APP promotion on mobile phone

Wireless applications recklessly occupy a large portion of territory in big data. However, big data analysis on mobile reading and mobile music is limited to specific apps. In reality, apps that promote each other also exist, and we name it the App Mutual-Push Alliance.

APP push each other to explore the hidden relationship between the App, an App is pushed to another related class App, the entire mobile app has hundreds of thousands of applications, the information contained is of course very rich. If these data are effectively utilized, we and our partners can adopt more effective ways to push each other's apps. With these big data, the mobile advertising platform will be more efficient.

Database Modeling is a database modeling function, any huge data requires a correct analysis mechanism

Big data analysis of the problem

① global data get through

When we retrieve individual App data, these data are isolated islands and we are less concerned with what kind of behavior the user has in other App. When we have hundreds of thousands of APP, if we can get through the App data, then we can get the use of endless.

On the PC side, the browser can record the ID of a user by means of cookie, flash, etc. This method is not good on the mobile phone side because users on the mobile end use the App much more frequently than the browser.

However, we found out that each user's mobile phone MAC number is generally unique (in rare cases not unique), so we can get through the MAC number to the user.

② information completion

One of the tricky problems we encountered in the wireless music app of the previous chapter was that the user information was incomplete and we could not effectively push the right song for it. The main reason for incomplete user information is that the amount of information left in the APP is small. If we can make use of the information of users in other APPs to supplement the user information, the user experience in the APP may be greatly improved.

App's push each other

In the absence of access to App data, App mutual push are generally based on random principles, popular App principles, similar to the principle of similar recommendations. There are many problems with these methods, such as different user App recommendations, the recommended App user preferences are low.

When multiple App data get through, the original App push method can be significantly improved. Because on the one hand, we can get the user's global information, we can use this information to better make personalized recommendations for users, on the other hand as the App get through we can better cluster users and App analysis, so more Similar users like to facilitate the likes of similar applications.

App in the app

In the absence of open App data, App ads are generally based on the App's keyword principle, different users may receive the same advertisement when they log in to the same App. One obvious problem with this delivery method is that it only catches the match of the app and does not capture the exact match of the person on the app. For example, when a person just browses an application that plays Maternal and Infant to another music application, the music application App mostly delivers music advertisements rather than maternal and infant advertisements.

The app is designed for ad placement and content early in the design process

When we open the App data, because of the relationship between global information allows users to have a memory function. When an appeal occurs, a music site can serve ads that are closest to its likes based on the user's preferences. In order to more accurately grasp the user's psychology, so that users, companies and advertising to achieve win-win results.