Taking the operation of WeChat public account as an example, we will teach you how to find user growth points!

Every business and operations team is pursuing user growth , but how exactly is growth achieved? Is there any underlying methodology that can support high-probability growth?

Today, after consulting with a user growth expert, I will introduce a methodology to you, teaching you how to find the most promising user growth points in the fog.

Core Methodology

The core of this methodology is: Test A-Result-Test A1.

I believe that when facing the growth of a product, there are many product operation strategies and promotion channels in front of you to choose from. They may not look good at first, but you must find a good way out by testing each method. The core of the methodology I am going to introduce is that you have to choose a relatively effective plan for test A, get effective conclusions from test A, and then make a decision for the next round of testing, either testing the A1 optimization point under plan A, or shelve or abandon plan A and start testing plan B.

It sounds simple, but in practice, few teams can really achieve this core goal.

The worst team is the one that can't think of any reasonable test plans ABC; the average team tests the ABC plans but cannot clearly know whether these plans are effective and how effective they are. As a result, the next test plan seems to have nothing to do with the previous test conclusion. It seems that they are very smart and have a big imagination and come up with another L plan. At this time, the effectiveness of your test is very low, and many attempted actions are a waste of time.

So, remember the core of the method below: conduct effective tests to obtain clear conclusions and provide decision support for the next test.

Principles of Growth

The principle of this methodology is: the judge is the data, not you.

Many team leaders make decisions basically based on their feelings. Especially when a project is going from 0 to 1, you will find that you have many strategies and many things to do. You just think that doing this or that "should be effective", but you never know how effective it is. Many teams even adopt a strategy without knowing whether the strategy is effective or not, and then they keep trying like a group of headless flies, never knowing what the specific results of the attempts are. As a result, many strategies that are obviously problematic are insisted on being implemented for a long time, which is really a waste of the lives of a team and a group of people.

Basically, all strategies, if you want to do them, you should assume that they are effective. If all strategies are based on the feelings of one person or some people, then your entire team is betting on whether this feeling is accurate. It’s not that it feels 100% unreliable, but the probability of success is relatively low compared to scientific data decision-making methods.

So, when you first think of a strategy, you can base it on experience and feeling. But when you start running tests, you must remember that all decisions should be supported by data and data estimates. When you think "it should be effective", you need to answer the following three questions:

• 1. Will this strategy improve your core data goals?
• 2. To what extent can it help improve this core data goal?
• 3. If this strategy is effective, what data shows its effectiveness? What data can prove that it is effective?

It is important to mention here that in actual implementation, you will find that many strategies cannot be measured by data at all, and many good and bad can only be judged based on feeling. Never fall into such self-deception. You must find a way to detect the data. If it cannot be detected, it means that this matter cannot be evaluated for the final project results. It is better not to do it. It will be in vain if you do it. In other words, since it cannot be verified, it may not be a key strategy.

The logic of this methodology is to find the key and use data to detect this key. If you think you can’t detect it, then you can’t implement the growth hacker methodology, so you’d better not read the following content.

Specific implementation method

The specific methods will be introduced below. For the sake of convenience, let’s take the operation of a public account as an example to explain how to implement this strategy.

The first step is to locate the core data goals of the project

There will be many small goals when doing a project, such as how much retention , how much growth, and how much revenue. If you have many goals, then the strategy will be unfocused and it will be easy to get stuck in the process. So the first step in executing this strategy is to clearly identify the only key goal that all the strategies you want to implement are going to serve? In other words, what are the key data that the project needs to break through at present? You must focus on only one goal. For example, in the operation of a public account , I assume that my current goal of operating a public account is to increase fans and reading volume. However, after careful evaluation, I find that increasing fans is actually a higher priority and more important factor at the moment. So I will clearly set increasing the number of fans as the core data goal.

Step 2: Disassemble key process data

Usually the core data goals of a project will be composed of several key process data. For example, the core data we just set is the increase in fans, so there may be several process data in this process, such as "number of followers" and "number of unfollowers". If you have decided that your main way of growth is to promote each other with other public accounts , there may be many key data in the process, such as "number of readings of mutual promotions", "number of public accounts promoted by each other", etc.

Step 3: Develop strategies for important data

After clarifying the above project core data and key process data, the team should then enter a full brainstorming phase. Give the team a limited time to think through what strategies and methods can be tried for these data, list them all out, and then think specifically about: How to increase the amount of mutual reading? How to reduce the number of unfollows? How to find other sources of growth?

Finally, the strategies that are considered more likely to have a huge impact on the core data goals are prioritized.

It must be emphasized here that when we are thinking of solutions, we will think of many solutions, many of which seem to be effective and must be implemented. For these solutions, you need to ask the three questions mentioned above:

• 1. Will this strategy improve your core data goals?
• 2. To what extent can it help improve this core data goal?
• 3. If this strategy is effective, what data shows its effectiveness? What data can prove that it is effective?

If the answers to these questions are all positive, you estimate that you can effectively improve the core data goals and have also figured out how to test its effectiveness. Then these are the strategies that you really need to implement.

For example, we believe that public accounts can increase followers by finding similar accounts to promote each other, placing GuangDianTong ads, doing interactive forwarding, and following hot topics to induce forwarding. These methods can all help improve core data goals. OK, then we will prioritize these solutions based on the estimated effectiveness.

Some other solutions may emerge in the process. For example, someone may suggest that the current public account avatar does not look advanced enough and a more advanced avatar should be made. Some people think this is useless. Instead of arguing whether it is useful, it is better to think about: what process data will be affected by the change of the avatar? Will it affect the attention conversion rate ? If so, how big is the impact? Then the debate might come to a conclusion and either be abandoned or a viable testing plan might emerge.

There may be many solutions like this: you think the instructions to follow the public account in the tweet are not conspicuous enough and need to be more conspicuous; you think leaving a message for a lucky draw can increase activity; you think forwarding the tweet to various groups at a regular interval can increase exposure and bring in more fans...

In short, there will be many strategies and methods in front of you, and all you have to do is ask yourself three questions.

Step 4: Prepare monitoring methods and expectations for the strategy

If you have already thought of all the strategies and decided which ones to implement next, then you need to estimate the data results of these strategies. Think clearly about which data can prove the effectiveness of the results, how much data you expect to be effective, how long is the expected time period for recovering data results, and when will you get the conclusion?

Let’s go back to the case of increasing followers of a public account. For example, we estimate that mutual promotion can increase followers by 3,000 per week. The process data may show that mutual promotion generates 50,000 external readings. If such data results are achieved within a week, we believe that this strategy is effective. The data is collected on Friday night, and the conclusion of the strategy test is obtained on Friday night.

Step 5: Implement testing

There is nothing much to say about this step. Just implement it and make it clear to the people. If this cannot be done, no matter how many methodologies there are, they will be useless.

Step 6: Compare the collected data results with the expected results

If a complete team has the role of BI ( data analyst ), BI should be involved in the entire process when formulating strategies and estimating data goals and recovery methods. At this time, BI needs to recover and organize all the established relevant data according to the previously determined recovery time and compare it with the expected data.

For example, through mutual promotion, our official account gained 2,300 followers in a week, and external readings generated more than 70,000.

Step 7: Decide whether to optimize or abandon

After getting the data results, it is time to test the team. The team needs to find clear conclusions from the data prompts.

The public account’s goal of increasing followers did not reach 3,000, but external readings reached over 70,000, which shows that this strategy is feasible and relatively effective. However, it is found that the conversion rate of increasing fans is still relatively low and lower than expected. So it depends on the team's decision whether to optimize the conversion rate or abandon this plan and find a more efficient plan. In this case, it is basically believed that A1 can be optimized based on this plan, that is, there are A1, A2, and A3 plans to optimize the conversion rate. If the 123 plans are all ineffective, then consider whether to abandon the big A plan.

According to this plan, you can record each strategy method, and then you can get the following table:

The fields of the form may include: implementation time, main executor, main behavior, main questions to be answered (what to test), core testing indicators and expectations, expected data recovery time point, conclusion time point, actual effect record, and effect analysis.

If you strictly follow this methodology, you will eliminate team confusion, you will have a clear grasp of the project focus and project rhythm, and you will be clear about the conclusions reached throughout the process, the successful experiences and the failed experiences gained. At the same time, team members will also be very clear about what everyone is doing now and what results they are trying to achieve, and there will be no misunderstandings between decision-making and execution. All of this provides an effective foundation for you to find opportunities in your constant exploration.

Author: Guangyu, authorized to publish by Qinggua Media .

Source: Guangyu Notes (yaoyaobiji)