Shulou(Shulou.com)06/03 Report--

The central station is to "let people who can hear the gunfire summon gunfire." in the face of the raging tide of the construction of China Taiwan, only after explaining the question of "how to use the data", it is necessary to further answer the questions of "how to get the data" and "how to store the data."

Why do you need a data Intelligent Model

Zhongtai, one of the hottest words in 2019.

If the data center is compared to an aircraft carrier of modern enterprise data management, there is no doubt that this aircraft carrier is still a lone giant, with no escort fleet, no combat cluster, let alone nuclear submarines.

The problem now is that we have built a "central power station" only to find that there is a lack of "light bulbs" that can release its huge capacity.

Do we have a computing infrastructure with tens of thousands of servers in a single cluster, just to make reports run faster or eliminate data islands? This is undoubtedly the biggest waste of resources for massive computing power, and the market calls for data applications that can release these huge computing power.

What kind of system can match the huge computing power? What kind of system can really and effectively solve practical business problems? Data intelligent model!

What is the data Intelligent Model

What is a data intelligent model? What's the difference between him and the traditional information system? Where is intelligence embodied?

Here we need to explain the difference between "information system" and "intelligent system". The essence of "information system" is to edit database. If the core of a system is to rely on manual decision-making and rely on a large number of human interactions to complete tasks, then it's an information system. On the other hand, the "intelligent system" relies on the machine to complete a series of operations such as "data cleaning-problem location-business decision" with the task as the input and the processing result as the output.

According to this standard, the shape × × intelligent systems on the market only mix beads with fish eyes in the name of intelligence.

The intelligent degree of the intelligent system can refer to the following figure. L0 to L4 means the higher the intelligent degree.

In the four stages of the development of enterprise management, the enterprise must complete the transformation and upgrading of the whole information project from L0 to L4.

Intelligent data Model-- Intelligent dispatching and replenishment system

The information construction of enterprises is actually the projection of the upgrading of enterprise management. Of course, this is still a bit abstract.

Take the most basic commodity replenishment link in commodity operation as an example.

1. L0 stage: in the early stage of the enterprise, the replenishment of regional stores and the transfer of goods between stores is not a very serious problem. A person can list a few forms and take some thought to solve it, and only standardized management can be achieved.

2. L1-2 stage: with the expansion of the scale, when the number of stores reaches hundreds, it is necessary to set up a commodity department to coordinate the commodity shortage and regional imbalance among hundreds of stores, which requires process management. Semi-automatic or automatic management system can assist commodity operators to form unique operation style and strategy of the enterprise.

3. L3-4 stage: the transmission efficiency of data decreases marginally with the increase of nodes (personnel). With the further expansion of the scale, we want to rely on adding people to manage the whole batch of goods offline. It has become a problem of bloated staff, low efficiency and difficult to measure effectiveness, accompanied by high staff training costs and the risk of core staff leaving. Intelligent system also arises at the historic moment.

A group we have recently served is in the transition from process management to automatic management, with more than 20 people in the commodity operation department, who need to spend an average of four days a week on replenishment and transfer data.

These more than 20 people are well trained and need human flesh to judge the profit and loss status of goods in 500 stores. Which type of stores should be given priority when supply exceeds demand, and how much should they be satisfied? When supply is less than demand? A skilled delivery specialist needs to consider more than a dozen measures at the same time.

Now a group is ready to open the × × model in the coming year, the number of stores is expected to expand to 4000, and the average training period of a skilled commodity specialist is at least 2 years.

When the store expands tenfold, so does the merchandise specialist?

Intelligent dispatching and replenishment system, which collects weather, region, location and other external data, combined with the industry's advanced experience in dispatching and replenishing goods, is expected to achieve:

1. Using the deep learning algorithm, the replenishment data originally required the cooperation of more than a dozen people, and the machine completed all the replenishment process in just a few minutes, even if the replenishment calculation of more than a thousand stores was fine, and there was no need for hard recruitment and training. and always be on guard against competitors to poach.

2. For the replenishment problem which was pulled repeatedly by the whole commodity department, the system automatically completes all the scheduling work according to the gross profit optimal solution, and is expected to increase the average weekly sales rate to 60% to 70%. The biweekly sales rate has steadily increased to more than 80%, and the number of inter-regional transfers has been reduced by more than 30% (only the replenishment logistics cost of a certain group in 18 years can reach 2.5 million).

3. The brainpower of the business staff is greatly released, and the precious core staff only need to manage the optimization direction of the algorithm model and data supplement, and have more time to think about the operation strategy of the commodity itself. the commodity department has also been upgraded from a cost department to a profit department.

A giant ship still needs a sharp gun.

The wave of cloud computing has reduced the cost of hardware procurement and directly spawned the ecological prosperity of China and Taiwan today.

Today, small and medium-sized enterprises in China can build their own data center at a low cost, and have the opportunity to correct the direction of data construction from the source, but the popularity of China Central Taiwan and the needs of business development will inevitably force the upgrading of the business foreground.

Under the China-Taiwan strategy, all aspects of the current enterprise are actually worth restructuring and upgrading.

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.