AARRR model case: How to use data to optimize channel delivery?

With the disappearance of the new population dividend brought by mobile Internet , the growth rate of new users in mobile Internet is getting slower and slower, the cost of acquiring new users is getting higher and higher, and the sources of new users are beginning to target the huge stock of users. The acquisition of each new user means that your competitors will have one less user.

QuestMobile active device count and growth rate

Data from QuestMobile shows that the growth rate of mobile devices has begun to decline, and the growth of mobile Internet users as a whole will face fierce red ocean competition.

From the perspective of cost optimization and new user quality, user growth has entered the stage of refined operations . How to acquire a large number of high-quality users at a lower cost is a difficult problem that all Internet product or operator personnel must face.

This article combines the currently commonly used online user acquisition channels , the AARRR model and data analysis to explore how to use data to optimize the acquisition of new online users.

1. Introduction to online channel types

Promotion channel type

There are many types of online channels, which can be generally divided into the following five types:

1. Social media : mainly featuring native advertising , such as Toutiao , Tencent’s Guangdiantong , Weibo’s Fanstong , Wi-Fi Master Key and other advertising platforms. After selecting the target audience, as well as appropriate selling points and materials, the cost of acquiring new customers can be controlled at 5 to 10 yuan per customer.
2. Search engine SEM: Baidu is the main one in China. It is necessary to do a good job in keyword optimization, search result copy optimization, landing page conversion effect, meet user needs and pain points, have novel content, and good experience. It can achieve a download conversion rate of more than 30%, and the cost of acquiring new customers can be controlled at 5 to 20 yuan.
3. Software store: The store has paid CPD promotion and daily event promotion. Do a good job of store software ASO and pay more attention to the store's theme activities. Participation in one event can increase the growth by more than 10 times.
4. Social media: well-known ones include Weibo, WeChat , etc. WeChat is suitable for explosive communication, while Weibo requires deep cultivation of fans and long-term investment to see results.
5. Traffic exchange alliance: The third-party traffic exchange alliance is relatively chaotic. When choosing partners, they usually exchange traffic through the apps of large manufacturers and guide traffic based on user scenarios. Such as the traffic flow between WeChat, QQ Browser and Tencent Video , and the mutual traffic flow between Taobao and Alipay . This method is suitable for cooperation with the same level of users. Otherwise, there will be a large gap in the exchange volume, which will affect the next step of cooperation.

2. Channel promotion model and core data indicators

There are many types of online channels and different ways of cooperation. If you don’t have core ideas and methods, you will be confused if you dive in head-on.

Here we can use the classic Internet customer acquisition model AR RR model for analysis:

AARRR Model

AARRR starts from the entire user life cycle , including Acquisition, Activation, Retention , Revenue, and Referral.

Applying AARRR to the acquisition of new users in a channel would correspond to:

Key user acquisition metrics

For each process, you need to have data thinking at all times during actual operations, pay attention to the changes in indicators at each step, and improve the conversion rate of each step, so as to reduce user acquisition costs and obtain better user quality.

3. Case Study

1. Search Engine Optimization (SEM)

Search engines have a great advantage in satisfying user needs: the keywords entered by users represent the current users' clear demands.

Once user needs are clear, the remaining core issue is how to combine the user's clear demands (keywords) with the characteristics of your own products, and display them as high as possible in the search results, while attracting user clicks with a higher conversion rate.

This step has three core indicators:

1. Keyword bidding: Popular keywords are more expensive, but have a huge display volume . Unpopular keywords are cheap, but lack sufficient exposure. In terms of overall strategy, the head keywords should account for 60%, and the remaining long-tail keywords should account for 40%. Give priority to satisfying users’ core demands (delivering mainstream keywords)
2. Search result CTR : Search results in Baidu are divided into two types: brand area, which is only open to brands and is guaranteed to be displayed in the first place . For ordinary result cards, bid in real time based on keywords, and those with higher bids are ranked first. Currently, search results are generally charged on a CPC basis, and no fees are incurred if there are no clicks. An excellent keyword CTR is the foundation. Only with enough clicks can there be the possibility of the next step of user conversion.
3. Results page conversion rate: After clicking, users enter the results page. The results page is the user’s first impression and the most important thing. Whether the user downloads the app and whether the user installs it after downloading are all reflections of the results page’s effectiveness. It is necessary to combine the current user demands and inform the advantages of your own product within 2 seconds, so as to achieve the effect of users wanting to download and try. The following is a case analysis of the placement of Zuoyebang and Yuantiku in Baidu SEM:

Baidu Keywords – Take a photo and search

Xiaoyuan search results page

On the search results page, Xiaoyuan Search Questions and Homework Helper are both on the first screen of results. From the result cards:

• Position factor: Both are displayed on the first screen, and Xiaoyuan Search, which ranks higher, does not have a clear advantage
• Result title : Yuantiku's "Primary and Secondary School Photo Search Tool" is clearer and easier to understand than Zuoyebang
• Content display factors: Zuoyebang displays more content on the result page, which is relatively cumbersome and users lack the patience to read

From the above three positions, the CTR of the result page here Xiaoyuan Search wins.

The results page also shows the attitudes of the two teams towards SEM:

• Zuoyebang: A simple page with no additional content. Unless users have a clear understanding of Zuoyebang, it is difficult to attract clicks to download.
• Xiaoyuan Search Question: The core of the product is clear at a glance, it is friendly to new users, and it hits the user's pain points, so the download conversion effect is naturally better. Here are the calculation methods of several core indicators:

Landing page click rate = download button click PV / result page display PV

Landing page download conversion rate = channel package activation times / download button click pv

Unit price for acquiring a customer (NU cost) = number of new users activated per day / total promotion cost per day. The calculation here ignores the number of users retained on the next day. If the product retention rate is low, the cost of lost users should be amortized into the NU cost.

The above indicators are used to evaluate the overall conversion effect of keywords and landing pages, and reflect the customer acquisition cost. Comparing Xiaoyuan Search and Zuoyebang, it is obvious that the Xiaoyuan Search team of Baidu SEM is more dedicated, and the conversion effect is naturally not bad.

2. Social native advertising (Toutiao)

Toutiao News Feed Video Ads

Taking Toutiao’s advertising platform as an example, Toutiao has several advertising styles in the information flow list or text page, and combines them with crowd portraits to achieve accurate advertising push to the target audience, with good overall conversion effect.

In actual operation, information flow ads are also divided into video ads and ordinary graphic and text ads. In terms of advertising presentation, video ads are obviously more expressive, and carefully produced videos are more conducive to user download conversions.

For social advertising, the core indicators that need to be optimized are:

1. Material CTR: material click PV/material display PV. The key to material click PV lies in the elements of the material itself, including product selling point extraction, copywriting optimization, and image production. Each step needs to be designed in combination with the population characteristics of the advertising platform. The click-through rate of materials is related to the advertising style, placement and precise audience strategy. The balance between click PV and display PV needs to be continuously adjusted in combination with the characteristics of different platforms to achieve the most optimized display combination.

2. Download (installation) conversion rate: download (installation)/material click pv. Advertising platforms can provide full-link data from download to installation to activation. Through the following indicators, problems and optimization points can be found:

• Download rate = download pv/material click pv
• Installation rate = installation pv/successful download pv, which reflects how many users who downloaded the software install it. The relevant factor here is the size of the installation package. The larger the installation package, the greater the possibility that the user will fail to install it (insufficient mobile phone space).
• Activation rate = activation pv/installation pv, which reflects how many users activate after installation. This indicator can be used to monitor whether there is channel fraud: the installation volume is huge, but the activation rate is low.

After the indicators of each link are optimized, the final data indicator is the reduction of the unit price of channel customer acquisition. The excellent result is a large number of new users, and the cost is controlled within an acceptable range:

Customer acquisition price = total number of new users per day / daily channel consumption cost

4. Channel user quality assessment

The user quality of a channel, in addition to the customer acquisition price, also needs to be combined with the product's own retention and user activity to comprehensively judge the commercial value of each user. By gradually analyzing the product data on the end, high-quality channels can be screened and data decision-making basis can be provided for product improvement.

Assume that materials are placed in the following four channels and run for a week. By observing the core indicators, the data are as follows:

Channel customer acquisition data

Judging from the data, channel A has the lowest customer acquisition price, channels B/C have poor user retention, and channel C has a low click-to-activation rate. Channel D has the best user retention, but the lowest number of new daily users:

1. There may be cheating in channel C : users click on the materials but the activation rate is low, and the user retention rate is the lowest. It may be inflated by third-party channels. It is necessary to reduce the budget or abandon the channel.
2. Channel B users may not be the target group of the product : User retention rate represents the basic dependence of users on the product. Channel B retention rate is relatively low, so it is necessary to verify the matching degree between the product’s selling point and the needs of this group of people, and adjust the selling point in time or abandon the channel.
3. Channel D has the highest user retention rate and good results, but the number of new users is low . We can increase the advertising costs in channel D, expand the number of new users, and further observe the user retention rate.
4. The overall indicators of channel A are excellent and can be deployed sustainably.

After analyzing several core indicators, the results after optimization are as follows:

Optimized channel customer acquisition data

According to the AARRR model, after user conversion, continued activity is still required to generate better revenue. At the same time, by guiding users to spread and share, new users can be used to attract old users, further reducing customer acquisition costs.

For AD channels, if you need to select the best from the best, the next step is to carefully observe the user behavior indicators on the terminal:

Average usage time per user: total usage time of the product on that day / number of users using the product. This reflects the user's stickiness to the product. The longer the usage time, the greater the user's dependence and the greater the potential for commercial realization (for example, the average usage time of Toutiao is 74 minutes per person).

Average channel time per person

Through the data, we found that for channel name A, its average usage time is higher. We further analyzed the usage of channel A in terms of product functions:

Function usage penetration rate: number of people using the function on that day/total number of people using the product on that day. Reflects the scale of each product's business functions among users

Channel Function Penetration

Average time spent on channel functions

By comprehensively comparing the usage penetration and average usage time of functions, we found that in channel A, function b has the highest penetration rate and average usage time. Therefore, the product should take function b as the core selling point, guide users well on the terminal, and attract more users to use function b.

5. Conclusion

The cost of acquiring customers online is getting higher and higher. Only through careful observation, good use of data, and bold assumptions and experiments can we carve out a refined path for user growth in the red ocean.

The author of this article @尹为业 is compiled and published by (Qinggua Media). Please indicate the author information and source when reprinting!

Product promotion services: APP promotion services, advertising platform, Longyou Games