Analyze these 4 factors of the viral growth model

What is Growth Hacking ?

This concept originated in the Internet industry in the United States. Its core is to use products or technical means to drive user growth based on data analysis. User growth is crucial for startups, so low-cost customer acquisition is what marketers are always looking for.

Fission is a typical means of acquiring customers with low cost and high growth. Achieving exponential growth in users through fission activities and generating a large number of new customers is what all companies dream of. But in reality, there are very few companies that can do this. Even if some companies do it by chance, it is just because they have the right time, place and people, which is difficult to replicate.

Of course, today we are not going to talk about the means of fission. We are mainly going to talk about a core formula behind fission - the viral growth formula, and its core indicator - the viral coefficient K-Factor.

This viral coefficient model comes from Adam Paineberg's "Viral Loop" (Zhejiang People's Publishing House, 1st edition). In the third chapter of "Viral Loop", the establishment of the viral marketing model is interpreted in a data-based way. In the tenth chapter, the adjustment of the viral coefficient, taking the Bebo social networking site as an example, the viral coefficient is used as an indicator of the website's user growth, and the impact on Bebo's development is analyzed.

We usually use the K value to represent the viral coefficient. To put it simply, the K value represents how many new users each current user can bring. If expressed using the most direct calculation formula, it is K=I*Conv.

I: Invitation, represents the number of invitations sent by each user (sharing rate); Conv: Conversion rate, refers to the success probability of each invitation (conversion rate).

Take a simple fission activity as an example. Suppose that after the activity is released, a user sends activity invitations to 10 friends, then I=10, and finally 5 friends accept the activity invitation, then the final Conv=5/10=50%, and the K value=I*Conv=10*50%=5.

Of course, nowadays, invitations may not be sent one-on-one, but invitation links may be shared through social networks or Moments.

Assume that a user’s circle of friends covers 1,000 users, and finally 7 users join the activity through invitations from the circle of friends, then I=1000, Conv=7/1000=0.7%, and K value=1000*0.7%=7. Generally speaking, this fission activity can only proceed when the K value is greater than 1.

Then let's go further and understand the virus propagation model:

• Custs(t): customers after time represents the total number of users after a period of time;
• Custs(0): initial customers represents the total number of users at the beginning of the activity (initial number of users);
• K：viral coefficient, also known as the virus coefficient;
• T: time refers to the total time the virus spreads;
• CT: cycle time: the time required for each infection cycle.

Let’s take a fission activity as an example. Around 2018, a new way of distributing courses emerged, which can be roughly understood as crowdsourcing distribution. KOLs are pulled into the same group, and they are asked to purchase courses at low or no cost. At the same time, more profit tiers are promised, so that they can promote the courses.

On the one hand, KOLs have greater influence and can attract more buyers; on the other hand, a bunch of KOLs posting posters for the same course is itself a good opportunity to promote the brand course (well, many of them are just collectively cutting leeks...).

We assume that there are 100 KOLs in the first batch, so Custs(0) in the formula means the initial number of users is 100.

Assuming that a KOL can bring in 5 buyers, then the K value is equal to 5, and the total duration of dissemination is generally either stopped by the activity organizer itself, or because the dissemination declines to a stop, or is banned.

Generally speaking, most of those activities that flood the friend circle with messages are banned by WeChat officials. So we assume that the value of T is 50 minutes. There can be a new round of dissemination in about 25 minutes, that is, the second group of people forward the distribution poster from the KOL’s circle of friends. Then CT=25, and the value of the formula we want comes out.

Theoretically, one round of dissemination takes 25 minutes, and three rounds of dissemination can be brought in 50 minutes (including the first round). Since each person can bring five users, one person can bring 31 people after 50 minutes, and 100 people can bring 3,100 people after 50 minutes.

Assuming that each person can bring 7 users, then one person can bring 57 people after 50 minutes, and 100 people can bring 5,700 people after 502 minutes. Assuming the time is extended to 100 minutes, it can bring five rounds of transmission, so one person can bring 781 yuan after 100 minutes, and 100 people can bring 78,100 people.

Each additional round of duration will bring an exponential burst (similarly, reducing the cycle time can also achieve the corresponding effect). Of course, this is just a theoretical valuation. In reality, it is difficult to achieve such data due to various factors. We will talk about this later.

From the above announcement and data, we can determine the impact of the three factors, K value, CT value, and T value, on a fission activity. We will not discuss the value of Custs(0) here, because there are not many techniques to expand the initial number of users.

The main considerations are the K value, the infection time of each round and the total duration. Because they are more operational and involve more skills, the most important thing for the operation of fission activities is to optimize these three factors - and this article focuses on the K value.

Let's use more specific numbers to compare and see the impact of changes in K values.

Assume that the initial number of users in an activity is 10, the propagation cycle is 25 minutes, and the total duration is set to six durations. Then substitute different K values ​​into the formula, and then draw a trend chart. Because the K value is too large, the result value is also too large, so the value above K>2 will not be displayed in the trend chart, otherwise the early trend chart will not show fluctuations.

Custs(t) change table

Custs(t) variation graph

The viral coefficient K means how many users an initial user can bring, and we can roughly see it from the above values ​​and charts.

When 0＜K＜1, as shown by K values ​​of 0.5 and 0.9: although the number of users is still increasing, this growth is very weak and belongs to sublinear growth. It is difficult for one user to completely bring in another user. If this is a fission growth activity, it is undoubtedly a failure.

When K=1, the user base is in a linear growth trend, neither hot nor cold, and the years are peaceful... One user can bring in a new user. Such growth cannot be considered a failure, but it cannot be said to be a viral fission either.

Frankly speaking, a company's normal growth should be linear growth, and only a few companies can achieve exponential growth. However, for fission activities, such growth is still a failure.

When K>1, the growth will show explosive growth, which we generally call superlinear growth or exponential growth. This is a successful viral fission, and one user can bring multiple users.

However, in this case, the ability of the operators will be tested even more, that is, how to make this successful fission controllable and sustainable.

Let’s use a few more numbers to derive the theoretical data: an activity with a K value of 10, an initial number of 100 users, and an infection cycle of 20 minutes.

From the above table we can see that, theoretically, at the 141st minute of this event, the number of users will reach 1.2 billion, which is the current total number of WeChat users.

If this activity were not restricted by the channels of transmission, the number of users would have reached 7 billion in 156 minutes, which is the current population of the earth; that is to say, two and a half hours plus six minutes, in less than three hours, all of humanity would have been infected.

Of course, we have always emphasized that these are theoretical values ​​under perfect conditions. Is it possible to achieve them in actual activities?

We can say without thinking that this is impossible to achieve. Not to mention spreading it globally, even spreading it throughout WeChat is impossible. WeChat banned this kind of crazy fission activity not because it was afraid that this fission activity would take over the entire WeChat, but just to ensure that the user experience of WeChat is not affected too much.

In real communication, it is often restricted by many factors, the most important of which are the following:

One is that not all users’ K values ​​can reach the K value of the initial users. In fact, each fission activity will look for KOLs or KOCs with a certain influence. However, for most users, both the number of radiated users and the number of users that can be converted are far from the level of the first batch of users.

A KOC basically has two to three thousand or even more than five thousand friends, while an ordinary user may have no more than 500 friends. This determines the base of the people they influence. Moreover, KOC friends are more related to industries and professions, while ordinary users’ friends are more likely to be relatives, friends and colleagues. The accuracy of the audience is not as good as the former.

At the same time, KOCs have been publishing professional content for a long time, which makes it easier for them to be recognized by ordinary users. Ordinary users usually post some updates about eating, drinking and having fun in their circle of friends. If they suddenly publish some fission activities, it will be difficult to get the recognition of other users.

Another thing is that not all users will be interested in this fission activity. All activities have their audiences. Just like our marketing teacher said, if someone tells you that the target customers of his product are everyone, you can just fail him.

In fact, most of the current screen-swiping activities do not go beyond the circle and disappear as soon as they go beyond the circle. The longer the campaign lasts, the more users who are not interested in it will be exposed, and the infection cycle will also become longer. The K value will become smaller and smaller, and finally the growth will approach a stop, and a fission activity will almost end.

During the National Day last year, the WeChat Moments was flooded with the message [Give me a national flag @微信官方]. More than 200 million WeChat users participated in this event, which ultimately brought millions of additional users to the event's APP.

The cost of this event is said to be just 200 yuan. The team sent a red envelope in a group and asked everyone to help forward it (but no one forwarded it after receiving it...); but even for an event with such a low entry threshold, in the end the number of participating users was less than 300 million, and the event was stopped.

One reason is that the people are not interested in this, or the elderly are unable to participate due to lack of mobility. So given all these limitations, what is the application value of the viral growth model?

I think the greatest value of the viral growth model is its application in activity planning, monitoring and review.

In the viral growth model, we can see several data, the initial number of users Custs(0), the viral coefficient K, the infection cycle CT, and the total duration of viral transmission T; and our operation of fission activities is to start from the four aspects of increasing the number of initial users, increasing their number of new users, streamlining and shortening the user's transmission cycle, and finding ways to extend the survival cycle of the activity.

The initial number of users of Custs(0) is equivalent to an initial traffic pool. The larger the traffic pool, the faster the volume will grow.

But in addition to the number of initial users, what is more important is the attributes of the initial users. The initial users are a group of people we carefully select, but for different fission plans, we also need to select different initial users. It does not mean that KOC/KOL must be better than ordinary users.

Indeed, as mentioned above, KOC/KOL has a wide user coverage and strong trust endorsement.

But at the same time, most of their friends are strongly related, and the overlap of the crowds is higher. In some activities that require breaking through circles to spread and have low participation barriers, it will be easier for people with more weakly linked users to penetrate into other circles. Therefore, before carrying out fission activities, it is necessary to plan the initial users.

The viral coefficient K has just been explained. It is a crucial factor. We can basically judge the success or failure of an activity based on the coefficient. When the coefficient is too low, you need to find a way to raise it, such as changing the hook.

However, don’t be too proud of how high the coefficient is, as you may be banned soon. You should control this value reasonably.

Judging from the activity of inviting friends to send courses or books, which has been popular since last year, three friends is a relatively safe number. It has a low threshold for user participation and is easy to form fission. At the same time, it will not exceed the official warning line.

If there are more than five people, the high threshold for participation will make it more difficult to spread the message, and the activity may be blocked, which poses a certain risk.

The infection cycle CT is actually an important factor in official monitoring. If the infection cycle is too short and the number of people participating in an activity within a certain period of time exceeds a certain threshold, it is easy to be banned.

But in fact, more often than not, operations may need to consider how to shorten this cycle. The main method is to find ways to streamline the conversion path of the activity. If the conversion path is too complicated, it will extend the entire CT cycle and block some users from the activity.

The simpler the path, the lower the user participation threshold and the higher the user's enthusiasm for participation.

The total duration of the activity is mainly controlled by the above three factors.

If the fission is unsuccessful, the activity will naturally die out and the total duration will naturally be short; if the fission is too successful and causes an official ban, the activity will end immediately and the total duration will not be long.

Therefore, only by controlling the above three factors within a reasonable range and making the entire activity reasonable and orderly can the total duration of the activity be guaranteed.

If the activity is too popular, you can reduce some of the self-sealed fission paths, replace weaker hooks to reduce the k value, add some paths (such as filling in personal information, queuing mechanism, etc.) to increase the CT time for control; if it is really uncontrollable, suspend the activity, so that at least you can ensure that the activity is not banned or the fans who have been successfully converted (fans to the official account) are not cleared by the official.

Of course, if the activity is too unpopular, you can optimize it step by step.

Finally, what I want to say is that if after an activity, the viral coefficient K>1, and the total number of Custs(t) users brought in is large, then this only means that the activity is successful, but it cannot guarantee how much value these users can create for you.

Especially in this era of red ocean market, user retention is greater than user acquisition to a large extent.

Of course, to ensure user retention, this is mainly related to the hook used during fission. An activity must be considered holistically. For example, the hook used during fission will not only cause changes in the K value, but also affect the attributes of the users generated by fission.

Therefore, the selection of hooks itself is a process of user screening. If you use cash as a hook, then what you screen out are general users with different attributes. And if you use operations courses as a hook, then the users you filter out will at least be those who are interested in operations.

The choice of hooks affects user attributes, and user attributes affect the difficulty of subsequent conversion and monetization.

Regarding this point, it is even more true for To B companies, because for To B companies, the number of users acquired has never been the most important thing, because the products of To B companies are generally not suitable for individuals, are difficult to develop, and even more difficult to screen and convert.

At the same time, the profit of one To B enterprise customer may be equal to the profit generated by thousands or tens of thousands of To C enterprise customers, so the quality of users will be much higher than the number of users.

Therefore, for ToB, user attributes determine the value that users can generate. If the user attributes are incorrect, the user may never come into contact with the products of the To B enterprise in his or her lifetime, and no value can be generated even if there are more fissions.

Therefore, if a To B enterprise wants to conduct a fission activity, it must carefully select the hook. Even if the final K value is less than 1, it is definitely cost-effective if it can acquire the target customers.

The significance of establishing a viral fission model is to use the model to calculate and judge, so as to adjust one's own operating behavior and improve one's own activities, thereby reducing the cost of trial and error and improving output results.

However, user growth has never existed independently of user operation. User retention and user life cycle should all be considered when doing user growth. The final payment made by users and the final value created are what affect the survival of the enterprise.

Author: Commoner

Source: Commoner