On September 17, Aurora BDS product manager Liu Jishi shared the theme of "How to achieve refined user access" through an online live broadcast. From the "big data of user access, the "5" factors of user access , How to achieve refined reach and refined reach examples" 4 aspects, in-depth sharing of how Aurora achieves refined user reach, and helps customers grow their business.

The following is the core content of sharing, friends who missed the live broadcast can review it again~

1. "Big data" reached by users

We selected the application of TOP1000 and did some data analysis based on the funnel model.

As can be seen from the above figure, the delivery rate, average reach rate, and display rate are all above 80%, but the click-through rate data is only 0.07%. It can be seen that in the click-through rate section, the loss is the most.

At the same time, since the average notification permission closure rate is between 30% and 50%, it has a relatively large impact on the final click-through rate. And the trend of the number of notifications and the number of uninstalls is synchronized. The more notifications sent, the more the number of uninstalls.

Therefore, a conclusion can be drawn: the extensive notification reach strategy will lead to the closure of notification permissions or the increase of user uninstallation.

It is understood that an average of 126 apps will be installed in each user's device, and the average daily operation messages of APP will be sent 7.5. As a result, users will receive about 945 operation messages every day.

In general, users receive a large number of notification messages every day, most of which are ignored or cleared, so the click-through rate is very low.

2. The "5" factors that users reach

  1. Object

Who are the users considering the reach, and what is the profile of the user?

  1. opportunity

When is the time for us to reach, that is, when and under what circumstances?

  1. content

Based on the first and second points, think about what users need and are interested in, and provide them with corresponding content.

  1. aisle

How do I choose a more suitable way to reach users? For example, if the user is currently active on mobile devices, do I use Push, official account, or email to reach users? If this message must be reached, should I use Push or SMS first? How to calculate the cost?

  1. Target

What is the ultimate quantifiable direct goal reached by the user? Do a good job of data attribution to achieve user reach goals.

The more accurate these five factors are, the higher our goal achievement rate will be.

3. How to achieve refined reach

There are four suitable for achieving refined reach: to the right object, at the right time, and in the right way to send the right message to achieve the attributable goal.

  1. How to find the right target to reach?

The common practice is to rely on big data, or cooperate with a third party, purchase, etc., and then make the user profile of application big data by yourself. The second is the production of application-oriented accurate portraits. Different rules can be used to make data burying points to produce smart labels.

Based on the above two steps, we finally need to settle into user groups. User grouping is the best way to delineate the target group. Combining prediction algorithms can delineate potential target groups; combining historical click rates and historical delivery rates can assist decision-making.

  1. How to find the right time?

One way is to deliver what users need and initiate timely interactions.

For example, if a female user is searching for eye cream, we can know that her preference is beauty makeup and skin care, and we can also know her approximate age, so that we can send a related product push. This is a common solution in e-commerce to find the right time.

The other two methods are scene assist conversion and preferential incentive conversion.

In the O2O scenario, customers have their own offline stores, and they will request a scenario to delineate a geographic target, such as about 500 meters from the store, triggering a message to be sent.

Judging at the right time is to promote and stimulate conversion.

3. How to use the appropriate channel?

As the operational positions become more and more dispersed, how to choose a suitable channel has become a difficult problem. At present, there are channels such as APP Push, WeChat Official Account, WeChat Mini Program, and SMS. When we need to implement the operation plan, we need to consider which channels are needed for this contact, what is the optimal contact sequence, and whether it needs to be reissued?

For example, when sending a must-reach message, in order to save costs, you can select APP Push first. If the user does not click or receive it, you can choose to send the SMS within a few minutes, which saves costs and ensures that the message must arrive. .

Take Push as an example. There are many secondary sub-channels, such as VIVO and Huawei. They all have their own vendors, which we call vendor channels. However, channels of different vendors have their own configuration restrictions, such as VIVO, which is the most restrictive of mainstream domestic vendors. The quota of operational messages that it allows to send every day is limited, and an error will be reported if the quota is exceeded.

In actual operation, due to the limitation of the manufacturer's channel, many messages will be intercepted in the manufacturer's channel, resulting in failure to push to users. Normally, the delivery rate of the vendor channel should be above 90%, but in many cases the delivery rate of the vendor channel is only 50%-60%.

This leads to another question: how to save quota, how to do smart routing and how to automatically reissue?

For example, when the channel quota of VIVO is full, can we be an early warning, get an intelligent route, and use, for example, the Aurora channel or a third-party channel for message delivery? This will not only ensure that our messages will not be lost, but also ensure the maximum delivery rate.

  1. How to set effective goals?

When we make a goal, we must ensure that it is an attributable direct goal. For example, when we push coupon messages, we only need to set one goal: whether the coupon is received. This is the most direct goal we reach.

The unattributable goals include product GMV, conversion rate sales, etc. It is not the most direct goal reached this time, but the long-term core goal.

4. Examples of Refined Reach

Based on Aurora’s accumulated technical and operational experience in serving a large number of customers, Mr. Liu Jishi shared a number of practical cases in the live event. The following 3 most representative ones are selected.

  1. Difficulties of scattered operations

When you have multiple operating channels such as emails, mini programs, and official accounts, how do you solve the problem of scattered operating positions?

It can be considered from three aspects. The first is to solve the problem of data unification. For example, what users have done in my APP, applet, and official account, and what their identity is, can be resolved in the data integration.

After reporting the basic data of the data source, there will be a problem-the problem of user unity, which is an old and difficult problem. For this problem, it is recommended to build a multi-level dynamic user identification system, which can solve the user identification problems of single device account, single account and multiple devices, and multiple accounts and multiple devices.

Finally, the user touches the unity. If the user uses more than a dozen small programs, they will receive more than a dozen touched messages, which is very disturbing to the user, because it needs to be touched to reduce the interruption to the user. .

2. Low user activity and active duration

During our operations, we will encounter situations where user activity is low and active duration is short. Here we recommend that you use the intelligent timing prediction algorithm. The smart timing of operational messages can be judged based on user historical click data, device activity status data, and industry click historical data.

If operated in this way, the click-through rate of push can be increased by 20% to 150%, which is a considerable increase. If it is a million-level daily activity, the increase in click-through rate may be very large.

The click-through rate is low. In many cases, this news is very common and has little in common with users and cannot resonate.

You can use variable-passing personalized copywriting in copywriting, for example, using the customer's name, plus some of their personal characteristics, and the click-through rate will be relatively doubled.

Finally, in the area of active users, it is recommended that you access the short video service, which is especially suitable for tool applications, or if the user is more frequent but stays in this APP for a short time, it is suitable to use this service.

  1. Important news must be reached, and the company must save costs

The following requirements must be met, but the company's costs need to be saved.

Operational requirement 1: The account information for successful payment must be delivered to the customer, but the company's annual cost of text messages is millions of yuan.

Operational requirement 2: The public announcement information of securities funds must be delivered to users and clicked to view the news.

Operational requirement 3: A user adds the product to the shopping cart on the app, but finally completes the payment through the mini program.

recommends that you use Jiguang's UMS unified messaging system. The above three requirements can be flexibly created according to the needs of the operation scenario, and various channel sending steps, triggering times and reaching conditions can be flexibly created.

V. Summary and Review

Refined reach formula: send the right message to the right person, at the right time and in the right way, to achieve the attributable goal.

Six elements of refined reach: accurate crowd, predicted crowd, operation plan, intelligent execution, reach channel and analysis improvement.

At this point, the content of how to achieve refined operation is finished sharing, thank you everyone!


极光JIGUANG
1.3k 声望1.3k 粉丝

极光(www.jiguang.cn)是中国领先的移动大数据服务商。其团队核心成员来自腾讯、摩根士丹利、豆瓣、Teradata和中国移动等公司。公司自2011年成立以来专注于为app开发者提供稳定高效的消息推送、统计分析、即时通...