Are more and more people unsubscribing? You lack such a push data analysis method

Message push is the best channel for APP operation. It is a fast, accurate and effective message reminder. The main function of Push is to activate users and increase user activity. A survey pointed out that the retention rate of apps with message push is more than twice that of apps without message push. For e-commerce apps, a push can even bring significant performance growth.

But push is a typical double-edged sword. If used well, it can help you improve your retention rate. If used improperly, it will directly affect the user's return on the same day, close the push, or even uninstall the APP. This is because most operators are not properly aware of the negative effects of push, nor have they established a method to evaluate the quality of Push.

The above picture shows the entire process of push sending, from sending push, reaching users, user clicks, user consumption to possible fan to hater conversion. The conversion at each level corresponds to the delivery rate, click rate , conversion rate and unsubscribe rate (uninstall rate), and the unsubscribe rate (uninstall rate) may be the data that operations colleagues are most likely to overlook.

I will share my experience working with push analytics data .

The above table is the header of the push data analysis. Almost all the data required for push analysis can be found in this table. Next, I will briefly explain each data:

How do we analyze so much data? Then I will explain how to use this table from the following core indicators:

1. Reach rate

Arrival rate = arrival volume / delivery volume. I will use the following table to monitor the arrival rate:

There are two main reasons for low reach:

1. Due to technical channel reasons , the arrival rate is relatively low. This data mainly tests the technical level because signal loss occurs in the channel.
2. The user actively turned off the message push. Don't be afraid, this may not be entirely your fault, because when you open the APP for the first time, a pop-up window will prompt whether to allow the APP to push messages to you. At this time, the approval rate may be less than 50%; if the push operation is not good later, resulting in users prohibiting your push messages, then you may never have the opportunity to send push notifications to users again.

First, you need to examine whether you have too many behaviors that harass users; second, don’t be too discouraged. As long as you operate it carefully, you still have hope of obtaining users’ push permissions during their subsequent use.

At this time, you must give users a clear reason and implement it through core functions that meet user needs in the product, such as takeaway order status notifications, queuing progress notifications, ticketing information reminders, investment income reminders, etc. The following is an example. When a user who has turned off push notifications subscribes to an article, a prompt will appear to turn on push notifications.

2. Click-through rate

Click rate = click UV / reach UV

In my opinion, whether the click-through rate is high or not can be summed up in one sentence: push messages to users based on their usage scenarios.

To break it down, first you need to understand your users, then choose the right time, and finally push the right content.

If you want to optimize your click-through rate, you need to control variables and conduct a lot of A/B tests and small -volume tests.

1. Copywriting

The most important factor in push should be copywriting. Many operation masters have explained the specific plans, so I will not go into details here. You can search for PUSH operation work on the Internet, which is basically related to copywriting.

2. Time

Let’s talk about the timing of push. Different apps have different user groups, so the timing of push will also be different, but generally speaking, the time to push should be when people are relatively free. For example, for office workers, the best times to send messages are on the way to work and during breakfast time (9-10 am), lunch break (12-2 pm), on the way home from work (6-7 am), and before going to bed (21-22 pm), so you need to have a very clear understanding of your user base.

My test plan is as follows: keep your push type, strategy, copywriting and other factors unchanged, and conduct a small traffic test every hour from 9 am to 11 pm for users with the same attributes:

Because our APP is about food, the results we have tested before are that pushing messages after dinner between 19:00 and 22:00 has the lowest click-through rate and results in a large number of unsubscribes.

You can combine the data of your own APP and choose a time period with a higher click rate and a lower unsubscribe rate to push.

The click-through rate is inseparable from the unsubscribe rate. I will explain more dimensions in detail in the section on unsubscribe rate.

3. Conversion Rate

Conversion rate = conversion volume / click volume

Push messages must link to relevant pages. Users will click to enter only when they are interested in the content you push. If what they see is irrelevant to the message content, they will be disappointed and uninstall the app. After all, the steps for uninstalling are much simpler than those for turning off push notifications. If the number of unsubscriptions/clicks is higher than your usual push data, then check your landing page immediately.

To monitor the user conversion rate, it is necessary to conduct comprehensive tracking of the push landing page. If it is a marketing activity push, after the user comes in, the user directly exits the APP or stays on the marketing page and then converts into a purchase, etc., the data of the entire chain should be monitored to see which link has the highest churn rate, and adopt certain product strategies and operation strategies for optimization.

For example: 90% of users stay on a page for more than 30 seconds, but only 20% of users go to the next level page. Then you need to focus on checking the content of this page. If a trading platform puts a video and there is no obvious click reminder after the video, the bounce rate will increase.

4. Unsubscribe rate (uninstall rate)

Unsubscribe rate = Unsubscribe quantity / Arrival quantity

Many operations colleagues are trying their best to optimize the copy, and they are complacent when they see the click-through rate increase. However, 99% of operations have misjudged the effect of Push.

Each push has a corresponding unsubscribe rate. While push increases daily activity, it also increases unsubscribes and uninstalls! From the monitoring data, you can find that after the marketing push is sent, the number of unsubscriptions immediately rises to a peak. I have also experienced that the daily number of cancellations is greater than the number of people who open permissions every day. In order to reduce the number of cancellations, I did the following analysis.

1. Push time

Just like what we said in the click-through rate section, keep all factors such as push type, strategy, and copywriting unchanged, and conduct a small traffic test every hour from 9:00 a.m. to 11:00 p.m. for users with the same attributes:

Time is actually the user scenario. The copy should be combined with the scenario. Do not send push notifications in the middle of the night. Users who are grumpy will delete your app directly.

2. Push frequency

I conducted the following test using the control variable method, with the frequency ranging from once a day, once every two days, to once every 31 days:

The unsubscription rate is declining in a straight line. The highest unsubscription rate comes from pushing once a day, and the lowest is pushing once every 31 days. So control the frequency. If you send four or five messages a day, it would be strange if users don’t uninstall your app.

3. User access frequency

I conducted the following test using the control variable method. The average interval between users' visits ranged from 1 day, 2 days to 31 days. The shorter the interval, the more active the user:

The unsubscribe rate is rising in a straight line. The push unsubscribe rate of users with higher activity will be lower, while silent users may find out that they haven’t deleted this APP because of your push ...

4. User age

Divide your users into several age groups:

According to actual measurements, users aged 20-25 have the highest unsubscribe rate for push, and the unsubscribe rate decreases with age.

5. Split models

The actual test divided the models and found that iPhone7 had the highest cancellation rate:

The analysis of unsubscribe rate and click-through rate complement each other, and each dimension should be considered simultaneously. Determine the push sensitivity of different user groups, perform graded operations for users with different sensitivities, reduce the disturbance of push marketing activities to push-averse users, improve marketing effectiveness, enhance user experience , and thus establish a set of push strategies that are more suitable for users.

Judging the quality of push is not just about the content of the copy. Systematic data statistics and analysis are more important for improving push operations. My evaluation of push mainly relies on the reach rate, click rate, conversion rate, and unsubscribe rate. There must be an anchor point when comparing these data. The first is that it is not convenient to disclose. The second is that the situation of each APP is different. The specific value of the anchor point also needs to be summarized by the operations colleagues after a large number of pushes. In the initial stage, you can use the highest value of the reach rate, click-through rate, conversion rate, and the lowest value of the cancellation rate as the anchor point to continuously improve its push ability.

Experienced operators are welcome to add us as friends to discuss and give feedback!