We have received your feedback.Thank you for taking your time in view of your busy schedule.
The user growth logic of Facebook in the early days
The word "growth" has become very popular in recent years, which shows the anxiety of many entrepreneurs about traffic in the stock market. In fact, the concept of growth has not changed. Under the pressure of the situation, we have begun to explore whether there is a formed methodology, trying to find a "shortcut" to growth.
Let's take an example of how Facebook is making user growth.
How to set goals?
The first step is to set a goal. Setting goals seems simple, but actually setting them well is difficult. A poorly set goal = no goal. The core goals should be different for different products and businesses. But it must be simple and clear enough!
For example, at Facebook, there are three general segments: growth, core products and commercialization. The goals of the business units and the company as a whole can be clearly expressed in a simple mathematical formula.
The metrics for the core of different businesses are different. For example, WeChat and MiChat, which are also IM apps, are assumed to be opened by 10 million people every day, but I send 100 messages with WeChat and I only send 10 messages with MiChat every day. Using DAU can't measure the real level of the two APPs, it should be more about the amount of messages sent and received. And if WeChat's goal is set to DAU, then the team will take measures to improve the indicator, which may mislead the direction of the business.
Then there is the quantification of the target. To set a goal that is relatively easy to achieve. Suppose Facebook has 10 million DAUs now, then what will be the target for next year. Of course you can set a target of 15 million. But if the target is set too high or too low, it will not be effective as a guide. Usually the target will be set more clearly: for example, 14.5 million, of which natural growth is expected to be 2 million. New users with X% sub retention will contribute roughly 1 million DAUs, recall level will have roughly 500,000 DAUs of space... All added up later it would reach the 14.5 million target.
How to segment your goals
Abstracted goals are not instructive. Turning each goal into an executable project is the essence. For example, you've been assigned the project of improving new user retention. The proposition of improving new user retention is very vague, so let's make it real. To improve new user retention, we need to understand what attracts the user when he comes to the scene. A user who can stay must think the platform "useful" ("useful" includes something interesting). So your goal is to let the user discover the value of the product as soon as possible (the so-called aha moment). There are many ways to do this, and we offer a solution to the problem from the perspective of data analysis.
We divide users on the Facebook platform into two categories: one for active users and the other for inactive users. We tried to analyze which behaviors/features of users are more likely to contribute to user retention. There are some common behavioral characteristics of users as follows:
You will find a very significant difference between active and inactive users that is the characteristic of having contacts or not. Inside the active users, there are especially many users who have added contacts. But in the inactive users, most of the users did not add contacts. You seem to have found a point where you have made a hypothesis that making new users add contacts as many as possible will significantly increase their retention.
So now the proposition goes from "improve user retention" → "get users to add contacts as soon as possible". That seems to be a bit of a direction.
At this point, you have to keep asking yourself. The platform actually has the ability to add contacts, but it may not be doing a good enough job for new users, and it's not leaning enough resources. So how do you decide how many resources are needed? We have to quantify the revenue as much as possible. You have to continue to think: how many contacts do new users have to add to achieve a consistently high retention rate?
Let's draw a graph: (horizontal axis is the number of contacts, vertical axis is retention)
We will find that users will be able to achieve a relatively high retention by adding up to about 10 friends.
So the goal has changed from a rather vague goal that "improve new user retention" to a very clear goal that "make users to add 10 contacts in 7 days". Of course, you can continue to subdivide here. What I have outlined here is a relatively simple model, the actual analysis will be more complex and rigorous. The overall idea is to transform a rather imaginary goal into a small executable goal.
Operation and review
There is a saying in Silicon Valley: "If you can't measure it, you can't grow it". A good growth team is a team that can fight and resist after numerous experiments. Now the goal is there: "Get new users to add 10 contacts in 7 days". Then people will have a lot of ideas: is it to recommend friends in the registration process? Is it a resource place for new users to recommend more friends inside the information flow? Do you want to add more new users to the referral friend list? How to use PUSH/SMS methods, etc...
How to prioritize projects? How to make sure the quarterly goals are met? This is the time when the project leader brings all these ideas together and combine past experience and data analysis. The projects are divided into two categories: deterministic projects and exploratory projects. Deterministic projects are those that produce a definite benefit to the goal and generally account for more than half of the team's resources. Exploratory projects is more unknown, but once a new idea is uncovered, it can bring a steady stream of revenue for the next few quarters.
There is a set of efficient and practical AB experiment platform and data analysis tools at Facebook. For each experiment, people from all parties can quickly measure and evaluate the specific effect of this experiment from various dimensions. Each analysis strengthens the team members' understanding of the users and the product. Over time, the success rate of experiments increases dramatically. A lot of experience has been gained, so that everyone knows that the path is workable. This adds to the overall "decision tree" and creates a positive cycle of continuous efficiency improvement among team members.
In the above example, we found that the "recommend friends" module was a big boost to invite 10 friends in 7 days. We will use this as the pivot point and continue to fork out more pivot points, such as "giving more traffic in different portals", "recommending more friends to new users in the information flow", and so on.
The AB experiment was used to get more information. For example, in the branch of "recommend more friends to new users in the information flow", we will analyze the balance between traffic cost and advertising revenue; refine the different stages of the new user population; and make differentiated strategies for the cultural attributes of people in different regions. These become a small branch, until the water source of this branch is completely tapped out, we will continue to explore a new branch!
What does it mean to you?
"Growth" is more of a mindset than a replicable methodology. If you want to summarize it in one sentence: it is to improve the efficiency of the system iterative evolution. The word "growth" is attached to the living organism. Growth of a living organism is a systematic and collaborative process. And this process must be in accordance with certain laws.
This is an article from WeChat official accounts LouisX (ID: louisxuwei).