Since the data in Taiwan fire, every day someone asked how to measure the data in Taiwan construction effect, how to prove that data in Taiwan construction is successful.
This is a very complicated problem, just like asking how to prove the success of enterprise digital transformation. Making enterprise data “use, run and turn” is the standard for the success of enterprise data construction.
The following case is a typical enterprise to build a data center using data, mining data value process, using this case, we look at how to build a data center from scratch.
Data application challenges
This enterprise is a traditional industry, and its business model is a typical four-layer structure of brand, distributor, store and consumer. It is a typical product plus service model, and it is highly dependent on stores (clients).
In the past, the enterprise has a strong competitiveness by product differentiation, is a typical seller’s market, the way of operation is quite traditional.
The market is big money, campaign advertising is rough management.
Sales of atmospheric pressure goods, dealers are controlled management.
However, in recent years, the industrial upgrading of the enterprise, customer demand is more and more diversified, stores are less and less dependent on a single product, dealers are under more and more pressure, the traditional model of pressure has been unable to cope with the existing model.
In order to cope with such challenges, the company has also made a lot of attempts to establish its own e-commerce platform and offline direct-sale stores, hoping to establish more direct contact with customers and consumers. It has built a lot of applications, but it still faces great challenges.
It mainly includes the following five points:
- Market cost spent a lot, what effect did not know to bring
- There is no client end data, there is only Sale In without Sale Out, and we don’t know who is using our products
- Online traffic cannot lead to offline orders
- Offline dealer data cannot be returned online to form a closed loop
- Companies pay for traffic, but conversion rates are low
How to solve these problems?
An iceberg model for meeting challenges
The reasons for the above business phenomena can be found from the perspective of data:
1, the market cost spent a lot, do not know what effect has been brought:
Investment in marketing, such as advertising and campaigns, is not directly correlated with reading, transmission, conversion and sales data, so it is impossible to measure and evaluate the effect.
2, only know the Sale In, not Sale Out, do not know who is using their products
In the traditional dealer agency model, the brand can’t get Sale Out data, so it can’t know the real inventory situation of the market accurately, so it can’t make sales forecast more accurately, and it’s not clear which stores and consumers their products are sold to, which is the problem of missing data.
3, online flow can not lead to offline orders
Due to the particularity of this industry, it needs the support of service, so it is less likely to place an order directly online and largely depends on offline communication. However, there are users visiting online, but which ones are high potential users, how to follow up, and which stores or dealers are more appropriate to be allocated to? These decisions lack data support, resulting in a low success rate of online traffic to generate orders offline.
4. Offline dealer data cannot be returned online to form a closed loop
Offline dealers’ order data are lack of effective means to collect them back to the brand owners, so as to match the corresponding Sale in data, so as to form a closed loop, which will lead to the lack of understanding of market forecast, inventory, dealers’ behavior and sales ability, so as to make more accurate guidance for the subsequent operation.
The market competition is more and more fierce, brands or earn money, in accordance with the smoke into the way and profit requirements for the Chinese market still maintain a certain growth, this case brands all kinds of pressure goods sales department in order to complete the task, the dealer not to earn money, the toes, and distributor of digital level is generally low, staff turnover rate is large, into a vicious circle.
Enterprises spend money to buy traffic, but the conversion rate is low
Some enterprises realize the importance of traffic, so spend a lot of investment in traffic, the quality of traffic is getting worse and worse, the conversion rate is very low, in the final analysis, the data of these traffic and internal marketing data are not integrated, resulting in quality traffic has not been identified
Three stages of successful construction of a data center
The whole process can be summed up in three stages: Getting the data, “working,” “Running,” and “turning.”
1. Put the data to work
The enterprise did not make a technology platform at the beginning, but made a light consultation. First of all, from the perspective of business value, it released a list of all kinds of valuable data utilization scenarios, and then sorted the value priorities of these scenarios, and selected the scenarios with higher priorities for in-depth analysis.
Explore the data required by these scenarios, verify the technical feasibility, make a feasibility priority ranking, and finally select the most valuable scenario with the best data base and technical feasibility from the scenario list. Start delivery development immediately as the least feasible product (MVP), so as to use the data as soon as possible.
Three intersections of lean data innovation
Putting data to use and generating business value is the first priority of data center construction. Whether the process is automated, supported by a big data platform, and advanced technology are secondary matters.
For example, as a dealer of this enterprise, I often need product parameter data. In the past, these data were sent by email, so it is not timely, and need to be manually processed after receiving them before they can be used.
This scenario was identified in the light consultation, and later, a data API was made to allow dealers to easily call the latest parameter data without having to call and send emails. This parameter data was used.
But, in fact, the demand of the technical implementation is very simple, not complicated, but in the early, even some backward, in order to quickly respond to business needs, how advanced the development of the background and not used, also is not automatic, but good artificial query, acquisition, processing parameter data into a folder below, and then use the program to read the file, I didn’t even use the database.
Enabling data to be called by business as a service, even if it’s manually configured, is the first stage of data centralization, making data usable.
2. Let the data run
The data center is the data product factory of an enterprise. Its function is to build an automatic data processing chain between source data and data products, so that source data can be automatically collected, processed, transformed and integrated into data products (services), which will be invoked by relevant business systems.
The whole process flows through the data value chain in the data center. The data center automates the process from production to consumption of all data products, forming automatic flow and thus making data run
Let the data run involves several layers of specific meaning: let the data run all the time to get the data to the right place
Get the data rolling
summary
The corresponding data application system, data technology system and data operation system are to make data “used, run and turn”. All three systems must be in place to build data-driven capabilities.