Why can “China Mobile” dominate the 4G era?

How important is China’s 4G market? The number of 4G users in China will exceed 1 billion in 2017, compared with around 2 billion worldwide. How strong is China mobile’s 4G market dominance? While China Unicom and China Telecom were still celebrating over 100 million REGISTERED 4G users at the end of 2016, China Mobile has become the first operator in the world to have more than 500 million registered 4G LTE users. China Mobile accounts for six out of every 10 new 4G subscribers in China and three out of every 10 new 4G subscribers worldwide.

What is the reason for China Mobile’s dominance in the 4G era? The two biggest rumors on the street are the TDD license dominated by early proprietary technology and the smooth 3/4G evolution of TDD and TD-SCDMA.

Myth 1: At the end of 2013, the Ministry of Industry and Information Technology issued TD-LTE licenses to the three major operators “at the same time”, and there is no initial monopoly of TD-LTE licenses by China Mobile. Two years later, the Ministry of Industry and Information Technology only granted China Unicom an FDD license. Just get TDD license, Unicom telecom did not follow up base station construction, always bitter FDD license. Relevant studies of Qualcomm and other companies show that TDD’s distribution frequency band is generally higher than FDD, and its coverage ability is inherently weaker than FDD. TDD is not the sweet property mentioned by some small editors, and the card given to China Mobile by the Ministry of Industry and Information Technology is not “four and two with two Kings”.

Myth 2: About the smooth evolution of 3/4G, some people believe that TD-SCDMA can smoothly evolve to TD-LTE, while Unicom Telecom cannot achieve smooth evolution. In fact, neither Mobile’s TD-SCDMA nor Unicom’s WCDMA nor Telecom’s CDMA can smoothly evolve to any 4G LTE technology. The misunderstanding is caused by the official name TD-LTE, which makes people think it is the LTE version of TD-SCDMA.

Without technological tilt, technological convenience, and even tD-SCDMA giving itself a daze and weakening skills in the first wave of the 3G era, China Mobile was able to achieve rapid growth in 4G subscribers. This kind of as wuxia small master said in “get through the two arteries of the governor” to bring the transformation of heaven, I am afraid that many operation partner psychological play and read a picture below – there is this kind of operation!

Here is a general background to explain: in the early stage of the development of mobile Internet, it was impacted by OTT services of social media software and faced with a pipelining crisis. China Mobile really understands the importance of user experience in terms of user growth and retention. Under the new environment and new situation, China Mobile realized that it had come to a crossroad.

There are many people standing at the intersection, but few of them are really walking on the right road. The fundamental reason why many tech and Internet companies in China end up becoming labour-intensive is that they fail to grasp the essence of “experiment-driven innovation and data-driven decision making”. China Mobile realized and found the right tools and chose excellent A/B testing tools (which ones?). , and widely used in its services and apps, not only “think”, but also “do”.

Let’s take A look at how China Mobile achieved continuous innovation and rapid decision-making through A/B testing tools in the 4G era through three groups of cases.

4G Butler is a mobile APP specially customized for 4G users by China Mobile. You can think of 4G Butler as an upgraded version of the China Mobile Mobile Business Hall App. After the launch of numerous XX mobile apps, 4G manager is still independently launched, and the adoption of new positioning and access to a large number of exclusive services is sufficient to prove that:

China Mobile intends to differentiate its product experience so that both old and new users can be distinguished from its previous mobile products. Therefore, user activity has become the core focus of 4G butler index, each modification should not only bold innovation but also prudent.

At this time, the product department put forward an update plan: change the business document from “phone bill balance” to “phone bill”, and make corresponding modifications on the home page and service page of APP.

According to the previous ideas, process could be roughly like this: file name demand with the product after discussion to determine “as bills”, the whole amount after the update, if the data is ok, if the index is not ideal, need emergency rolled back, so, what all don’t like a product decisions, more like a gamble.

Fortunately, At that time, China Mobile had begun to contact with A/B testing. They adopted A/B testing (called to science and Technology) single blind test to test the two versions at the same time and conduct behavioral control tests on users of different versions (as shown in the figure below). AppAdhoc A/B Testing adopts scientific test flow segmentation to make each group of test objects have consistent user characteristics, and user flow can be adjusted at any time during the test process.

At this point, China Mobile can observe the optimization index in three dimensions: basic index (PV, UV), key index (next-day retention rate), core business index (seven-day retention rate, click volume of balance entry, click rate of balance entry), so as to judge the data performance of this group from basic to core business.

After running the test for nearly three weeks (generally speaking, in order to obtain more reliable data results, the test cycle should be at least 1-2 full natural weeks), the background data has shown very clearly:

• In the experimental version, the index confidence interval of “phone bill” copy on android terminal and iOS terminal was significant, with negative upper and lower limits, resulting in a decrease of nearly 23% in core business data. In addition, the experimental version has a negative impact on some core data of iOS terminal; The acceptance degree of the experimental version is not high, and the overall function recognition degree is lower than the original version of the “phone charge balance” copy.

• The original version undoubtedly proved the rationality of the product through intuitive data. After comprehensively considering users’ usage of balance business and core value, the product team finally decided to give up this revision and maintain the original version unchanged.

Since then, China Mobile in 4G butler on the “anti-fraud SMS” function gray release. “Fraud message” identification of such functions has been very high, why China mobile to use the form of gray release?

On the one hand, the entrance of mobile terminal is too expensive. When a new function is revealed, it means that other functions are displayed at a lower level, which may lead to a decline in the overall operation effect. On the other hand, the added anti-fraud SMS function will cause great pressure on the workload of 4G butler’s manual service, so it is necessary to ensure the overall smooth transition under the condition of no abnormal use.

AppAdhoc A/B Testing platform supports the function of grayscale release. Enterprises can adjust the user flow at any time in the test process, and gradually expand the user group of the new version on the premise of stable version and no bug, so as to complete the smooth transition of product release.

Through the grayscale publishing function of AppAdhoc A/B Testing, the new function was successfully launched in this experiment within about three weeks and positive data changes were obtained. The test results show that:

• The launch of the new function has greatly improved the user activity, and significantly improved the core business data of iOS terminal (an average increase of 68%).

• The anti-fraud SMS function is released without any abnormality, which effectively buffers the pressure of manual service and ensures the smooth and smooth online of the new function.

At present, 4G Butler application has completed its historical mission and has been renamed as 10086 (APP). However, China Mobile’s efforts in mobile terminal are not finished yet, and we will explain the contents related to 10086 in case 3. From these two examples, we can see that in terms of “trial-driven innovation, data-driven decision making”, A/B testing has unmatched advantages over empirical decision making — low cost, high efficiency, controllable risk and high accuracy. Industry experts have said that top PMS can only outperform half of A/B Testing.

It is predicted that wechat will reach 1 billion users in 2017… Yes, the total number of 4G registered users of the three major carriers… Do operators small friends heart shadow area (⊙﹏⊙)…… Almost all of China’s mobile users, if competition in this play wind up, the consequences are unimaginable. Let every wechat user into mobile users, this is China Mobile likes to see.

In public, the main service module and operations in each period, there is a special guide users to download the APP landing pages, the data of 10086 head, found that the current version of the download interface is very long (more than 6 screen), although each screen has direct download entrance, but the data show that in addition to the first screen to download into relatively more, The data on the other screens is far from ideal. After comprehensive evaluation, the product department decided to shorten the download interface of the new version to one screen to see whether the download transformation can be improved. The specific plan is as follows:

Original version: The page content is relatively rich, divided into 6 screens to describe APP functions from different perspectives, and each page has a conversion button; Test version: Integrate the original 6 screens into a single screen, and enhance the visual effect of CTA buttons.

In this test, A/B test and grayscale release were combined, and the initial flow was allocated 4% (1% for each version). After ensuring no bugs or errors in the real environment, the test flow was gradually increased, and the test period was 7 days.

After observing and comparing the test data of different versions, the product department came to a scientific test conclusion: the data of the new version entrance reached statistically significant in terms of both click and conversion, with a sample increase of 55%. From the confidence interval estimate, the diversion transformation after integration into a screen can increase at least 12.4%.

These data well verify the assumptions of the data department and the product department. Based on the existing users of China Mobile 10086 public account, too much function introduction will distract users. Users will naturally download products if they simply express the value that mobile terminal can bring to them. The valuable experience brought by this test also provides a clearer direction for the subsequent product optimization.

For years, netizens have been “flirting” with 10086 SMS self-service for fun, perhaps because China Mobile SMS self-service is really good. In the hearts of many netizens, 10086 is synonymous with China Mobile. The 4G Housekeeper mentioned above was officially renamed as 10086 later, which also proves the brand value contained by 10086.

How to continue the service reputation of 10086 on 10086APP is not only a simple service access and function integration, but also involves a big problem of whether the experience migration is smooth — without the basic disk, how to start with the mass 4G users?

At present, the customer service page of 10086 APP is displayed in the form of “hot questions expansion” by default. The product department concluded that this display mode would lead to the failure of displaying “points” and other types of problems and dialing 10086 buttons in one screen, thus affecting users to find relevant knowledge quickly and resulting in a low usage rate of the customer service interface. May be several problems down, the user is still a face meng forced, helpless to leave.

In order to increase the click volume of each question column on the home page and the utilization rate of online customer service support, reduce the number of times users dial 10086 in the APP, and reduce the pressure of 10086 customer service center, the product department decided to modify the layout and buttons of the customer service interface in the application through A/B test, and decided to create three test versions. The specific plans are as follows:

In this test, A/B test and grayscale release were combined, and the initial flow was allocated 4% (1% for each version). After ensuring no bugs or errors in the real environment, the test flow was gradually increased, and the test period was 14 days. At the same time, click on the question column in each section and click on the button of “contact online customer service” as the core indicators of this test.

In this test and release, the red data results were so important and influential that let’s take a look at what we came up with:

• Whether the “hot question” is expanded or not, or whether the “dial 10086” button exists or not, there is no significant direct impact on users clicking different types of questions. If you want to improve the click of the question column, pay attention to the quality of the specific question and the answer;

• In terms of guiding users to click online support, trial version 3, which did not expand the “hot questions” and removed the “dial 10086” button, had the best effect, with statistically significant results and improved by nearly 25%.

• For the experimental version 1, the data results show that when the “hot questions” bar is folded, the customer service and dial 10086 button at the bottom of the page move up, and users’ interest in hotline and online customer service will increase correspondingly, among which the increase of users who choose to dial phone is more obvious.

• These data prove that users’ utilization rate of buttons at the bottom of the page (including dialing 10086 and using online customer service) is related to the expansion of “hot issues” to a certain extent. The product department can determine the direction for subsequent optimization based on the valuable experience of this experiment.

10086 new user interface: let’s take a look at the online service is recommended home page, and drama, as one of the three TAB label below the top outstanding artificial online customer service, to 10086 hotline, launched a popular answer, select the above title, and add the tags under each category, almost every modification is based on the experiment the conclusion.

A/B

“The more backward the management style, the more empiricism prevails and the more product decisions depend on operational and supply chain management capabilities rather than product innovation. Even for dominant companies like China Mobile, the core competitiveness still comes from product innovation and experience. To improve product experience, we need to learn to drive innovation with experiment and decision-making with data — A/B testing is A data analysis tool and product decision-making tool widely used by world-class technology companies. Scientific use of A/B testing is the only chance to beat competitors, beat giants, beat the industry.”