“This is the 25th day of my participation in the First Challenge 2022. For details: First Challenge 2022”

There are five steps to implementing a growth strategy at work:

First, collect enough data

We need to collect as much behavioral data as possible. For example, what pages you entered, how long you stayed there, and what functions you clicked on. In addition to these common data, there are some business data that need to be buried. For example, how long it takes to open the page, how long it takes for the image to load, etc., these data provide data reference for our subsequent optimization.

In the early days of development when you are short of resources, you can use a third-party data platform, we just need to embed the SDK and add buried points. The disadvantage is that some flexibility is not enough, and some special business data still need to be realized by themselves.

Second, analyze the data

There is a saying that data are objective, but those who analyze them are subjective.

Macro data is relatively easy to analyze, which is to analyze the transformation effect between each step according to the business process. We can visually see where the business process has changed a lot. For example, in many products, users lose their apps when they open them. This may be because the front page of the app does not “activate” users to see the value of the product.

The more subjective analysis is based on some behavioral data of users. The important thing to note here is that users are very different from one another. For example, new registered users and existing users for many days, WiFi users and mobile data users, registered users and non-registered users, etc. Pay attention to the segmentation of users.

Three, put forward multiple hypotheses

After analyzing the data, we need to make some assumptions. Then propose solutions or validate solutions based on assumptions. At this time, as many hypotheses as possible, because many phenomena are caused by multiple results.

For example, it’s clear from the data that when a new user opens an app, a large percentage of users simply close the app. Therefore, we propose the following hypothesis:

  1. Do these users have anything in common? Same phone model, same location, same channel, etc. If so, the corresponding possible problem is compatibility, right? Telecommunication line problems? Is there a bug in the channel pack or is the channel itself grinding data?

  2. Does the content on the page reflect the value of the product? For example, your store focuses on cheap and good goods, and the results of the home page are very high to show the sale of high-priced goods, which is not in line with the value of the product to reflect. On the other hand, if you focus on high-end products, the presentation can’t be too low-end. The price should not be too cheap.

  3. Does the business process create obstacles for users, such as having to log in to view content?

The more hypotheses we put forward, the more likely we are to find the real problem. The important thing here is to make the logic clear when proposing a solution. You need to be clear about what type of users you are targeting, what specific things you are doing, and ultimately what data you have to show that your strategy is working. Don’t blur it.

Make a plan and prioritize

Having proposed different solutions based on various assumptions, we also need to prioritize those solutions. With limited resources, we need to focus on solving the core problem of cost-effective solutions first. We can draw a two-dimensional four-quadrant diagram of the required resources and expected output, and prioritize those with low cost and high output.

Five, online test

The last step is testing. We usually use small A/B tests to see the results. But if you have fewer users to begin with, full testing is fine. The key is to be able to compare the results of the new strategy against the data.

To summarize, the growth strategy is to collect user data and analyze that data. Then make some assumptions based on the relationship between the data, and then design some verification schemes or solutions based on the assumptions. The priority should be established according to the plan and put into online verification after the completion of development. Effective growth strategies grow out of this cycle of validation.