With limited resources, how can entrepreneurs make a good App?

[[153097]]

Today, the cost of acquiring users for apps is rising, while the life cycle of apps is getting shorter.

"In March 2015, the cost of acquiring a single O2O user was between 120 and 150 yuan, while that of ride-sharing apps was 180 yuan and that of P2P finance reached several hundred yuan," said CEO Wang Jun of Testin Cloud Testing.

On the contrary: "The average life cycle of an app is only ten months, and 85% of users will delete the downloaded apps from their phones within one month. After five months, the retention rate of these apps is only 5%." iMedia consultant Luo Yuanmei told the media.

This means that even if users are acquired through high operating costs, they may still be lost in the end. So how to observe users keenly at every critical period of App development? This article attempts to give you some suggestions from a data perspective.

The simpler the indicator, the more valuable it is

Traditionally, there are six dimensions to measure an app:

1. Terminal attributes (platform and manufacturer, model, OS version, resolution, language used, resolution, etc.);
2. User sources (channels, advertising, etc.);
3. App related properties (App version, etc.);
4. Region (country, province, etc.);
5. User attributes (gender, age group, etc.);
6. Behavioral dimensions (new and old users, whether paying, frequency of use, etc.).

New users, active users, and retention are the most basic indicators for measuring apps. New users are a subset of active users, and retention rate is a slightly more complex calculation based on the active rate. The number and duration of user sessions, which assess the depth of the active rate, are more detailed indicators for measuring active users.

Yan Hui, Vice President of Product at TalkingData, believes that different types of apps should be combined with indicators from different fields. For example, for gaming apps, the most important data indicators are registration conversion, revenue, ARPU, ARPPU, and lifetime value (which refers to the average revenue generated by a new user after the user reaches a certain lifetime number of days). The more important indicators in the pan-e-commerce field are: customer order conversion, customer unit price, and ARPU.

The more core the indicator is, the more it can reflect the development status of the App. After the DAU is broken down in detail, the proportion of new and old users can be calculated. The change in the user proportion reflects the changes in the product life cycle.

The life cycle of an app can be divided into four stages: start-up, growth, maturity, and decline. The data focus of each period is different.

Liu Liming, product operation of Umeng, suggested: "During the promotion period when the App is just launched, in order to increase the popularity of the App, a series of promotion activities need to be carried out to increase the user scale and collect user feedback. During this period, the focus should be on the growth rate of users and the optimization of the product. If the number of new users does not match the promotion intensity, it means that there is a problem with the promotion. It may be that the promotion method selected is inappropriate or the promotion population is not accurate enough. Or the reason lies in the problem of the product itself, the product interaction design is not user-friendly, there are bugs, etc."

"During the growth stage of a product, we must strike while the iron is hot. We cannot be as extensive as during the promotion period. We must carry out detailed promotion with the goal of acquiring high-quality users. Therefore, we must pay attention to new users, retention rate, and user behavior data (data such as usage time, usage frequency, and pages visited). If the retention rate and user usage data (duration, frequency) perform poorly, it means that the channel quality is poor, and we must adjust the promotion strategy appropriately."

"The user base in the mature stage is basically stable, the user scale and active data have stabilized, the threshold for attracting new users is already high, and the focus should be shifted to existing users. During this period, it is necessary to focus on the data of active users and churned users. Once there is a large fluctuation in active users and the proportion of silent users increases, it means that the number of churned users is increasing, and there may be problems with the operation strategy. At this time, developers need to pay attention to and strengthen the operation of the product to maintain the enthusiasm of old users for the product."

A new measurement method that is worth learning from abroad

In recent years, popular data indicators in Silicon Valley include NPS (Net Promoter Score). For example, to measure the popularity of an app, you can calculate how many users of the app will recommend it to their family and friends after using it. In addition, Yan Hui mentioned similar indicators, such as the K communication factor, which calculates the average number of people each active user will share with.

Liu Liming believes that there will be more new measurement standards for mobile games, such as eCPA (effective user acquisition cost), ARPDAU (average daily revenue per active user), ARPPU (average revenue per paying user), and LTV (lifetime value of users).

eCPA is an extended indicator of user acquisition cost, which refers to the effective user acquisition cost. That is, the total cost of acquiring each user plus the download volume attracted by the user. It can also be understood as the viral spread ability of the application, which is somewhat similar to the K spread factor. For example, if a promotion costs 5,000 yuan and brings in 2,000 users, but brings in a total of 4,000 users that month, then the CPA is 2.5 yuan, while the eCPA is 1.25 yuan.

LTV (Lifetime Value) refers to the average cumulative spending of each player in the game. This indicator includes both paying and non-paying players. It can be calculated by ARPPU (Average Months Per User).

There are different ways to measure apps from startup to different financing stages

"From round A to round C, investors focus on the entrepreneurial team, product concept, business model, product scalability, replicability, and profitability. After a company completes round A financing with a business plan, round B financing is mainly to verify the business model. Round C financing is a further verification of profitability and scalability."

"Therefore, during the B round of financing, investors will focus on the company's operating cost situation, team expansion speed, product growth rate and other data that reflect market attributes. During the C round, investors will pay more attention to the explosion point of products and markets, sustainable operating capabilities, and the ability of projects to expand in preparation for listing," said Chen Benfeng, founder and CEO of Yunshipei.

Based on this point of view, some data dimensions can be divided from the angel round to the C round.

Angel round. Investors are most concerned about the DAU trend, whether it has a high slope and a fast explosion. The proportion of new users in the DAU composition, whether new users are obtained through natural increase in communication or channel promotion.

Round A. Still DAU trend and stable retention rate. Also pay attention to quality indicators such as frequency and duration of use.

After rounds B and C, the data will be monetization data or data related to business models.

"Three months ago, good data and user accumulation might have impressed investors, but now investors are more concerned about the financial situation of entrepreneurs and whether you can survive before and after you "burn" the investors' money." Testin CEO Wang Jun suggested: "The cold winter is coming, and entrepreneurs should also pay attention to the market share in the sub-sectors, the share of active users, and the share of coverage to ensure their survival."

User portrait analysis

User portrait research is increasingly becoming a key area of ​​App data analysis. By adding tags to users, we can find people with greater consumption potential and accurately push products or activities based on user portraits during promotion periods to improve conversion efficiency. According to some data indicators provided by Umeng and TalkingData, user analysis can have the following dimensions:

Demographic attributes: gender, age, marital status, property status, vehicle status; for example, income status, education level, occupation, etc.

Hobbies and Interests: Describe a person's preferences in various aspects, such as playing games and traveling.

Recent intention: If you are interested in certain items or content recently, or have already consumed something, you will be interested in related products in the short term.

Location attributes: The simplest is the division of provinces and cities, and more valuable are the areas where people often move around and the types of places they often visit, such as airports, destinations, tourist spots, etc.

Living habits: daily routine, eating, dressing, living and transportation characteristics.

Traditional industries have always had a strong demand for portraits, such as Ping An and China Merchants Bank, which have transformed from traditional businesses to the mobile Internet. In the Internet circle, the types of apps with greater user demand are all related to profitability and are also related to the current hot fields; recently, they are more interested in Internet finance, mobile games, and various O2O.

Interview notes: How to choose a reliable data service company?

Currently, there are different companies providing data in the App market. Due to different data collection sources, the data provided by each company varies greatly.

Testin CEO Wang Jun believes: "Data provided by iResearch and Analysys are more similar to the opinions of authoritative experts. They adopt the sample statistical method. The accuracy of the data depends on the number of samples. The corresponding foreign institutions are similar to Gartner."

"The data service provided by Testin is like App Annie, which analyzes data through third-party data import and developer data services. Testin does not embed SDK into third-party apps to obtain data."

"TalkingData and Umeng obtain third-party data through SDK. They each have their own characteristics in terms of the types of apps they cover."

The focus on data is only a reference direction in the product development process. Yan Hui, Vice President of Products at TalkingData, believes that there are several aspects to consider when choosing a data platform:

First the functionality:

A. Business analysis function. All the necessary indicators should be available, and it can support multi-dimensional screening and analysis such as version and channel. In terms of business analysis, the functions of various platforms will not differ greatly.

B. Data capability. Data capability is not as simple as business analysis. You need to find a company that has experience in data processing and is very knowledgeable and concerned about the legality and security of data. You need a platform that can help you integrate data from the entire industry, identify the behavioral characteristics of each user, and provide more data supplements that you cannot obtain from your own business.

The second is scale:

You need to find a relatively large and reliable partner; start-ups and small data companies will have higher risks. Once they have any problems, it is equivalent to the loss of important assets for entrepreneurs. In addition, scale also reflects the stability of the platform from another aspect. The accuracy and stability of data are the first priority.