Open statistics: the key to enabling fine-grained App operation and fine-grained app operation

Open statistics: the key to enabling fine-grained App operation and fine-grained app operation
Data is the lifeline and the most important intangible asset for companies, small and medium-sized entrepreneurial teams, and individual developers. Currently, App development is increasingly dependent on data support. Most developers use data as the basis for product decision-making, operation decision-making, and promotion decision-making to carry out refined operation.
When it comes to refined App operations, I always think it is very unpredictable and there is a sense of height! In fact, how can refined operations Lack mobile statistics assistance? In the mobile Internet Arena, how many brushes do they dare to mix? Companies that provide mobile statistics services at home and abroad have accumulated a certain number of partners and typical success stories over the past two or three years, so new entrants do not have the opportunity to enter the market? No, as long as the services you provide are considerate and considerate, and the capabilities are excellent, as long ......, There is always a market for you!
Voice cloud open statistics, as an emerging mobile statistical analysis platform, uses the powerful cluster computing capability and Big Data Technology of xunfei voice cloud to build a statistical analysis SDK based on the data indicators that developers are most concerned about, with simple integration, your applications can easily learn the development trend anytime and anywhere! The basic analysis functions, such as application trends, channel analysis, terminal Property Analysis, behavior analysis, custom analysis, and error analysis, are all available.
In addition, the speech cloud open statistics, combined with the unique voice semantic data of xunfei, can provide Big Data Analysis Based on voice semantics, helping applications quickly run in the mobile Internet wave, this is a special feature not available on other platforms! At the same time, if the application does not have the ability to integrate speech semantics, open speech cloud statistics will also be open to you!
What are the unique advantages of open voice statistics over other platforms? Without comparison, how can we know?
Advantage 1: although the sparrow is small and dirty! The minimum size in the industry (the Android SDK is only 49 K), but its functions are comprehensive ~~~
Advantage 2: multi-dimensional comparison, how to compare data ~~~ Time Comparison: You have a large family of users. For more information, see application, channel, and version comparison!
Advantage 3: key data is refreshed in real time in 3 seconds, which is so fast! Three key indicators, including new users, active users, and number of startups, are refreshed in real time within 3 seconds. There is no latency in operation response, and the exclusive one is so cool!
Advantage 4: The report content and various forms are sufficient to meet various requirements. There are many daily, weekly, and monthly reports. The report content and recipients can be customized. You can decide who you want to view and what you want to view!
Advantage 5: only four steps are required for integrated development! Import the SDK, configure the AndroidManifest file, add code, and edit and run the file. The entire process takes only 5 minutes!
That's all? No. How can I beat it if it's just a little skill! There are more considerate service functions ~~~ 1. Indicator warning: Keep An Eye On Data fluctuations. Is it possible? You can customize important indicator fluctuations to keep you updated on indicator fluctuations!
2. Milestone management: What should I do if my application has a major change or a serious Bug that has been around for a long time without logging? Milestone management allows you to easily record inflection points in a trend chart and recall historical events at any time, making reading a chart easier.
3. Custom by collaborators: What should I do if I want other related personnel to view the data? Collaborators can be easily managed. developers can add collaborators, browse permissions, and set permissions on the web ....... All configurations can be customized as needed, and permissions can be revoked at any time ......, All in all, DIY!
If you talk more about it, you might as well experience it yourself (experience the Demo) in your applications. You may feel less exciting and want to make your applications stand out from each other, so that you can control the development trend of your applications at any time, what else does it provide data support for refined operations? Voice cloud open statistics joins hands with you to complete what you once wanted to do without refined operations. It not only has powerful technical support, but also has considerate service support!

How can we analyze user behavior through application statistics to achieve refined operation in apps?

I have a wide range of questions and can only give a rough answer: (1) the goal of refined operation. For example, if your product is just a tool, I'm afraid there will not be too many refined operations, generally, regular user behavior analysis and user qualitative research are well performed to guide product design. If it is a content-based product or a product with both functions and content, you must consider it. 2. design the statistical framework. If you frequently interact with and use functions on your app and browse or generate content, you must design the statistical framework, design your statistical framework. Ii. Brief procedure 1. data collection first lists the data items you need, and then evaluates which data items need to be reported by the APP and which data items can be counted in the background. Generally, the APP must undergo careful verification and testing before reporting the collected data, because once the version is released, the data collection is faulty, not only has the previous work been done in vain, but it also brings about a lot of dirty data and may reduce the running efficiency of the client, which outweighs the loss. 2. after data collection, you need to process various raw data into visualized data that product managers need. Here, you need to perform some basic data logical association and presentation, so we will not go into details. 3. data analysis is based on the statistical framework designed at the beginning. You can clearly see the data you need. Of course, the above is just A basic analysis that cannot be further analyzed. For example, if you get the data, you can analyze the user who uses the function and also like the B function. The two are highly correlated, whether integration or interface adjustment can be considered during front-end design. For example, analysis of click streams, what is the path for most users to access or use the APP, is the core function hidden too deep? For example, we can analyze different user attributes, such as male users and female users. Are there obvious differences in user behavior? And so on. The data analysis methods and models of different products are very different, so we cannot make it clear at once. So the above are more examples. 3. Some principles that need attention 1. the data itself is objective, but the data to be interpreted must be subjective. The same data may be analyzed by different people to draw the opposite conclusion, therefore, you must not analyze the data with your opinion in advance (for example, if you already have assumptions, then use the data for demonstration); 2. data collection by APPs must have a low priority. Data collection cannot affect product performance and user experience, not to collect users' private data (although many domestic apps do not); 3. data is not omnipotent. You still need to trust your own judgment.

How to Implement refined APP operations? Starting from setting the target and proposing assumptions

Especially in the mobile Internet era, the window of any product will be shorter and shorter. In the recent off-line activities of umeng, Jia Enbo, the product manager of umeng, analyzed how to use data for refined APP operation. There are a lot of dry goods, and we recommend them to the Operation staff. The following is a transcript of Jia nbo's speech: Let's start with the data-driven APP operation. But from the perspective of APP operation, as we have said just now, this is to say that the operation is not an independent thing. It must be infiltrated into the full process of product creation. Even we thought that if we were going to make a product, we had an idea at the beginning. Then, through this idea, we may need to find people and then design them, then the design goes to development. After the development is complete, it may go online, and then we will do the promotion. After the promotion is complete, we may think that when the number of users reaches a certain level, my data reaches a certain level, and I can rely on it to make profits, and then I will develop like this. But in fact, we do this, but when we think about it, we must think about it. That is, we are sure that the first foothold is this thing. How can I do it, in the future, I will be able to make money to support my team, support myself, and think about how I need a large number of users, how can I do promotion, and a large number of users, then let's move on to the idea that if there are a large number of users who must recognize me and recognize me, how can I meet his needs with my products. Then, let's push forward, that is, how can I implement this product and provide the core value to the user to meet this requirement. In fact, if you think about it, the entire operation is actually the means of the entire operation that can be reflected at the end of your mind, in fact, in the design phase of the APP, you should think about it, and then let it go online, you may have planned it. I want to see what kind of indicators it has, it indicates that it meets my expectations and what kind of indicators I can achieve after I launch it. I think it can be promoted, these things must be planned in advance. Therefore, operating it is not a specific task. It must be something we should consider throughout the entire product process. Starting from setting goals and making assumptions, I will talk about data operation today. We will take several steps, because there should be many guests behind me, they should share some operation cases, so I will not talk about the case here, mainly about how to use the data operation to get a general thinking model. There are several steps. The first step is to establish the target and propose assumptions. Then we need to select some key indicators, collect data comparison and analysis, then make assumptions in the curve, and finally iterate and improve, I will follow these steps. First, we need to implement operations. The first and most important thing is to establish goals. In fact, I propose my tasks in stages. I need to know that I am at different stages, after this stage, what is the purpose of my operation? Then, based on my goal, I will propose the assumption of different operators, that is, how can I collect data, this should be a prerequisite. For example, at the early stage of our launch, our operation goal may be to say that our product was just launched, and I don't know whether this product is actually acceptable to users, are there some serious design defects in the product. Therefore, in the first stage, I must obtain early users after the product is launched, and then obtain the target of early users. I just want to verify the logic of my product because my product is just launched, without a successful experience, I must think that I want the user to do this, but in fact, the user does not do this, so I must say, put forward a hypothesis that is to say, I want to verify it. Using data is to check whether the APP is used in this way, whether it really refers to my data, the information I want to pass to it, and what it understands, I really want him to use this function. Is it acceptable to him? Is it true that he can use this function as I want? My APP is available every day and he can see it every day, as a weather application camp, I use it every day to check the weather. So in the early stage of the release, it must be assumed that the entire logic of our product should be verified. Then, after a period of time, it will produce... the remaining full text>