This is an article that I wanted to write a few months ago, but on the one hand, there were too many things last year and I couldn't make time for it.
This is a common thinking trap. We often dare not take the first step to get 10 or 100, but in fact 0 and 1 are fundamentally different .
So this is still a rough article, but instead of 0, it is better to have 1 finished product.
01 Insurance also needs [growth hacking]
AARRR (Acquisition-Activation-Retention-Payment-Viral) is a well-known user growth model for Internet products.
Internet insurance has also explored the application of this model. The biggest fanfare is the 15-year Ping An, the "four markets, two clouds, and one door" around the needs of "medical, food, housing, transportation, and play" . strategy.
Today, major applications in this country email list strategy, such as Ping An Haofang, Ping An Hao Car, and Ping An Hao Car Owner, have all died down.
The practice of building a movie theater to sell popcorn has proven to be unworkable.
However, the insurance model of 1 yuan in the first month turns insurance products into a model similar to Internet products. The low threshold of 1 yuan in the first month carries the function of [acquiring customers-activation] , and a large number of users with low threshold join, how to get the follow-up user retention. , Repurchase has become a problem that major Internet insurance players have to pay attention to.
This is when the theory of growth hacking finds its best use in insurance.
02 Make a hypothesis
This point I shared a little in the article "How to Make Good Hypotheses in User Growth".
Last year, when solving the problem of customer retention in an important channel for the first month of the first month, I first delineated the boundary of version 1.0: that is, to improve the renewal rate (success rate of deduction) of M2 (the next month after purchase).
There are two types of churn:
Turn off withholding for third-party payments
Before doing any substantive analysis, I put forward the first hypothesis: the distribution of churn nodes should conform to the 28 principle.
So I asked the data students to do an analysis, and it was found that the closing withholding and surrender on the day of purchase accounted for more than 80% of the lost users.
At the same time, I put forward a second hypothesis: either there is a swipe, or there is a serious cognitive inconsistency before and after the sale.
I communicated with the channel BD, and first ruled out the possibility of brushing.
Then I experienced the product sales process of the channel in the whole process, and found that the sales page is relatively weak for the premium payable in the next month, and the payment success page is very clear and informs the user of the premium payable in the next month.
Then I put forward the third hypothesis: before the loss of users, they thought that the insurance was 1 yuan per month. After purchasing, they saw that the actual premium payable in the next month was much higher than the first month, so they lost it immediately.
Combining the assumptions, we negotiated with the channel to remove the next month's premium payable on the successful payment page, and made a series of matching strategies.