Retention or virus? What should a product focus on in its early stages?

Many companies ask the question “How do we acquire more customers?” But we should be asking “How do we better retain the customers we already have?”

When dealing with growth, it’s easy to assume that all you need to do is acquire more users. After all, that seems to be our definition of growth. However, if you take a step back and think of growth as maximizing the number of users per week over time, it quickly becomes clear that focusing on retention is more important than acquiring new users. It's more of a sustainable growth mentality. Rapid growth and rapid loss of users is an unsustainable growth indicator. High user retention over time is a good indicator of product-market fit, which is what you want to achieve anyway.

Viral factors and persistence

Retention is much more important than virality because if your users churn, they won’t be able to invite others for a while. If your product has high retention but no virality, you can sustainably grow your user base over time; conversely, if it only has virality but no retention, there will be no sustained growth in users. In between these two extremes, it seems a bit complicated. For a more detailed explanation, let’s first look at a few terms: viral factor and retention. This will give you a better understanding of the internal growth model in this article, and the charts we quote later are derived from it.

Viral factors

It describes the rate at which existing users of a site or app invite others and is often referred to as the k-factor.

• i = number of invitations sent by each client
• c = conversion rate of these invitations (c = number of registrations / number of invitations)
• k = i * c

On a weekly basis, the viral factor usually looks like this. It varies from product to product, but I see this shape over and over again in the products I work on. It tends to be this way for three reasons:

1. Validity of the login invitation process

Onboarding is one of the few points where you get a lot of attention from your users when it comes to completing a specific goal (user registration), and they generally don’t think much about it when you instruct them to invite other users.

2. User Excitement

Humans are most interested in new things, and this also applies to Internet products. The excitement users get when trying a new product leads them to share more, but this sharing fades as the product becomes commonplace in daily life.

3. Low invitation saturation of user networks

When a user first starts using a product, they know that there are more people who don’t use the product than there are people who do. Over time, they’ll share your product with people they know. Eventually, even a very enthusiastic user of your product will find that there is no one to continue sharing it because many people have not heard of it, causing the virality to decrease over time. This can be a problem if your company grows large, but it’s a good problem to have.

Retention

It refers to the number of users who remain sticky from one period of time to another. There are two ways to express retention, total retention and weekly retention:

1. Total retention

Total retention is cumulative over time. If your total retention rate in week 3 is 30%, that means 30% of your users stayed from week 1 to week 3. This is how retention is often expressed when discussed within a company.

2. Weekly retention

For growth purposes, it is often used to look at user retention on a weekly basis. Weekly retention refers to how many users stay from this week to next week. If there is a 40% total retention in the second week and only 30% total retention in the third week, then the weekly retention from the second week to the third week is 75%. If the weekly retention rate is lower than 100%, it means we are still losing users.

The weekly retention curve usually looks like the following, with the first week to the second week being the lowest and gradually approaching 100% over time.

Why is retention so important?

In order for virality to be more important than retention, your viral factor must be greater than your total retention at that time. I will use mathematics to prove this later, and the mathematics is difficult to simplify exactly, but there is a basic principle that you can follow and get close. If you take away only one point from this article, it would be this:

Unless your total retention is stable, don’t focus on improving virality until it stops decreasing after a reasonable period of time.

To help illustrate this, let's look at a few examples:

1. Your product has a very high instant viral factor. If your product front-end loads invitations and is accepted at a high rate, you can achieve an instant viral factor k > 1. In this case, if your invitation acceptance rate is fast enough, your monthly active users will continue to grow even if your website/app has zero retention after the first use. But since the exploitable k > 1 instant viral loop will eventually expire, high virality, low retention growth is almost always unsustainable. This explains the downfall of many companies that seemed to grow quickly, make a lot of money as they grew, and then die just as quickly. For example, Viddy...

1. Your total retention is high and slowly decreases over time, but you have strong virality. In this case, a long-term focus on viral growth can indeed benefit the product, but it is usually a false signal. The only time it’s worth focusing on virality is if you think you can get it higher than total retention. If you switch focus too early, it will result in sub-par growth metrics. This is because the compounding effect of improving retention is much stronger than that of improving viral performance. This can be easily illustrated using a retention/virality simulator, with weekly retention and virality set equal. We can then measure the impact of a change in proportion on one of them based on the number of users we have at some point in the future. In the real world, it seems unlikely that virality would equal weekly retention, but this setup helps illustrate our point in the clearest way possible.

Base case - Viral factor equals retention : In week 7 there are about 88k users, which can be obtained from the stacking of the bars of 44k retained users and 44k viral users.

20% increase in virality in the first week : 110k users in week 7, 53k retention and 57k virality

20% increase in virality in the first week : ~125k users in week 7, 65k retention and 60k virality

As you can see, changes in retention have a longer-term impact that is greater than an equivalent change in virality.

Your total retention is high and stable. If you’ve maximized retention to the point where you think you can increase virality rather than total retention, it makes sense to focus on virality even given the compounding impact of retention.

Prove it!

It's easy to model growth in Excel, and there are some shared help online, the one I used was downloaded from http://bit.ly/growthmodel. If you plug in some numbers it gives you a good overview of the growth you can expect in terms of user counts which is great but a little hard to conceptualize growth in actual work. To get a different perspective, building a treemap can be very helpful to see exactly where the users that exist during the week come from.

In the tree below, w0 represents the set of users starting at time 0. Each level of the tree represents a week, and at each subsequent level you’ll get users from retention or viral invitations. The number of users at a given node is the product of all nodes leading to that node. The coefficients represent the viral and retention factors over time, v2 is the viral coefficient for the second week, r3 is the retention coefficient for the third week, etc.

The number of users at any given level can be simplified to a recursive equation.

As you can see, this is consistent with what we saw in the dendrogram above. The main viral factor (current virus i) is related to the total retention (the product of n's).

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

This article was compiled and published by @哟呼科技 (Qinggua Media) when it is reprinted. Please indicate the author information and source!