China Mobile: China Mobile Terminal Big Data

At the "China Mobile 4G+ Terminal Industry Summit", China Mobile, while introducing the latest industrial policies, released "China Mobile Terminal Big Data" for the first time, disclosing a set of statistical information in the form of PPT.

When this set of data was released, almost all the audience raised their phones to take photos of the PPT. What kind of information attracted so much attention from these industry professionals? Let's look at the picture first:

1. The brand with the largest “user base”

2. The brand with the greatest growth potential

3. The most powerful brand model in the region

4. The most loyal brand

5. The most durable brand

6. The most “MAN” brand

7. The brand with the most “user value”

8. The most popular APP

For example, the apps with the largest user base are still browsing and instant messaging, while the rapidly rising video category ranks third.

Among instant messaging apps, WeChat still has not surpassed QQ and can only rank second; while Tencent ranks first in the video category, largely due to the traffic diversion from WeChat.

There are many things worth analyzing and pondering here, but due to space limitations, I will not go into details.

Similar terminal statistics and big data analysis reports have been released many times before . So what is different about this release of terminal big data by China Mobile? Let’s get to the point.

1. Where does China’s mobile terminal big data come from?

【Data acquisition】

When China Mobile released the report, it emphasized that both the terminal information and other data were directly obtained from more than 800 million customers in the network, and all were real situations in the live network. How was this data obtained?

Let’s first talk about how China Mobile obtains terminal brand and model data.

Since 2003, China Mobile has been researching and promoting the collection and analysis of terminal data, requiring equipment suppliers to add the user's IMEI (International Mobile Equipment Identity) information to the call records generated by the switches.

The IMEI field was originally optional, but it has become a mandatory field for network access at China Mobile's request. In other words, all mobile switching equipment used on China Mobile's network after 2003 must provide a correspondence between users and terminals, so that as long as a user makes a call with a new terminal, the operator can detect it.

According to international standards, IMEI contains information such as manufacturer, model, place of origin, serial number, etc., and is the unique identification number of a mobile phone. However, in the early days, the accuracy of this number was not high, especially when copycat phones were prevalent. Some manufacturers did not set IMEI at all, and some models could change IMEI codes at will, so the recognition accuracy was not high. When the system was first launched, more than half of the terminals could not be recognized.

Later, customized phones from operators began to appear, and IMEI became the basis for terminal management. All parties increased their management efforts on the standardization and accuracy of IMEI. Only when the data source was accurate could the accuracy and objectivity of terminal analysis be improved. Today, the accuracy of IMEI recognition across the entire network has exceeded 90%.

The accuracy of customer information has also been gradually improved. Before the full implementation of the real-name system, the operator's customer profile was often the result of analysis and was not completely accurate. Now, the correspondence between numbers and customers is completely standardized, and the accuracy of analysis has been further improved.

Customer behavior and tags are another major type of data. This information mainly comes from user communication usage records recorded by the business support system, user activity status collected by the network side, and Internet access information generated by users during communication.

In order to collect, process and store this information and records, operators have invested heavily in building IT support systems in order to accurately classify and profile customers.

Data processing

To reflect the value of data, the premise is to clean the data and remove the noise data. The data generated by communication networks and equipment is relatively primitive. Before playing with big data, it is like panning for gold. The data must be cleaned, converted, and loaded in a series of processes.

Some data processing is standardization, converting the call records generated by various devices into a standard format; some data processing is filtering, deleting some incomplete information or erroneous records. These processes are to facilitate subsequent storage and analysis.

Some data processing is to establish the corresponding relationship between data, for example: how to determine whether the user has changed the phone - perhaps the user is just temporarily borrowing someone else's phone, or some data information is wrong. Only when the customer uses the same new IMEI for a period of time can it be considered that the user has changed the phone.

For mobile, the data processing stage also includes horizontal and vertical data summarization and organization.

Horizontally, it includes business interactions and data collection between various business lines, and vertically, it mainly refers to data transmission between provincial systems and headquarters systems. If it is data interaction between the two levels, there are also some non-technical factors involved, which sometimes really requires a battle of wits and courage.

2. What are the advantages of China’s mobile terminal big data?

【Authenticity of data】

First of all, from the situation introduced above, we can know that most of the information in mobile terminal big data is real information of more than 800 million customers in 31 provinces. Not only is the data sample size large enough, but it is also collected directly from the Internet.

Secondly, the accumulation of these data is not a one-day effort. Before this disclosure, the processing, analysis and presentation of terminal analysis and other data were all done within China Mobile, and were circulated in a small range in the form of internal reports or analysis reports. Not only are there thousands of cross-check rules to check the authenticity of the data, but after so many years of continuous accumulation and analysis, if there is any change in the data, analysts will work with technical personnel to make judgments and handle it.

Third, for some uncertain data, the internal processing principle is to prefer less data samples to avoid the embarrassing situation of "false data + real analysis". For example, in this analysis of APP application, only the data of some provincial companies were selected, rather than using all the data.

Data continuity

Most analysis reports use static data at a certain point in time because the cost of obtaining data is relatively high and it is difficult to maintain it for a long time. For example, questionnaire surveys require a lot of manpower, and network-side packet capture and collection requires ultra-large-scale storage and preprocessing.

Various user behaviors will continue to occur on China Mobile's network. The collection, processing and routine analysis of data have been fully automated. Therefore, mobile terminal big data can continue to add new information and even carry out targeted statistics and analysis based on new situations and requirements.

On the other hand, China Mobile's resource advantages are fully demonstrated. It has a huge amount of historical data, and the granularity of this data is fine enough, and the computing resources are relatively abundant, making historical comparative analysis possible.

【Relevance of data to customer information】

In terms of conventional analysis, China Mobile's big data already has advantages in terms of data volume, data authenticity, and continuity, and its ability to correlate with customer information is a unique competitive advantage.

What are the differences in behavioral characteristics and consumption when similar customers use different terminals? What are the behavioral characteristics of users and what kind of terminals are suitable for them? Especially the analysis of dual-SIM dual-number customers. In the actual operation within the enterprise, the correlation analysis between terminals and customers is far more than what is presented this time.

Based on the requirements of relevant laws and for the sake of protecting customer privacy, operators will perform data desensitization and other operations as required, whether for public disclosure of data or big data cooperation. However, within the enterprise , personalized data analysis and application of customers is legal and is also a future development trend.

3. The significance of China Mobile’s release of terminals to send big data.

At the 4G+ Terminal Industry Summit, China Mobile released an important industrial policy: all-network terminals will enter the mobile customization library.

From a certain perspective, this policy was forced out by the market. China Mobile's move means that the influence of customized mobile phones by operators on the terminal market is getting weaker and weaker. So how can we maintain influence and even control over our partners? The public release of the terminal big data report at this time is very meaningful.

An interesting thing is that in order not to affect the stock price, what was released this time was not the entire analysis report we provided. Some content that might affect the stock price of our partners was temporarily withdrawn.

These data and analyses exist objectively and can withstand questioning and scrutiny. In today's market where exaggeration is rampant, the real data can completely slap certain manufacturers in the face.

On the other hand, it is the first time that the data analysis originally provided internally has been packaged into a big data report for public disclosure at China Mobile headquarters.

Today, the construction of the centralized economic and financial system at China Mobile's headquarters is nearing completion. This big data platform acquires dozens of TB of new data every day, and its analytical and processing capabilities are top-notch.

If China Mobile adopts a more open mindset and model in the future and provides open capabilities including big data to the outside world, it will be a great driving force! (Shangru Inn-Ning Yu)

The following is the full text of the PPT "China Mobile Terminal Big Data":

Source: Data view from China Mobile PPT and Shangru Inn-Ning Yu manuscript