Current location - Health Preservation Learning Network - Fitness coach - App data analysis, what do you want to analyze?
App data analysis, what do you want to analyze?
According to popular methods, product life cycle (PLC) is divided into initial stage, growth stage, maturity stage and decline stage. At each stage of the product, the work weight and analysis emphasis of data analysis are different. Let's use the case in stages.

First, the initial stage

The focus of the initial stage is to verify the core value of the product, or to verify the hypothesis of the product: a problem can be solved for a specific group of people through a certain product or service. At this stage, according to the idea of MVP (Minimum Variable Product), the idea of starting a business with minimum cost should be verified, and the solution should be quickly iteratively adjusted according to the feedback from users, and finally verified on the data.

Case:

Take the social application of a mobile phone forum abroad as an example. During the idea period (around 12, 13), it was found that users of the forum often said that accessing the forum from the Wap page of the mobile phone was slow and there was no mobile adaptation at all. So we put forward the idea of making an app to connect the forum system with users, so that forum users can enjoy a smooth forum access experience on the mobile side, and

So in the early days, the whole product was completely excavated around the two core scenes of reading posts and posting posts, and was publicized in the forum. The price is $65,438 +08, and it is found that many users pay, and the retention rate of these users reaches 60%+ (of course, it is related to the user's payment), and half of the users use it for more than 70 minutes. At that time, some competing products came out one after another (Vbulletin team, the largest forum system at that time, developed a mobile App to solve the same problem), but it didn't take long to lag far behind us, just because the whole team followed the idea of MVP and concentrated on polishing the smooth experience of reading and posting repeatedly according to user feedback, which won a very good user reputation, led the market and won the investment of a famous Silicon Valley investment institution.

Key Data-Portrait of Target Group

In addition, in the initial stage, you can access some third-party application monitoring SDK to understand the portrait of the initial user group, and verify from the side whether the characteristics of the user group are consistent with the assumed target user group, as well as the common demographic attributes (gender, age, education, region).

Case:

At the beginning of April this year, I chatted with the product manager of a fitness APP in China. At first, this application was a tool application for tracking fitness and exercise. In the early stage of the product, the retention of new users is at the industry average level. When observing the portraits of target users, I found that there are obviously more female users than male users, and the retention of female users is also obviously higher than that of male users. Therefore, it is decided to tilt towards female users in product strategy, focusing on women's fitness, fat-reducing beauty and other functions and content recommendations. The overall next-day retention rate of products increased by nearly 100% compared with before.

Similarly, I recently served an internal customer of a goose factory, who developed a new product for young people, only to find that the age distribution of its users was mostly teenagers and the elderly:

This is only related to their user channels. It turns out that they have a product for teenagers and the elderly. In order to bring the first users to the product, they directly drained users from the old product, and found that they were not the target users of the product.

Critical data retention rate

When the current users meet the characteristics of the target audience, the core pays attention to the retention rate, duration/frequency of use, user stickiness and other indicators of these users. The retention rate has been expanded here.

Retention rate has many dimensions (7 days, 2 weeks, 30 days, etc.). ), according to the product characteristics to choose. If the product itself meets the low-frequency needs of the minority, the retention rate should be two weeks or even 30 days. High retention rate means that users recognize and rely on product value. Generally speaking, the hypothesis can be verified, and the retention rate below 20% is usually a dangerous signal.

This paper introduces a data-driven leading indicator model, which can guide product design by finding leading indicators, thus improving the retention rate. Let's first look at the definition of leading indicators. Precedent index refers to a product behavior of new users in the early stage of using products. This index has a very high linear correlation with the user retention index, which can predict whether users will stay in the product.

Describe it with your own summary formula, which is roughly as follows:

Possibility of positive prediction (%): indicates that the user performs this behavior, and then the possibility that the user remains active can be predicted.

Negative Prediction Possibility (%): indicates that if the user does not perform this behavior, it can predict the possibility that the user will no longer remain active.

Finally, the credibility of leading indicators = the possibility of positive prediction x the possibility of negative prediction. Let's look at the case directly.

situation

Assuming the previous forum social App and assuming that "users added more than 7 friends within 10 days before registration" as the leading indicator, we calculated a set of data:

Among them, if the user adds more than 7 friends before 10, the possibility of staying on the 30th is 99%; If the number of friends added is less than 7, the probability of not staying (losing) within 30 days is 95%, and the reliability of comprehensive indicators is 0.9405.

Similarly, calculate the credibility of the following two leading indicators:

Finally, we get a comparison:

The above data is only hypothetical. In fact, we need to compare a dozen or even twenty behavioral indicators to find out the behavior with the highest credibility.

The first rule of this model is "add more than 7 friends within 10 days after new users register", which is the classic "aha moment" of Facebook. The so-called "aha moment" is the moment when users realize the core value of products, which is our "leading indicator".

(Facebook, Instagram recommended friends screenshot)

In addition, leading indicators should meet the following conditions:

Second, the period of rapid growth.

After the initial stage of product polishing, the product retention rate is good, and at this time, the product begins to enter the spontaneous growth period. In the product stage of spontaneous growth, we still need to pay attention to data such as user retention, user duration and user portrait changes. However, we can focus on the management of users' whole life cycle, including the growth, activation and trigger of new users to the full funnel analysis of stable and active users of products.

Growth and activation of new users

Among them, there are generally two ways to add and activate new users. The first way is to construct the virus transmission coefficient of the product to make the product grow spontaneously. The classification of user virus transmission mentioned in the book Lean Operation Data Analysis is very interesting:

Protovirus, that is, the way to spread the attracted new users through the invite friends function of the App itself;

Word-of-mouth virus, that is, through word-of-mouth communication, users actively become new users through search engines;

Artificial virus, that is, through manual intervention, such as prize invitation and other incentives to encourage users to invite behavior.

An indicator of concern here is called "virus transmission coefficient", and interested students can learn more about it themselves.

New user download-> Activate-> Aha moment'-> The product is stable and active.

After the product begins to grow spontaneously, it is necessary to pay attention to the life cycle of users from new users to active users (after retention) to core users, and refine and refine the key indicators of each process.

situation

Take the previous forum social APP as an example. When new users enter the product, they will see a welcome page (as shown at the bottom left). After registering and logging in, they will see the home page of the product (as shown at the bottom right). Most applications have a similar process:

The process from a new user entering the welcome page of the App to becoming a core user is probably like this: new user (exploring and discovering product value)-> Onlookers (gradually recognizing the product value and having a certain sense of participation)-> Producer (identify with product value and actively participate):

According to popular methods, product life cycle (PLC) is divided into initial stage, growth stage, maturity stage and decline stage. At each stage of the product, the work weight and analysis emphasis of data analysis are different. Let's talk about it in stages with the case:

At this point, the user behavior at each stage is decomposed into indicators:

New user &; Explore the discoverer:

Welcome page bounce rate

New user registration rate

Conversion rate of new user guidance process

Initial seed feed page bounce rate

Search result conversion rate

Push license opening rate

Onlookers (passers-by):

Average number of attention cards per user

Average number of other users followed by each user.

Average likes/shares per active user

Number of feed cards displayed

Number of hits on the Feed card

Subscription content push click rate

Content producer:

Average number of posts per active user

The average number of photos and videos sent by each active user.

Average time spent by each user in the forum.

Behavior distribution of active users in the forum

The fine splitting of behavior indicators in the early and middle stages of the user's life cycle will help the product to constantly polish the details during the rapid growth period and continuously improve the user experience from the new to the core. At the same time, after the data of each node was perfect and stable, the students of product operation began to carry out various promotion, expand the plate and occupy the market.

Third, maturity

With the rapid growth of users and the continuous improvement of products, the focus of data operation began to shift from the first half of the user's life cycle (attraction, activation and retention) to the second half of the product before and after its maturity (loss and return).

Here, we share a data template called "Daily Net Change" (from john egan @Pinterest), which focuses on growth and maturity. Different from only focusing on DAU and MAU data, only focusing on the increase or decrease of the number of active users. This model can help to visually observe the growth factors of users or the changes of users' plates, and show the addition, return and retention of products through a picture.

The net change = new users+returning users-lost users.

New users refer to how many new users join in that day.

Return users, that is, how many old users have been useless for 28 consecutive days and started using them again today.

User churn refers to how many existing users just used the application for the last time 28 days ago.

Loss and backflow

In the process of paying attention to the loss and return, the data will reveal a change in the current user sector. For a detailed analysis of the reasons for the loss, please refer to the following process:

The core idea is to determine the reason of loss through qualitative return visit and data verification, change the product operation strategy, prevent users from losing or pulling back, and thus promote regression.

In addition, for some stable distribution channels, the common improvement methods may be limited. At this point, more accurate channel analysis can be conducted to optimize and improve the return on investment:

Case:

Improve the return on investment

Fourth, the recession period.

Finally, when the product enters the recession, there are generally two ways to take before entering the recession:

1, scale

It often appears in the retail industry. For example, if you open a massage health store and get favorable comments in a certain range, when the products are mature, you can start the franchise chain model, and form a brand effect and barriers by rapidly and widely expanding the market. Resist the risk of recession at this time.

2. Ecology

When the product grows or is close to perfection, it is easy for a single product to have the problem that the demand is too vertical and users cannot form dependence. We can develop new products with synergy, build a complete product ecosystem, and let users who can't satisfy or lose interest in current products drain to new products and become new users of new products. At the same time, users of new products can also drain back to old products, forming an interdependent chain between products, and end users can effectively flow and form an ecology.

This article is reproduced from Sohu, author: Shangzhu Technology, link:/A/217398072 _ 501610.