The characteristics of mobile big data

Mobile devices have gradually become the dominant information age, and the differences between mobile Internet and traditional Internet have become more and more neglected. If we are to bring a whole new paradigm to new technologies and new data analytics solutions, we will need to break our old mindset and re-examine big data on mobile phones.

Although the mobile Internet comes from the Internet, as a new thing, the mobile Internet has its own unique brand.

Differences in user characterization

Mobile Internet "fragmented"

The Internet has tethered people to computer tables, and the mobile Internet has liberated otaku housemaids so that they can return to the real world.

In the station waiting for the bus, holding a cell phone reading the novel; standing in the business district, holding a cell phone search hot spots businesses; or before going to sleep, holding PAD to see what entertainment information, look at the articles like, both pull The long reach of users, but also broaden their use of the network time.

Undoubtedly, relative to mature Internet applications, the mobile Internet application is mainly to fill the commute, work on the way, before going to bed at night and other fragmentation time. As a result, the "kill time" app is also one of three main categories of apps, with a wide audience.

Fragmentation on the phone

Although it has a large audience and a high percentage of time, it does not yet have a good profit model due to its "fragmented" nature. Of course, this is also an inevitable process of market development. Let as many users as possible first use, temporarily do not have to consider how to profit, it is too early.

Mobile phone reading time

Mobile video time

As you can see, cellphone reading and cellphone video peaked at 10pm to 11pm. The typical pre-sleep time reflects the difference between a mobile application time and an internet application.

The complex situation brought by mobile internet

Mobile Internet is more than a simple extension of wired business to wireless. Is bound to produce some growth in the business geometry, give full play to the characteristics of mobile Internet, such as LBS, such as O2O.

The development of the mobile Internet makes some things that may seem unrelated to string together, I feel a faint faint. Yuan Fang, how do you see?

Therefore, the data processing of the mobile Internet is by no means a simple statistical analysis, but a variety of more advanced algorithms to find the truth hidden in layers of fog.

Differences between internet analysis and mobile internet analysis

Web Analytics has been widely used by a variety of Internet companies, e-commerce businesses, and corporate websites in traditional industries. Mobile Analytics is still a fresh thing. It has an inherited relationship with Web Analytics, but with a clear distinction.

(Usually referred to as Mobile Analytics mainly refers to the Mobile Application Analytics, which is the statistical analysis of the native applications on a variety of mobile devices.)

In Web Analytics, the majority of users are based on browser cookie statistics.

In other words, two people who actually use the same browser on the same computer to surf the Web will be counted as a unique visitor, while the same person who uses both IE and Chrome on the same computer will be counted as Two independent users. Mobile Analytics has different audiences. It counts on mobile devices like cell phones, and for the most part users per mobile device are unique. So it is more accurate than Web Analytics to humans, which means it can be based on more personalized service and more accurate marketing.

Data analysis on the phone

This is not the only fascination with Mobile Analytics. Browser cookies are easy to clear or overwrite, but device-based statistics are relatively more stable and long lasting. This makes Segmentation and Cohort Analytics more accurate and useful. We can filter out a group of users by some criteria (for example, adding new users in March of that year and using more than 20 minutes) to analyze their patterns of behavior.

The big difference between Mobile Analytics and Web Analytics is that the former statistics may be off-line or poorly-received, and the network is unstable, which causes the statistics to not be uploaded immediately. Wait until the data can be uploaded, it may have been separated by a few hours to several days. This is not the case in Web Analytics. You can not access the Web without access to the Internet. If you can access the Web but can not connect to the statistics server, the statistics will not be resent. This makes Mobile Analytics need a more complex data compensation strategy.

In terms of data volume, Mobile Analytics is no less than Web Analytics at all. It counts a lot of data that Web Analytics does not have, such as device model, application version, promotion channel, or even location, along with a lot of developer-defined events. The growth rate of total mobile devices (including tablet PCs) is far greater than the total PC (including laptops) growth rate, the number of mobile applications on each mobile device growth is also very fast. So the Mobile Analytics platform that delivers public services is a typical big data scenario.