As we all know, Google, which is A search-based company, must rank its Market in the same way (a x a % + B x B % + C x c %) the value calculated by the formula for ranking. Developers can provide guidance on product design, R & D, and promotion based on their ranking rules.
1. What are the indicators in the ranking formula?
What are indicators A, B, and C? What are the weights of a, B, and c? The answer to these questions should be desired by every App developer and operator. After knowing this answer, some "smart" people will, like SP practitioners in the past, brush those indicators with high weights to promote product rankings. Many discussions on this issue on the Internet are endless. Android Market App Rankings, a relatively in-depth article, is just conjecture that it is based on the growth and retention of "installation volume. Unfortunately, this author is one step away from the correct answer ......
Maybe you are waiting for it, haha. Now let's talk about my conclusion. Indicator A = "Total installation/total download", that is, the download-to-install conversion rate; B = "rating/5", that is, the full score of the product to the Market; C = "retain installation/total installation", that is, the retention rate of installation. Don't rush to ask me how much a, B, and c are equal to. The exact values are only known by Google and can be tuned. I can only tell you that they add up to 100, and B> max (, c) (why? Please be patient ).
According to this formula, if every downloaded user has been installed, and has been given 5 points, and has not been uninstalled-OK. no matter what a, B, and c are, your product has a full score, which is No. 1!
2. Why are they?
Before telling you why it is A, B, and C, let me tell you why it is not D, E, F ......
First read data, Android Market-communication-free-TOP25:
With the product name removed, we can see more pure data.
First, we can confirm that the ranking is sorted by "downloads". The number of products with more than 10 thousand downloads (9th bits) is 5 million + (11th bits). Imagine a 11th-digit product shouting: "Why? Which of the following is better than me ?" -- Answer: see "rating ". Ah? It is 5 points! That's right. The 9th-digit product scored a full score. Oh, while excluding D, we have gained B.
Second, we can exclude the "number of people who have scored ". The reason is that when the "download times" and "scoring" are both of the same magnitude, the comparison of a number of samples is not systematic-it is really irrelevant to the "Number of people scoring. E is killed.
Again, we want to exclude "acceleration ". Many people speculate that there will be an "acceleration" indicator, or newly downloaded or installed. This speculation stems from "encouraging new products", but it is a radical idea. Market should encourage new products, but "Protect" new products and "Take good care of" old products that are indeed powerful. A large user base will surely slow down the "acceleration" of the old ox product. If so, the top 10 of every day will be totally out of sight (there will be "acceleration" in the App Store ranking rules ", so the TOP ranking is easy to accommodate every day ). Therefore, acceleration is not really fair.
Excluding D, E, F, and B. Where does A and C come from?
Take a break and ask a question? How do we judge the "good" and "bad" of an item? Is there any of the following components? There is something in front of you, and it starts your desire to try. After the trial, you may not be able to stop using it. Lu yaozhi Ma, it took a while again, and you have a new judgment. If this is really good, then you will be inseparable from it ...... Returning to the essence of our judgment on the standard of an item is nothing more than "accepting it", "affirming it", and "keeping it ".
To put it bluntly, whether the user "accepts" an App-the download is "head up", and the installation is "nod ". If the single "installation" is used as the indicator, it will make old products snowball, unfair. Therefore, A (download-to-install conversion rate) emerges. "Install" as the denominator and "Download" as the numerator. On the one hand, the higher the downloads of the top products and the higher the installation volume. if the product is not good enough, the higher the downloads may become the denominator, the 11th-bit product in the example above is an example (the installation volume cannot keep growing at the same time under a large download volume, so it is overwhelmed and rankings fall-spoiler, the fifth place is KakaoTalk, a product that once ranked 11th in the "communication" field ). On the other hand, as long as A (and B, C) is good enough, some new products can come up with A header.
B (score) has been analyzed just now. Here we need to add that the score is not one-time. We will find that the "rating" in the Market can be modified multiple times. This aims to spur the product to be "good" continuously ", users have always mastered the product's "killing and killing power ". In addition, the reason why B is greater than a and c is that there are only two scores below 4 in the top; among the top 25, there are zero scores below 4; and among the top 10, there is only one score lower than 4.5 (to avoid the phenomenon that the Miss diary in the App Store occupies a high rank For A Long Time)-Simply put: to sit in the front row, the product must first make users feel "good ".
This article describes the source of C (installation retention rate. To put it bluntly, it is to control the traffic loss rate. If the "Total installation" is very high, but there are very few "retained installation", it cannot be called a "good" product. Therefore, the installation retention rate is another indicator that comprehensively measures the advantages and disadvantages of the product.
3. What are they?
I have already mentioned that B is bigger than a and c. What about a and c? This depends on the policy of the Android Market. If "a is bigger than c", it means "adding more importance" (this is the most important option for any store when it is opened first) if "c is bigger than a", it means "retaining is more important" (for an old store that has already thoroughly penetrated the customer flow, it will be more important). Therefore, this can be adjusted, it depends on the development stage of Android Market and the self-recognition of Operation managers.
4. Comments
If the Android Market is really like this, it can be described as both old and new, both attack and defense.
It combines old and new because it enables new and good new products to come out quickly and withstand the scrutiny of time. This is because it will make the dishes bigger and bigger, and it also prevents "fake" (it can be flushed up, but if the product is not good enough after it is flushed up, will make the product worse), make the plate more and more powerful. This fair rule also provides another powerful support for the Android Market to increase the application speed beyond the App Store.
The invisible referee in the Android Market ranking rule tells us:It is the right way to continuously improve the product!