I take the game of the App Store as an example, the content may not be entirely operational, but the main attempt to analogy.
There should be two level one is the more traditional transactional data: downloads, monthly receipts, the ratio of views to downloads (representing the attractiveness of your app), the ratio of downloads to paid users (representing the ability to absorb gold), the amount of ad clicks (if any), and so on.
The other level is the user behaviour level:
Through what channels does the user enter your app on the App Store page? -is the navigation through the Mobile Web page? or keyword search? or the App Store recommendation?
If it is a keyword search, what are the keywords? (Search optimized)
User retention rate? (Software quality)
Average user usage cycle? (Promotion of content)
Analysis of user activities in the game? (Changes to the content of the game)
Peak period of user activity? (the corresponding time period launches the activity)
Wait, wait.
Could you give it a little bit more power?
If these are simply to ask for the average of all users, then the effect is still limited. If you want to further improve the efficiency of operations, you need to classify users to determine the user's ability to pay and retention cycle
Establish a user scoring mechanism:
For example, the user completed the novice task in X minutes + 1 points
User pays 1 cents on X day
User builds Guild + 1 points in game
User >x + 1 minutes online in x days
Users use software for +1 minutes during x hours
And so on and so on and so on, such a set of mechanisms built completely, in the operator has a lot of room to operate. Through different fractions, give users different classifications, so that can be expected, in this expectation, take different ways to maximize the rate of payment and user retention rate.
Could you give it a little bit more power?
You can magnify operations by sharing data with other companies and working together in a way that gives an example here.
A game (really forget ...) With a clothing store (looks like Zara?) With this collaboration, after getting the data from the user's location (the location service opens), if you find that the clothing store is around the game user, you can enjoy a 20% discount in the clothing store by showing a page of the game. It is alleged that the activity redemption rate as high as 60%.
The same idea can be extended to other data, but this data involves privacy, cautious ...
Could you give it a little bit more power?
In this internet age, how can we ignore the huge role of "social network"? By digging through social networking data, you can find out what the user is discussing about your app on social networking sites and finding different points of attention, and you can correct or open the champagne with criticism or praise.
Wen/Sun Weizhan