6 steps to quickly get started with growth hacking

Growth hacking is not a new concept in China. However, after reading many books and articles, I found that most of them are based on the different stages of Growth hacking, and there is little discussion on how to combine it with the actual work at hand. I have been chatting with some friends around me these days and found that many people are in the state of "knowing a little, being interested, but not knowing how to start".

In the first article of this series, I mentioned a similar point of view. The first half of the Internet experienced explosive growth accompanied by the demographic dividend, but in the second half, the extensive development model has made it difficult to replicate the previous growth rate, and many Internet companies have not yet entered the refined data-driven working mode.

A few years ago, I didn’t think Growth hacking had anything to do with me. Fortunately, my current team has been inheriting the Silicon Valley genes, believes in the power of data and work processes, and applies many Growth-related ideas and methods in daily work. And it turns out that sometimes just a decision, a parameter applied to a small requirement, can combine the Growth hacking methodology with actual work.

This is also the reason why I wanted to organize this series. I hope to use simple methods to discuss with everyone how to quickly get started with Growth hacking at work.

Before we delve into the specific strategies of Growth hacking, it will be more helpful for us to first understand the working steps of Growth hacking. The content shared in this article, on the one hand, is a synthesis of more than ten relevant domestic and foreign materials, and on the other hand, is based on the current situation of domestic Internet companies, plus my practice at work, and is combined into these 6 work steps.

I hope that through these 6 steps, I can help you have an overall understanding of the growth hacking workflow, and know how to apply it when talking about specific growth strategies in the future.

Step 1: Define your action targets

"Lean Data Analysis " mentions a concept called the first key indicator ( OMTM , One Metric That Matters), which refers to the number that is above everything else at the current stage and requires your full attention. As mentioned in the previous articles in this series, the difference between growth hackers and ordinary marketers lies in their focus on growth. This growth is not the growth of a single indicator such as page views/user volume/profit, but the continuous adjustment of growth goals as the stage changes. At this stage, it is sufficient for you to define only one OMTM.

This indicator needs to meet the following three conditions: ① accurate and clear; ② operational; ③ achievable. This requires you to break down the seemingly abstract word "growth" into smaller, achievable, and actionable tasks. For growth hackers, without a clear focus, subsequent work is meaningless.

So how do we disassemble OMTM to be focused enough?

Let me give you an example to illustrate my point.

When I first entered the industry, I created a community product from 0 to 1. So for the early stages of this product, the KPI is to increase the number of daily active users, or DAU (Daily Active User).

This seems to be fine, but upon closer inspection, this KPI is useless: there are too many factors that affect DAU, and it is impossible to determine what to do next to have the most direct impact on DAU. Later, the KPI was further refined to the user's next-day/7-day/30-day retention . However, like the DAU problem, this KPI also does not meet the three conditions of being accurate, clear, operational, and achievable. It was precisely because there was no clear OMTM in the early stage and the functional performance could not be quantified and evaluated that the project was aimlessly developing new functions for a period of time in an attempt to influence DAU or retention rate .

Looking back at this project now, I would set the KPI to the daily speaking user rate, that is, the number of people contributing content in the community on that day/the number of active users on that day. This is the core indicator that can most directly affect user DAU, and it is clear, operational and achievable enough. As it turns out, for community products, the depth of user participation is the key factor that directly affects user activity and stickiness.

Reversing the process of determining OMTM is actually a mindset of constantly breaking down and screening core indicators until they are broken down to a level that cannot be further subdivided. The subsequent levels will become your action plans. As shown in the following figure:

Step 2: Track and analyze action indicators

Once you have your action targets in place, you must track and analyze them. Tracking refers to obtaining the path data of user behavior, and what is usually needed is front-end data tracking. Of course, if it is not a project from 0 to 1, you just need to fill in the gaps.

Analysis helps you extract valuable data and correct your action indicators. Data analysis and action indicators influence each other. As mentioned earlier, you need to constantly adjust your growth goals as things change.

Let’s continue with the example of personal experience.

In this community product, assuming that OMTM has determined to increase the daily speaking user rate, let's see whether the current data tracking is in place:

After a new user opens the product, do you track the completion rate of the onboarding guide and the success rate of publishing the first piece of content through the onboarding guide?

When new content is generated, do you track the arrival rate, opening rate, and success rate of jumps to the corresponding pages of update reminders?

When doing user exposure, did you track the display rate and click rate of the area?

When a user's post is replied to and a notification is triggered, are the notification's arrival rate, opening rate, bounce rate, and re-reply rate tracked?

After getting this data, you may find that things are not quite what you imagined, and you are not quite sure how to solve the problem. It doesn’t matter, and no growth hacker can find the right tool to solve the problem at the beginning. All you have to do is recognize the facts and identify the problem first. This in itself is already a huge progress.

Step 3: Leverage existing strengths to find solutions

If you want to find a solution to a problem, don't let your mind wander. You need to focus on your own/team's current strengths. Generally speaking, you can look for your own advantages in these aspects to see if they can help with the current problem:

• Traffic advantage: Generally applicable to peripheral products of large companies that draw traffic from the parent product, or the founders themselves have a certain degree of fame and can bring traffic to the product;
• Resource advantages: The team has irreplaceable resources, which may be connections, channels , content, or the team members themselves;
• First-mover advantage: first to occupy the user's mind in a certain field;
• Capital advantage: sufficient funds to support you in achieving your goals or suppress your competitors;
• Cost advantage: If you can achieve the same goal at a lower cost than capital advantage, you can still achieve the goal;
• Technological advantage: technological capabilities that are ahead of the industry or competitors;
• Vertical advantage: In fact, it is another kind of resource advantage, but considering the second half of the Internet, vertical penetration into a traditional industry will become a very important advantage.

The scenarios that may arise in this step are rather complicated, so I will not elaborate on them here (for example, in the example I gave above, the advantage we used was exclusive content resources). You just need to understand the significance of starting with your advantages: on the one hand, you need to maximize your own advantages to solve the problem, which is the most likely way to successfully solve the problem; on the other hand, leveraging existing advantages can help you save time on preliminary work and get started quickly.

Step 4: Start the test

In the first three steps, we have completed defining the goal, locating the problem, and determining the solution. So step 4 is to start testing the solution. Before actually testing, you need to pay attention to the following issues:

1. Write down your hypothesis before testing

Writing a hypothesis is a simple thing, but in fact it is this simple step that helps you clarify the goal and significance of the test;

2. Don’t overthink testing

I know that many companies already have limited development resources and many demands are difficult to schedule. In this case, it is difficult to allocate some manpower for project testing. At this time, you need to obtain the support of your leadership, ensure that relevant colleagues know and understand the importance of this test project, and find ways to optimize the test plan to make it as simple, fast, and effective as possible;

3. Tests will often fail, accept that

Many times you will find that the test results cannot verify your hypothesis. This is normal. There is no shame in overturning the initial inference. You just need to change the method and keep trying until you find the solution that can help the product enter explosive growth.

4. Failure and success are both worth learning

Even if the data doesn’t confirm your initial hypothesis, the data collected throughout the testing process may help you make other decisions later.

Step 5: Testing

When developing a test plan, it is not as simple as just putting the plan online. In order to make the test as objective as possible, you need to use the following tools at the beginning of the test to improve the test itself:

1. Control Group

The test must have clear experimental and control groups, and the control and experimental groups need to be basically the same in terms of scope, characteristics, etc. Commonly used methods for selecting test groups and control groups: two groups of users from the same channel; two groups of users with the same statistical characteristics; two groups of users from the same cohort (the concept of the same cohort will be introduced and analyzed in the official account Pokeclub in the future); two groups of users with the same depth of product participation; and so on.

Without a control group, it is difficult to determine the exact reasons for the test results. Suppose you find that the DAU of the community increases significantly after you cold-start it with your superior content. It is difficult to know whether your content delivery strategy is effective or just because a group of people cursed at each other in your community for two days during the observation period.

2. A/B Testing

A/B testing is a commonly used testing tool for Growth hacking. It shows different plans to the control group and the experimental group at the same time. Other conditions of the two groups of plans remain unchanged, and only the impact of a certain attribute (such as the position of the picture, the color of the button, etc.) on the test users is examined.

A/B tests aren’t complicated to design, but the results can be hugely rewarding.

3. Multivariate Testing

Although A/B testing is simple and easy to use, it has high traffic requirements. Without huge traffic as support, the test results may be inaccurate due to the small sample size, or because there are too many elements to test, they can only be divided into multiple A/B tests, resulting in a long test cycle.

At this time, it is necessary to adopt a multivariate testing method, test multiple attributes at the same time, and use statistical methods to isolate the correlation between a single influencing factor and the improvement of a certain indicator in the result.

Step 6: Try again

When you have completed a full round of testing, you need to make adjustments based on the results and conduct another round of testing, or choose a new method for testing. The core of the Growth hacking workflow is to continuously test the minimum viable product, quickly test, quickly analyze and adjust, and quickly iterate, so as to continuously approach the goal or even correct the goal.

So when do we stop testing? There are two situations: one is that the hypothesis has been verified, then there is no need for further testing and the system can be launched directly in full to boost growth; the other is that it is found that better testing results cannot be provided under the existing conditions and resources, so it is better to free up manpower and material resources and move them to where they are more needed.

But generally speaking, it is difficult for growth hackers to stop testing. After all, the motto of growth hackers is to stick to the end and never give up :)

Mobile application product promotion service: APP promotion service Qinggua Media advertising

The author of this article @莔莔有神 is compiled and published by (Qinggua Media). Please indicate the author information and source when reprinting! Site Map