A formula to explain Xiaomi's marketing secrets

The gradual saturation of the incremental market is no longer a new topic.

In 2019, the scale of mobile Internet users in China reached 1.135 billion, with the growth rate falling below 1%, and user growth has basically reached saturation.

The growth dilemma has also shifted from simply "acquiring customers" to how to "revitalize" old users.

The so-called activation refers to increasing user activity and improving LTV (Life Time Value) by reaching out to existing users.

In order to help customers further break through the growth bottleneck, Xiaomi Marketing adopts the thinking of "user life cycle management" to effectively "attract activity".

The necessity of the business of pulling in business can be easily understood through a case.

In early 2020, before a certain online service advertiser launched its activation strategy, its DAU and penetration rate in the Xiaomi channel were relatively low. Therefore, Xiaomi Marketing used the oCPX model in the initial stage to accumulate data samples, and at the same time selected advertising resources that supported Deeplink activation, combined with full media deduplication, as well as technical capabilities such as RTA and real-time filtering, and finally made reasonable optimizations to the materials. After launching Xiaomi’s marketing activation strategy, this customer’s activation efficiency increased by 15-20% and DAU increased by 113%.

Based on the thinking logic of "user life cycle management", Xiaomi Marketing has summarized a set of effective activation formulas, namely: activation efficiency = traffic*CTR*first call success rate. Effectively improve activation efficiency and promote user activity.

From Xiaomi’s marketing perspective, the first step to attract users is to divide them into three different types: active users, silent users, and churned users. At the same time, for major nodes such as the launch of new features within the App, e-commerce promotions, game version updates, etc., different targeted optimization strategies are used to deliver different information about the App to users in different life cycles, cultivate user usage/consumption habits, and improve LTV.

In fact, Xiaomi marketing has inherent advantages in achieving accurate activation of customer base. The integration of various data can effectively deepen the understanding of consumers and build basic judgments on users' information and consumption behavior. Multi-scenario traffic coverage allows advertisers to efficiently reach their target audience.

When users click to download the APP recommended in the advertisement, it is because they are interested in the advertisement content. When activating the APP, they hope to reach the landing page that matches the advertisement content in time, rather than the APP homepage. Therefore, it is also very important whether the traffic supports deep link technology so that users can directly access advertisements after activating the App.

After knowing who the consumers are and their active scenarios, we still need greater traffic to cover a wider user base so as to reach users through different resources. In addition to Xiaomi's own traffic under the Xiaomi ecosystem, such as Xiaomi video, music, pictorials, themes, browsers, app stores, system notifications, etc., Xiaomi's all-media traffic also introduces a large amount of traffic resources from the Xiaomi Advertising Alliance. And these mainstream media all support the Deeplink activation function.

The first-time success rate directly reflects the efficiency of the advertiser's advertising spending. However, the "live-pulling" business has always faced the problem that in most cases, the advertising platform can only confirm whether the app has been installed by the user, but cannot obtain whether the user is active.

If advertisers want to target inactive users for promotional campaigns, in most cases the advertising platform cannot complete this type of targeting due to lack of data. Although uploading audience packages can solve this problem, advertisers’ delivery strategies and audience selections change in real time, so it is necessary to continuously report manually compiled data, which still cannot solve the problem of timeliness.

Xiaomi Marketing uses system-level data capabilities to quickly determine user activity. On the one hand, it realizes the life cycle management of the global APP, including activation, secondary retention, activity and deeper conversion. On the other hand, since most advertisers evaluate the first-time activation indicator when launching campaigns to attract activity, Xiaomi Ads can accurately activate the campaign based on the activity of the APP. It can also more accurately distinguish high-frequency, silent or users who have already uninstalled the app, so as to formulate targeted promotion strategies.

For example, to recall users who have uninstalled the app, Xiaomi Marketing can identify the user's installation status and configure Deeplink for the ad, which will automatically activate the ad after the user downloads it, achieving "what you see is what you get".

After the launch, Xiaomi Marketing can also use system-level capabilities to deduplicate clicks across all media and avoid invalid reach within 24 hours. That is, after a user clicks on a activation ad for the first time, the system will mark the user and will not send new activation ads to the user within 24 hours. That is, it reduces repeated interruptions to users while maximizing budget utilization, thereby helping advertisers save costs.

At the same time, Xiaomi marketing supports real-time filtering through system-level device parameters, excluding active users across all channels, and then accurately locking in the first-time target population, effectively improving the first-time success rate.

In order to solve the needs of some advertisers who need to formulate delivery strategies for each delivery channel, or need to ensure data security as much as possible and choose different delivery strategies and plans for different users, Xiaomi Marketing chose to use RTA (Realtime API) technology to solve this problem.

The main purpose of RTA at present is a directional capability that can realize personalized crowd screening and real-time traffic optimization. Simply put, the decision of whether to place an advertisement is given to the advertiser. The basic targeting and bidding of the advertisement, as well as the bidding eCPM estimation are all completed on the advertising platform, and the advertiser is mainly responsible for evaluating whether the user/ad space can be converted into the target user or whether the ROI is reasonable.

This means that advertisers can filter or exclude some traffic based on their own needs before delivery, thereby reducing duplicate or invalid delivery. At the same time, advertisers can make real-time judgments on platform users, greatly improving the efficiency and effectiveness of advertising, reducing advertising costs, and maximizing the value of traffic.

After accurately identifying the target users, regardless of whether they are within the Xiaomi ecosystem or not, and when there are relatively complete optimization solutions in terms of data, algorithms, and channels, the next step is to start a dialogue with the users. The most challenging part of this stage is to understand users and the attributes of their media, and to impress them with appropriate creativity.

Therefore, creativity becomes the last lever to leverage conversion. Only by constantly iterating and optimizing creativity can we maximize the effectiveness of our delivery. So, how does Xiaomi Marketing optimize creative materials?

Nowadays, user habits are changing at an ever-increasing pace, and so is the speed of advertising creative iteration. According to data from DataEye, the material iteration cycle has been shortened from 30 days in the past to less than 7 days now.

To gain sufficient traffic advantage, you can place multiple sets of creative materials for a single ad and launch them simultaneously, gaining more user reach opportunities with the advantage of quantity, and achieve rapid innovation and iteration through data-driven development.

At the same time, in order to ensure the quality and quantity of the materials, Xiaomi Marketing divides the entire material delivery and testing into three stages, and finally screens out suitable advertising materials through multi-dimensional calculations such as advertising bids, click-through rates, and historical quality.

The first step is to test different types of materials in the initial trial stage, and initially place them in channels that are easy to obtain volume. By fully testing various types of materials, the materials that meet the standards can be screened out.

The second step is to place the materials that have met the requirements across all media resources, screen out the appropriate media, and test the matching degree between the material type and the media channel.

The third step is to expand the same type of materials that meet the standards, retest them as new types of materials, and conduct secondary screening. At the same time, qualified materials will be placed on applicable channels, and channels with poor results will be abandoned. The resource placement structure will be optimized and adjusted to achieve the goal of further controlling UV costs and improving the call-up rate.

Growth is an eternal topic in the business world.

In today's rapidly changing mobile Internet environment, the importance of tapping the value of existing users is gradually becoming apparent for some applications that already have relatively high traffic. The business of pulling in customers is a business model that was born out of the integration of Xiaomi’s marketing capabilities and market demand.

Author: Innocent Roland

Source: Innocent Roland