The difference between APP live advertising optimization and attracting new users!

As the number of Internet users and mobile Internet traffic gradually "peaks", for those top APP applications that already have high traffic, the user scale has reached a level close to "saturation". It is time to manage the existing users well and tap the value of existing users . Advertising has become one of the means to attract users.

Especially in the past two years, the demand for attracting business from leading enterprises or leading apps has become increasingly stronger.

The competition between them is more about active user indicators such as monthly active users (MAU) and daily active users (DAU) . Whoever can retain the most users will become the final winner.

1. The phenomenon of attracting new customers and attracting active users in daily life

The broad definition of "user growth" includes two layers:

One is the growth of new users (attracting new users), and the other is the growth of active users (attracting active users).

New User Example

I have never downloaded Pinduoduo , but I often see Pinduoduo’s “new user acquisition” bombardment to me.

Of course, those advertisements about buying electric cars or mobile phones for 9.9 yuan are well-known click-optimization methods and did not sway me at all.

As someone who has worked in advertising for many years, I am generally not impressed by advertisements and they cannot fool me.

But there is only one way, if I fail, my reputation will be ruined.

One day, I opened the 360 ​​Smart Camera APP and saw the following prompt: "No network found, content lost". I really thought there was a problem with the network, and subconsciously clicked "Refresh". I didn't expect to enter the H5 page of Pinduoduo, which had an attractive "Cash Red Packet Reward", but I closed it immediately...

(Real picture of Pinduoduo’s new advertisement)

I have mentioned before that in the advertising delivery chain, clicks are the first step in promoting user conversions. The example here is to let users click unconsciously, and then use various incentives on the H5 page to make users fall step by step.

Pulling example

One day I was searching for dishwashers on Suning.com and saw a Siemens dishwasher. Then when I was browsing the Douyin APP, an advertisement for Suning.com’s Siemens dishwasher appeared, and it also showed the region I was in at the time, “Meizhou”.

(Real picture of Suning.com’s live advertising)

The way this is achieved is that Suning.com uses the user data of its own mall (I have seen dishwasher products) and accurately targets me in the Douyin traffic pool, showing me advertisements in the hope of stimulating me to complete the purchase.

2. The similarities and differences between attracting new customers and attracting active customers

Since the "pulling live" advertisement entered the industry's field of vision, some questions have arisen in the circle:

• Which advertisers are investing in live advertising?
• How much does it cost to do a job?
• What is the difference between the optimization methods of APP activation and new user acquisition?
• …

Which advertisers are investing in live advertising?

We learned from the media that currently, live-streaming advertisements mainly appear in industries such as games, platform e-commerce, finance, and online services, such as Taobao, JD.com, Pinduoduo, Alipay, and Meituan.

Taobao is probably the one that is investing the most in attracting customers in the market right now. Some people may be curious: Taobao’s brand and market share are already so high, is it still necessary to attract customers?

In fact, this is related to the changes in current purchasing platforms and netizens' purchasing habits. In addition to Taobao and JD.com, there are also Pinduoduo, as well as the second-tier e-commerce platforms on various information flow platforms. User loyalty has become very low, and it is easy to switch platforms because of new things or new discounts. Therefore, everyone needs to use various red envelopes, discounts, bargaining and other incentives to grab or retain users.

How much does it cost to do a job?

The cost of attracting old customers is much lower than the cost of attracting new customers. After all, "the cost of maintaining an old customer is much lower than the cost of developing a new customer."

Currently, the cost of pulling work on the market ranges from a few cents to a few dollars. For example, the cost of attracting new users for a mainstream APP is as high as 100, but the cost of attracting active users is only more than 1 yuan.

Because you already know that they are your target users, you can directly and accurately target these users, saving the advertising waste of non-target users.

In addition, since this user is an old user, your customer education costs are much lower, and the user conversion path is relatively simpler.

What is the difference between APP activation and new user acquisition in terms of optimization methods?

Currently, most of the new customer acquisition advertisements on the market fail to simply attract new customers, but instead attract a mixture of new and old users. When creating an advertising plan, old users are not removed. At most, you can set up a filter for converted users at the company/account level (here you can only filter the converted user data since you opened an account on the platform).

There are two major disadvantages of mixing new and old users:

• Waste of advertising costs. The goal of attracting new users is to generate more new users, but because there is no distinction between new and old users, part of the advertising costs are also consumed on old users.
• This gives old users a bad impression. They are obviously your users, but you keep advertising to get me to download your APP? The worst thing is "killing the old users". There are many benefits for new users, but old users are not eligible to enjoy them, but they happen to see this advertisement.

Therefore, the new customer acquisition ads are still too crude overall and need to be optimized, without missing any details that can reduce costs.

Active advertising is to redirect existing users , which sounds much simpler than attracting new users. However, optimizers who have actually delivered live-streaming ads will find that the simpler the delivery, the more it requires sophisticated operational capabilities.

When it comes to attracting new advertisements, the focus lies on the data capabilities and algorithm capabilities of the media advertising platform. As for live advertising, it depends more on the advertiser's own data and algorithm capabilities.

Let me summarize the similarities and differences between attracting new customers and attracting active customers:

-> Similarities

• Multiple accounts and multiple plans : Whether it is to attract new customers or to increase activity, it is necessary to stack accounts and plans reasonably. For different testing strategies, it is best to divide them into different accounts or plans for comparison and verification. For example, to attract traffic, you can use the offline data packet upload method, the RTA method, or the DPA method. The impact of different methods on advertising effectiveness should be distinguished and analyzed scientifically by using accounts and plans.

(Screenshot of the small airplane that I used to launch an app to attract people to the project)

• Eat creativity and materials : Whether it is to attract new customers or to increase business, creativity and materials are the decisive factors in advertising optimization; whether it is to pile up accounts or plans, never do meaningless repetitive creation, but pile up on the basis of differentiated materials. In addition, with the ever-increasing speed of Internet communication and the ever-changing needs and preferences of netizens, the life cycle of a set of materials may be less than 7 days, so the speed of advertising creative iteration must keep up.

-> Differences

• Applicable objects : Attracting new users is applicable to all people, while attracting active users is mainly aimed at advertisers who have a certain number of old users and whose data are relatively valuable.
• Advertisement targeting : Attracting new users mainly involves setting up user targeting for the advertising platform, while attracting active users mainly involves targeting the advertiser’s old users, who need to be sent back by the advertiser to the advertising platform.
• Conversion cost : The KPI cost of attracting active customers is much lower than that of attracting new customers, but the conversion benefit may be much higher than that of attracting new customers. The cost of doing business that I have come across so far is basically between a few cents and a few dollars.
• Bidding method : The mainstream platforms now generally use OCPM and OCPC bidding to attract new customers, which is relatively simple and convenient. However, the cost of pulling business is lower, and using OCPM and OCPC may not have enough volume. There is another situation, that is, when a third-party agent helps advertisers to place ads but cannot obtain the advertisers' conversion data, it is recommended to use CPC.
• Conversion indicators : The indicators for attracting new users vary from industry to industry, such as the number of registrations, downloads, forms, fans, etc. However, the most important thing to attract new users is to arouse the number of UVs, with a focus on the arousal rate (arousal rate = number of aroused UVs/number of clicks). This indicator is related to the activity of the old user group itself and the degree of overlap with the media group. The way conversion metrics are passed back to advertising platforms is also different.
• Conversion chain : Since the users targeted at attracting new users are mainly those who have not downloaded the APP, the users targeted at attracting active users are those who have downloaded the APP but have not logged in for a long time or have even uninstalled the APP. For example, the new conversion chain is ad exposure-click-app market-download-activation, and the active conversion chain is ad exposure-click-deeplink to open the APP or through the middle page user clicks the deeplink to open the APP (you need to find the media to activate the optimization whitelist and direct link entry whitelist).
• Data and algorithms : Attracting new customers mainly relies on the data and algorithms of the advertising platform, while attracting active customers mainly relies on the data and algorithms of the advertisers themselves. Advertisements must be able to formulate clear data strategies to apply the data of old users. Especially in the case of supporting RTA bidding decisions + bidding, the specific price that should be offered for the current traffic requires advertisers to have a lot of ability to predict user conversion value.

3. Detailed explanation of the application of relevant technologies for live advertising

Since live advertising cannot be separated from the advertiser's user data, the advertising platform naturally needs to provide a channel to receive the advertiser's data. The main ways are:

• Advertisers manually upload offline audience packages to the advertising platform, and optimizers target the audience packages when adding advertising plans. It is suitable for advertisers with a relatively small number of old users and who do not need to update the audience package frequently. When the audience package needs to be updated, it needs to be manually re-uploaded.
• Advertisers upload offline audience packages through the MKT API interface , and optimizers target this audience package when adding advertising plans. It is suitable for advertisers with a relatively small number of old users and who need to regularly update their audience packages. The technology can achieve regular and automatic uploading.
• Advertisers collect population data through the advertising platform DMP interface , and optimizers target this population package when adding advertising plans. This method generally involves adding DMP code directly to the website or APP, and the DMP system of the advertising platform collects data. Optimizers can customize and refine audience tags in the DMP for advertising delivery.
• Advertisers use the RTA interface of the advertising platform to make real-time pre-bidding decisions , and optimizers set RTA strategies when adding plans. It is suitable for advertisers with large amounts of user data, high data value, and the need for multi-dimensional and refined operations, but this will also place relatively high demands on the advertisers' technology and consumption.
• Advertisers transmit crowd behavior data back through the DPA interface of the advertising platform. Optimizers set up DPA plans when adding plans, which can achieve accurate user targeting and control of creative content output. It is mainly used in the e-commerce industry and requires docking with the DPA product library. For example, for users who have collected products within 7 days, you can accurately target them with advertising creatives that display the products they have collected.
• Advertisers connect to the RTA+DPA interface at the same time, which is also mainly used in the e-commerce industry.

The RTA concept has been quite popular in the past two years. After talking to the advertising platforms, I found that RTA is currently mainly used in platform e-commerce, leading Internet finance, and large media apps because they have large amounts of data and high data value. Careful readers will notice that there is no mention of games here...

Practitioners who have not used RTA may be curious:

What is the difference between “RTA delivery” and “directly uploading data packages, splitting them into different directions and creating advertising plans with different prices”?

There are four major differences between the two:

(1) For advertisers with large amounts of data, high update frequency, high data value and high confidentiality , uploading data packages is not feasible.

I also revealed in the article "Let's Talk about MKT API and RTA of Advertising Platforms from the Shallow to the Deep" that for a certain live-pulling project I launched, because we had not yet applied for RTA docking with Massive Engine, we could only upload it manually every half an hour or an hour, and encountered many problems in the process:

• From the platform level, firstly, there is an upper limit on the number of crowd packages in the background. If the upper limit is reached, no more can be uploaded, and the previous ones must be deleted manually. Secondly, the upload time is slow, and the data package should not be too large. In addition, the platform has so many customers uploading every day, and everyone has to queue up in the same process to take effect (I have tried waiting for a day before it takes effect). Thirdly, even if the package is successfully uploaded and takes effect, there will be a delay in the advertising bidding to call the new data. The specific delay is only known by the internal rules of the platform.
• From a manual perspective, we are essentially doing repetitive work, and data packaging will become cumbersome and inefficient. At the beginning, I had to get up at 6 o'clock in the morning to upload, then go back to sleep, and then continue uploading after a while. Although we later achieved automatic uploading at a specified time through our own tools, it is still limited by the number of crowd packages set by the platform itself. So, the value of RTA comes out.

(2) Accurately target or filter data in real time, divide internal user data into more fine-grained categories , screen high-, medium-, and low-value users, and make real-time bidding decisions for users of different values ​​to gradually improve advertising effectiveness. Combined with DPA, we can truly achieve different faces for different people . This is very important. It is impossible to achieve this by simply relying on split targeting to build an advertising plan.

(3) Controlling the frequency of same-media/cross-media delivery can reduce repeated or invalid exposure and avoid advertising waste. Avoiding repeated exposure of the same user not only wastes advertising, but also may cause the user to dislike the ad or even the brand.

(4) Avoid repeatedly targeting users who have been successfully converted in one channel as new users in other channels . If real-time RTA is not possible, it is possible that users who have been converted in channel A will be repeatedly exposed as new users in channel B or the wrong creative content will be used, which will lead to the "killing the familiar" phenomenon mentioned above.

PS: When attracting new users, RTA can also help exclude old users and filter out blacklist users.

Author: Advertising Practitioners Study Room

Source: Advertising Practitioners Study Room