This piece of analysis is often the product manager describing the strategic positioning and development goals of the product in the BRD or MRD, and then tracking the goal through data (as shown in the figure below).
Of course, many company's basic product managers do not know what these documents look like (or simply write them casually). Therefore, the current strategic goals can only be judged by some simpler and direct methods, such as the most commonly used product life cycle method (as shown in the figure below).
The analysis at the strategic level should be the first to start, and a clear analysis is required first. Because each subsequent specific function, each specific design, and each iterative optimization serve to achieve the strategic purpose. But this is precisely the missing part of many product analysis.
Because in many companies, the following factors will lead to the failure of strategic analysis, resulting in a lack of sense of purpose and subsequent problems. l The boss himself is the real product manager, others only dare to follow suit.
According to different tasks, it can be divided into core functions and auxiliary functions. For example, the core function of transactional products is very focused: closing transactions. It's just that in order to improve transaction efficiency, functions such as text grass, video grass, active check-in, tree planting and watering will be added, so pay attention to the usage rate, bounce rate, conversion rate of each function, and GMV that facilitates transactions. .
The analysis of content-based products is slightly more complicated, because the time for users to complete a traffic behavior may be very short. After logging in, a series of actions will be performed. And content-based products often provide a variety of topics (current affairs, technology, sports, two-dimensional...), a variety of products (video, graphic, voice...), and various forms (PGC, UGC), so user behavior is more messy .
At this time, users are often stratified first to distinguish light, medium and heavy users, and then look at the user's overall active market and the amount of content, and then look at which functions are dispersed in specific behaviors, and the usage of each function.
At the heart of the tactical layer is addressing priorities.
As for what it will be like, there is no specific expectation at all (or just write one at random). Subsequent analysis is even more confusing.
Battle layer: the design of a specific page/button/process/play .
The most well-known analysis of the combat layer is ABtest. Because you can go directly to ABtest, you can end the quarrel at the design level without thinking, and use whichever data performs well. ABtest is so popular that many factories, even data analysts, are too lazy to use traditional analysis methods, "In trouble, ABtest!"
ABtest is really the quickest way to resolve disputes, especially when it comes to design issues like the left/right side of a button on a page. When product managers are arguing, going directly to ABtest can make everyone shut up quickly.
But ABtest is not a panacea. Because country email list user behavior must be the result of the combined influence of multiple factors, the longer the user behavior path is, the more influencing factors will accumulate to the end, especially the process involving cash transactions.
Therefore, if the user behavior is really broken at the end, it is difficult to break it down in a short time. The actual environment of the enterprise is not as clean as that of the laboratory. Factors that cannot collect data, such as the cross-influence of user word-of-mouth, current events that are popular in the current season, public opinion trends, and marketing preferences, will affect the results. Therefore, ABtest is more suitable for solving the design problems in the early stage of the process. The further back the test results are, the easier it is to fail (as shown in the figure below).