Here are seven key points about APP message push

So, how to build a good APP message push mechanism? This article may help you.

This article will describe the following seven aspects:

• Part 1: Message Type
• Part 2: The impact of push notifications
• Part 3: Timing and scenarios of message push
• Part 4: Content of message push
• Part 5: Message Push Strategy
• Part 6: Data indicators of message push
• Part 7: Technical aspects

Part 1: Message Type

Based on product form and business type, from a broad perspective, message types can be divided into "IM type" and "non-IM type". Non-IM type can be further divided into the following different types according to its application in actual scenarios:

IM

Including WeChat , Momo, QQ, etc.

Non-IM

1. News and information: such as NetEase News , Toutiao , Tiantian Kuaibao , ZAKER and other news and information;
2. Marketing activities : such as Tmall APP pre-sales, big sales, full rebates and other marketing activities information;
3. Product recommendations: such as Tmall and JD related product recommendations, new product recommendations, and other personalized recommendations based on big data and algorithms;
4. System function category: such as Tmall delivery or arrival logistics notifications, birthday greeting reminders, coupon expiration reminders, and other system message pushes associated with personal information characteristics or behavioral attributes.

Part 2: The impact of push notifications

Whether the message push mechanism is used well or not will often have a significant impact on users' user experience and product functions. Specifically, the good and bad effects of message push can be summarized as follows:

benefit

1. Information notification and reminder

Message push serves as an information distribution channel . The platform can push information about product updates, content updates, event news, etc. to users through the message push channel, allowing users to be aware of such news in the first place.

If it is a promotional activity such as online events, e-commerce promotions, etc., users can also know and participate in it at the first time.

2. Promote activity and enhance stickiness

Message push is also a battlefield for user operations , and many products use it as a window to interact with users. Proper use of push messages can effectively promote user activity, increase user stickiness, and encourage users to open your APP from time to time to play.

3. Awaken silent users and improve retention

Some users may leave the app while playing it, and may not open the app for several months. Through interesting push messages, perhaps some of the lost users can be awakened. Even when users are about to leave, it can serve as a valve for the loss warning mechanism to prevent users from leaving too early.

4. Improve the utilization rate of functional modules

When a product updates its version, users can be reminded immediately through push messages, and guided to open and use new features, which can increase the usage rate of specific functional modules.

harm

1. Harass users and increase uninstall rate

Some operations colleagues are simply under pressure from KPIs, or have not established an effective operation strategy for message push, and do not understand user behavior habits. They simply proceed from their own interests - they just want to increase the frequency of users opening the APP.

Therefore, frequently pushing messages to users and disturbing them is often not worth the cost, accelerating user loss and even causing users to uninstall the app directly.

2. Overdrawn trust, the story of "The Boy Who Cried Wolf"

In order to attract users' attention, some operations colleagues have worked hard on push titles and come up with many crooked ways. They often use false titles to attract users. When users open the APP and see the content, they find that it is far from what is described in the title.

Users often feel "cheated". Over time, they understand your tactics and no longer trust you. When you have real content to tell users, they no longer open it. Because they have subconsciously determined that - damn it must be a lie again!

3. Too much worthless content causes users to feel disgusted and numb

Message push can attract some users to open the APP, but how to make users have fun in the APP and like your APP is a question worth thinking about.

Of course, this depends on many factors, including product functionality, content value, visual aesthetics, interactive experience, and so on.

But one thing is very important. Since it is a content push, the pushed content must of course be what users like. If most of the content is worthless to users, it can easily lead to user disgust.

Part 3: Timing and scenarios of message push

opportunity

The timing here refers to the push time period of each day. Generally speaking, there are several time periods for reference:

1. 9:00-10:00 a.m.: People have just arrived at the office and are generally not too busy, so the push message opening rate is relatively high at this time;
2. 12:00-14:00: This time period is the lunch break, when people usually check information, news, and use WeChat, etc. During the peak usage period of the APP, push messages and the opening rate is also relatively high;
3. 5-6pm: People are getting ready to get off work during this time. They are more lazy and relaxed. If they have time, they will generally choose to play with their phones. At this time, push messages will attract users' attention and will not be ignored.
4. 9pm – 10pm: After a day’s work and dinner, people can finally take a break. When people are reading news, chatting or playing games , push a message and the open rate is guaranteed.

Of course, these time periods are not absolute, there are definitely exceptions, and this time is for reference only. Another method is that we can reverse the push timing based on the user's actual usage scenario.

What does it mean? It means putting yourself in the real user scenario. For example, what do you usually do during lunch? Open WeChat to deal with the private messages you didn’t have time to reply to all morning? See what people are talking about in the group? Open the news app to see what major events happened today? Or play a game of King of Glory?

By using your own real usage scenarios to infer the push timing, you may be able to achieve good results.

Scenario

What does this mean? It actually means that we need to analyze the user's historical behavior and decide in what scenarios it is more reasonable to push messages.

For example, Tmall’s delivery or arrival notification. For example, you bought a product on Tmall last night, and you want to know whether it has been shipped this morning. Normally, you have to open the APP to check it yourself. But if the Tmall APP pushes you a message at this time - "Your product has taken special plane No. 10086 to see you!" Would you be very happy? Doesn’t it feel like a great experience?

The same principle applies to other possible application scenarios, such as coupon expiration reminders, birthday greeting reminders, etc.

Part 4: Content of message push

This is the top priority of the message push mechanism. A good title is crucial to grab the user's attention and attract the user to open the APP in the first place. Regarding the design of title copy , there are several suggestions and references:

• The copy should be concise, hit the point directly, be strongly related to the user himself, and stimulate the user to trigger.
• Title design should follow the " AI DA rule". (What is the AIDA rule: attention-attract attention, interest-generate interest, desire-arouse desire, action-click or purchase behavior).
• The title can be in the form of "numbers + emoticons". The expressiveness of numbers is far greater than that of text itself, and emoticons can enhance the user's sense of freshness and pleasure, attracting users to open and view.
• The content should be related to the product attributes and business model, and be careful not to blindly touch upon politically sensitive topics or gossip news.
• Form a scenario-based copywriting paradigm, such as Xiaohongshu : "scenario + related users + numbers + judgment words = a Xiaohongshu push".

1. Scene: Country name, transportation and other geographical environments or conditions, such as Australia, on the plane, before a date...
2. Related users: nouns and pronouns related to the user, such as woman, you, foodie, etc.
3. Numbers: ordinary numbers, rankings, percentages, and all other words related to numbers...
4. Determining words: such as equal to, should, enough, etc. (Note: According to the new advertising law, extreme words cannot be used)...

Define the push landing page

It is necessary to define which page the user will go to when opening the message. If you should go to the homepage, go to the homepage; if you should go to the details page, go to the details page; if you should go to the activity page, go to the activity page.

If you initially expect users to go to the activity page, but you define the push landing page incorrectly, users actually go to the home page. This is quite embarrassing. Not only will your intended purpose not be achieved, but users will also find it confusing, and it also reflects that you are not careful and professional enough.

Part 5: Message Push Strategy

Message push frequency

Before talking about the strategy, let’s first talk about the frequency of message push. So how can we control the frequency of push notifications so as not to cause user disgust? Here are some suggestions and references:

1. Adhere to "restraint"

Remember one thing, only when you like someone will you be unrestrained, only when you love someone will you be restrained. It is not a good thing to push messages to users repeatedly.

2. Determine the frequency of message push based on user usage frequency

The product form determines the frequency of use, and the frequency of use determines the frequency of message push. For example, tool apps (such as Amap) are usually opened when users go out, and the frequency is generally low. Another example is communication apps, such as WeChat. Just think about how many times you open it a day and you will understand.

Therefore, we can infer the frequency of message push based on the frequency of users using the APP, find out certain rules, and improve the accuracy of message push in the time dimension to avoid disturbing users too frequently.

3. Reference: Generally maintain 3-5 pieces per week

Message push strategy

In the process of message push operation, constantly summarizing the rules, analyzing user behavior data, and establishing and improving a relatively standardized push strategy can achieve twice the result with half the effort. Here are some suggestions and references on how to do it:

1. Tag push

That is to label people. The basic user tag categories include device information, user information, behavior information, discount information, other information... Various tags are given to each user, and then users are classified according to the tags, and differentiated messages are pushed to users with different tags.

2. Account system push (alias push)

A user database is constructed using demographic attributes such as name, gender, age, education level, and region known to the platform. Filter and classify users in the database and make targeted push notifications.

3. Multi-dimensional push

Taking Umeng as an example, there are several dimensions for reference:

1. APP version, distribution channel, geographic location (currently through IP positioning, support accurate to province and municipality);
2. Consumption power, purchasing power, groups (Central China and South China, tall, rich and handsome, fair, rich and beautiful, short, ugly and poor, single nobles, losers, etc.);
3. User activity (active for x days, inactive for x days), model (including popular models and all models), gender (male or female);
4. Multi-dimensional combination, take the intersection or union and push.

Push notifications for silent users

Content can be specially designed to be pushed to silent users, awakening them, improving user retention , and reducing disturbance to other users.

Push of upgraded version

When a new version of the product is released, a message will be pushed to users of the old version to remind them to update the version and experience more new features.

Limit sending speed

The amount of information pushed at one time is generally quite large, reaching hundreds of thousands, millions, or even tens of millions. In order to reduce server pressure and avoid high concurrency, you can limit the sending speed when pushing, and the number of messages sent per second.

A/B testing

In actual work, we often have multiple options to choose from, but we cannot be sure which one is more reasonable.

Therefore, we can conduct A/B testing for different groups of people. It includes two dimensions: population structure data and historical behavior data.

The operation methods of all A/B tests are similar. Just control one of the variables. I won’t go into details here.

push the crowd

The push population includes targeted push and full push. Although targeted push has a narrower coverage, it can significantly improve the push matching degree; while full push has a wider coverage but lower accuracy.

It can be said that each has its own advantages. As for how to do it specifically, operations colleagues need to slowly explore and make timely adjustments in their daily work.

Define multiple push priorities

As mentioned above, there are many types of messages. Corresponding to actual work, we also have many types of messages that need to be pushed to the user side. So how do we evaluate the priority of these messages? Reference: General system function category > Marketing activity category (there are exceptions, such as e-commerce promotion)

Copywriting Horse Racing Mechanism

Another form of A/B testing. How to do it specifically?

The operator writes a number of copywriting (for example, 5), and then extracts some experimental users in advance (for example, 10% of the total number of users), and sends each copywriting to a part of the experimental users (10% ÷ 5 = 2%, that is, each copywriting is sent to 2% of the experimental users). The click-through rate of each copywriting over a period of time is observed. The copywriting with the highest click-through rate wins, and then this winning copywriting is sent to the remaining users (90%).

Part 6: Data indicators of message push

Regarding message push, there are several data indicators for reference:

Reach rate (number of people reached/number of messages sent)

Here we need to pay attention to the reasons for the low arrival rate. There are two main reasons:

1. Technical channel reasons
2. The user actively closes the message push permission

Open rate (number of people who opened the page/number of people who reached the page)

Conversion rate (number of conversions /number of opens)

Uninstall rate (reference: number of people who uninstalled after 1 hour of push/number of people who reached)

Why is it 1 hour? This is a conclusion drawn by someone based on data analysis and experience in actual work. I forgot where I saw it.

Retention rate of users who received push = 2 * retention rate of users who did not receive push

What does this mean? It means that over a certain period of time, the retention rate of users who received message push is twice the retention rate of users who did not receive message push.

Part 7: Technical aspects

Finally, let's briefly talk about what we can do on the technical level. There are two references:

Semi-automatic push

In the beginning, the product was immature and the user base was small, so there was no data to refer to. Message push was in the testing phase and could be pushed entirely manually by operations staff.

As the product develops, based on our understanding of users and the collection and analysis of their historical behavior data, coupled with the time constraints on message push, it is impossible to leave everything to manual processing.

We need to standardize a mechanism for message push to free up some manpower - standardize and productize push, form push events, and send them through a semi-automatic system.

What are semi-automated events?

The "event" of the semi-automated platform describes the entire operation behavior, which is composed of "specific groups of people", "trigger conditions", "operation rules" and "scheduling mechanism". Events are real-time and sudden.

An event includes: what conditions are met, what is done, what exactly is done, when and how to do it.

Scheduling mechanism: such as scheduling push system, user tag system, and associating system-related data.

Deduplication Filtering

As mentioned above, we will have multiple types of message push. If we don’t handle it well, a user may receive many messages every day, causing user harassment.

Therefore, we need to formulate rules to ensure that a user receives at most N pushes per day (N is usually 2), and that pushes with higher priorities are received first.

For example, multiple push notifications such as coupon expiration push notifications, event launch push notifications, birthday greeting push notifications, and personalized push notifications need to be deduplicated and filtered.

Conclusion

The ideas for this article come from daily APP usage observations and experience, the author's project experience, discussions with peers, and analysis and reference of online articles.

The purpose is to compile an article about APP message push from multiple dimensions as comprehensively as possible, and it is not entirely original. Please be advised.

Finally, I would like to share the mind map of this article with you. Hope this helps.

The above is what I want to share with you.

Mobile application product promotion service: APP promotion service Qinggua Media advertising

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