Top Growth Hackers: How to Use Data Experimentation to Build a Growth Engine

Growth hackers are the real source of revenue for a company. Business and other aspects are necessary, but the best people are those who really use their brains.

Products create value, and growth is about allowing more users to experience the core value of the product more conveniently and frequently. From this perspective, the purpose of products and growth is to benefit users. The process of realizing product value is itself a win-win process.

We should naturally pursue maximizing product growth - growth hacking.

Growth hackers in Silicon Valley have summarized their years of practical experience into eight efficient growth strategies:

1. Is your product PMF? Build high walls, store up food, and slowly become king

Here we need to introduce a concept PMF (Note: Praod uc t/Markert Fit). This concept is very basic and every product manager , every CEO, and every founder needs to understand it.

Let’s not talk about growth until we achieve PMF. It is a critical point. From zero to one is PMF, which means the market feedback obtained in the initial stage of the product. When you can see some data showing that your product does meet the needs of users or potential customers, you have reached PMF, and then you can start working on things.

Here I would like to share an experience with you:

If you are a 2C product, and you don’t do anything special, and you have a thousand unfamiliar users, you are almost a PMF; the same is true if you are a 2B business. If you have 10 unfamiliar customers willing to pay for you, you can do something.

We need to lay some solid foundations at the beginning : integrate AB testing and statistical tools . You can also put in things like user grouping, lifecycle maintenance, etc.

At this stage, the most critical thing is that there are some early indicators that you must pay attention to. You must use data to confirm that you have achieved PMF, and only then can you start to do things.

If you recommend your friends, each of them can get ten dollars. This will increase dozens of times. We all want to do something like this.

In fact, it is very difficult for such a thing to happen. What is so difficult about it?

The difficulty lies in that we have to discover this opportunity from a small place, so that this opportunity can become bigger. The small thing is that you have to give your first user a good experience, and maybe you can find a way to do big things from it.

You need to create a value proposition: it is very important that users can understand what you output and what you tell them.

Build a community so that your early customers can form a circle, and they can be of great help to you - of course you also need to make sure the data is correct and the infrastructure is in place.

In the past, the pain point of many software products was that it was difficult to collect accurate data, and they were unable to enjoy the benefits of the tools mentioned later, so they could only use traditional methods.

Therefore, it is particularly important to establish a good growth hacking habit from the early stages of the product.

2. Experimentation is the strategy itself - keep experimenting and keep growing

If you don’t do AB testing you’re dead, why?

Some people will definitely think that this statement is alarmist. Perhaps it is the inertia of the labor-intensive culture that always makes some people stubbornly believe that relying on brute force and scale is more efficient than technology even when they come to the Internet industry.

Before the birth of the Internet, David Mackenzie Ogilvy , known as the "Godfather of Advertising", once said:

Never stop testing, and your advertising will never stop improving.

Since the rise of the Internet, data can help us complete efficient experiments. Why do we have to understand customers through cold means?

In fact, as suppliers or merchants, we don’t understand consumer users. Users are really powerful. Users don’t care about you, they only care about themselves, and they don’t care about anything you do. Your attempts to understand users often backfire. We can only find ways to truly understand users' real thoughts and needs through data.

In this process, you must do something very painful - AB testing.

From this you can understand what he likes and dislikes.

Growth hacking must put users first, which is actually a mentality of listening to users and believing in users.

We can see the different feelings of the two products, product iterations or releases. One may be based on my own ideas; the other may be based on my understanding of my users. Iterate through experience and traditional means. Sometimes it will be very good, sometimes it may not be very good, and it may even stagnate.

If you run an AB test before launching your product every time, you will know first that the corrections I made can indeed bring an improvement or decrease to our user data. If it decreases, it will not be launched online, and it will be launched only if it increases. The latter is the product iteration and growth idea for the long term and future.

Sometimes we make mistakes in this matter. The first part requires us to change our habits, but the second part does not require us to suffer.

But sometimes we forget. Many people say that we do this when we make products. We look at the data to show what our shortcomings are and then we make changes. This is a very stupid thing to do.

You should first see where the data is most attractive and has the highest conversion rate . You see what users like the most and invest heavily in it; if users like this feature the most, you work crazy on it. If users are not interested in your product, it is not feasible to expect them to like it by optimizing it.

Sometimes you can talk to users about why they use you and what they really like.

If the user already has a habit and likes to use your product in this way, don't try to change him or make him understand the new process. This is actually stupid.

There is a particularly good channel . If you invest one dollar and generate 2 dollars, just invest it like crazy and don't think about anything else.

Of course you can take some time to test new approaches, and that's fine.

Data is a driver rather than a reference. Amazon's Bezos expressed this more thoroughly:

Experimentation is not one of the development strategies; experimentation is the strategy itself.

Sometimes when we work, we are still accustomed to the traditional model. I think that looking at competitors' products can bring me growth, and then I try to imitate them - this is actually quite inefficient.

The best thing is to use data to drive you. By looking at the existing data - whether it is your current situation or the results of your experiments, you can find out what points are worth investing in and betting on.

You can try it, run experiments, and get this thing rolling.

Within this wheel, including research, analysis, behavioral analysis, and hypothesis proposing, I can tell you that a particularly useful function is called personality assessment. After you propose a price hypothesis, you need to run tests. The data will tell you whether it is really effective.

Of course, running experiments is the most painful. Because you really need to take out some traffic , for example, 10% or 5% of the traffic to give it a try. This way you will definitely affect some users, and this is the only possibility to move forward, so you have to do something like this.

Experimentation culture, product + growth equals a good culture.

This sentence is very good. To form an experimental culture is actually to encourage everyone to innovate more, encourage innovation, encourage thinking, and use scientific methods to verify your innovative points.

This is a theory, and it is very simple to put it simply. Your current product or operation status is stable, it is version A, and you can make some improvements based on the existing data. I will try version B and version C, and compare the data online at the same time. This is a repetitive process.

If data is the fuel of growth, then experimentation is the engine.

This is very simple. It requires you to get the data from the very beginning and think about how to develop it based on the data. From the perspective of long-term use, this is not the case. The growth curve of Modao is like this (Note: Modao·1 year·500% | Building a self-growing product from 0 to 1).

What exactly is growth hacking?

The most important thing is your product.

Creating a product that people love may be the most important thing. The most important thing for your hacker is to make your product the most popular among users, with the highest conversion rate and the best data. When you do this, let your core users participate in it.

3. Experimentation = Strategy Itself – Experimentation and Growth

The 8 points mentioned above are summarized by growth hackers. You just need to remember them. Many of these 8 points are related to experiments and AB testing.

Alice is the founder of Growth Hacker, and of course he is also a very famous hacker joint venture.

He has also invested in successful companies like Facebook, and he put it very simply:

The most important thing for a growth hacker is to run experiments. Without running experiments, there will probably be no output, especially growth output.

Run experiments, and not just one, but ten, one hundred, or one thousand. The frequency and number of experiments directly determine the extent of growth.

There were no good testing tools before, so the growth rate was slow. In fact, for many unpopular products, the prices drop very quickly, and the curve is basically like this.

• In 2011, there was a very powerful AB testing tool. Before 2011, the experiments were implemented manually, and only one experiment was run every two weeks, which was also very normal.
• After 2011, we could run ten tests every week, and two tests per month became 40 per month. The growth rate suddenly became like this.

For example, the website Alice created is called Growth Hacker, and he has done similar things.

Around 2014 or 2015, people who wanted to be growth hackers would go and take a look at him. He later did experiments crazily, and he also started to grow crazily, running a lot of experiments.

Why is AB testing the most effective for growth? It has three characteristics. In principle, they are:

First, the risks are controllable.

To put it simply, our AB testing tool takes out a group of representative users and compares their data with your typical users to determine whether it is effective. To put it simply, this is the principle and idea.

If you use a small number of users to conduct experiments and the results are not good, the evaluation function we launched online will be harmful to the users, and it will only affect the selected users, not all one million users. You may only affect 10,000 users, so you can control the risks and conduct a large number of experiments. They won’t be like us who boldly launch something, drive away a group of users, and then win them back.

Second, it is parallel and efficient.

You can run ten or a hundred experiments at the same time, and many of them may be useless, but if you find the useful ones, your efficiency will be improved. It won’t take a week or a month to reach a conclusion through one user version iteration. The efficiency is much higher, and a hundred conclusions can be drawn in a week.

Third, the results of the experiment have very high cumulative value for you.

You may have a vague understanding of users and consumers at the beginning, but slowly you will have a very deep understanding of them, which is an improvement for yourself.

The typical growth model is AARRR . A funnel, a user's traffic comes in:

The first step is to attract new users . There are some new indicators for attracting new users, such as: download and installation volume, and click volume .

The second step is called activation. Your users may come, but they may not really become your users. These may be the conversions in the second step.

The third step is retention , for example: payment; he opens it the next day, opens it every month, and finally there is income; he is willing to click on your ads, and he can help you monetize in other ways; some of these users will definitely contribute commercial value to you.

The fourth step is word-of-mouth communication. Because of the funnel, you will find that as long as you do the last thing well, you will be very awesome.

The so-called growth is to improve the conversion rate of each layer of the funnel, for example: finding channels in the stage of attracting new customers, big Vs who have not yet been discovered on Weibo, advertising, content, email, and EDM marketing .

Activation is often the best first experience for your users, and it’s a very important point for the first conversion.

A lot of valuable work is done at this step, and long-term users will definitely find fault with you. He must think that you are not good here and there, and have a lot of ideas, so you need some more systematic methods to digest them.

Monetization is actually in payment or final commercial conversion. There are many points that can be optimized here, which are all very related to the user's consumption or conversion experience. Recommendation: How can you let your old and loyal users spread the word about you to more people? There are many areas to explore here.

The most important thing in growth is experimentation. Why? We can see that growth is simply a circle and a goal.

For example: Our goal is one of our KPIs. If it were me, I might set it as our profit - sales.

With this goal in mind, we analyze our user behavior and think about how to get them to pay more or place larger orders. Then we gain insights and conduct experiments based on our past data. After the experiments, we summarize the results.

You find that in this cycle, experiments directly affect your decision-making. You ran the experiment and found that the results were different from what you expected. Many of our users have encountered this situation. We need to sum up our experience and make improvements. You have to stop first. The direction is not right, or there is something wrong with our ideas. We need to change them.

If the experiment is effective, it is also a very important decision - you will find that the idea is correct, and you will think about whether you can do better, and spend a lot of time on the following things, and the circle will get better and better.

The author of this article is @哟呼科技. It is compiled and published by (Qinggua Media). Please indicate the author information and source when reprinting!

Product promotion services: APP promotion services, advertising platform, Longyou Games