought by the homepage banner has declined. How can we deal with such a problem? First, we need to collect data, analyze data, train business model through historical data, and find out the reasons for the decline of Gmv through automated business model combined with user use scenarios, to see if it is UI problems and product selection Then, the model outputs new strategy suggestions to optimize

2. For an information product, such as today’s headlines, CTR clicks recommended on the home page have dropped by 3% recently. In case of such a problem, we first collect data, analyze data, find out whether the user portrait and content portrait match in advance, find out the badcase of recall layer and ranking layer, and what the dynamic information of the corresponding competitive products looks like, and then combine with the platform’s own business Optimize the scene

3. A social product, such as wechat, has recently changed its UI interface, and the data of users sending their circle of friends has decreased by 2% compared with the previous month. The solution is to collect data, analyze the data, find problems in advance from the dimensions generated by buttons and text chains, the click position of buttons, the shape / size / color of operation buttons, and the decision-making influence of color, train the model, and do iterative optimization

The second problem “what does it need?” requires data and experimental platform to find problems by analyzing the collected data, and then combined with business scenarios to do models and experiments, do a lot of experiments, and finally optimize problems, solve problems, and make decisions.

Maybe you feel that what I said is nonsense, but actually it’s not.

Data is like oil. Without it, we can’t drive. After collecting the data, we should promote everyone to look at the data and unify the team members’ cognition of the data and business. After all, it is useless to collect the data alone, but we still need to use it in a scientific way.

Third, how does it do it?

[concept] do you really understand the increasingly popular

First of all, in the era of big data, as long as the implementation of artificial intelligence projects is involved, the following basic elements are required: data, algorithm, scene and computing power. These four are indispensable, they will become the new core competitiveness of the enterprise, and then landing the same need to form a closed-loop system, continuous data collection, data asset accumulation, combined with business scenarios to complete data modeling, with the help of algorithms and computing power to provide support, to achieve the final automatic decision-making.

Secondly, data driver needs to have the following points for data acquisition:

  • Large amount of data: we should fully consider the growth of user scale and data scale, and be prepared for the accumulation of data assets
  • Full data: need to collect a variety of data sources, through a variety of methods for full data collection, collected data to run through the whole life cycle of users using the product
  • Detailed collection: it is necessary to collect enough data through multi-dimensional and multi index methods to make the accumulated data assets more high-quality
  • High timeliness: improve the timeliness of data collection, so as to improve the timeliness of subsequent data applications

[concept] do you really understand the increasingly popular

Finally, to borrow a sentence from Jack Ma’s speech at Hangzhou yunqi conference in 2015, human beings have entered the DT era from the IT era. Data has replaced oil as the core resource. In the future, data will become a public resource like water, electricity and oil.

Data is bound to become one of the most important resources in the future society, and it is also a valuable asset for enterprises in addition to technology, talents and resources.

As an enterprise, we should pay attention to the following points:

  1. In terms of cognition, we should establish the correct data values of the users, have awe of the data, store the data reasonably, ensure the security and privacy of the collected data, and do not attempt to use the data to do evil
  2. In the process, when making decisions, enterprise decision-makers need to add data statistics / analysis / data mining to the company’s decision-making process. Only in this way can data really make data decisions and provide value and influence for the business
  3. In terms of business, combined with the enterprise’s own situation and business situation, a complete set of data value system in line with the company’s long-term development system is sorted out
  4. On the ground, we need to collect, sort out, analyze / mine the data for a long time, extract the key business information, summarize the rules, and do comparative experiments on the data experimental platform, so that different ideas and strategies such as models, strategies, and designs can be continuously updated iteratively, so as to give full play to the real value of data

Original text:https://mp.weixin.qq.com/s/zC_dCJuMSF674iP_WwVLBA

That’s what we’re sharing today

See you

After three things ❤️


If you think this content is quite helpful to you, I’d like to invite you to do me three small favors:

Like, forward, have your “praise and comment”, is my creation power.

Pay attention to the official account of “Java didi” and share original knowledge without any time.

At the same time, we can look forward to the following articles

Recommended Today

Redis design and implementation 4: Dictionary Dict

In redis, the dictionary is the infrastructure. Redis database data, expiration time and hash type all take the dictionary as the underlying structure. Structure of dictionary Hashtable The implementation code of hash table is as follows:dict.h/dictht The dictionary of redis is implemented in the form of hash table. typedef struct dictht { //Hash table array, […]