4 user growth models: from simple to sophisticated!

The simplest growth model AARRR model:

Advantages: Easy to build, can quickly display all factors that affect the North Star indicator

Disadvantages: It is a qualitative model, so it is relatively rough. It does not explain the numerical impact of each factor on North Star Growth, nor does it mark the relationship between each factor.

Usage scenario: It can be used as the first version of the growth model to help the team have a general understanding of growth factors and to compare different products.

02

Semi-sophisticated growth model: full-chain funnel type and factorization type

Advantages: It can not only decompose the factors that affect growth, but also find the corresponding segmented indicators and values, and use simplified formulas to express their relationship, so as to find opportunities and perform simple calculations.

Disadvantages: Compared with the AARRR model, it takes more time to build and requires more data, especially the factor decomposition model; compared with the quantitative model, the formula is simpler and cannot predict future trends or conduct hypothetical analysis;

Usage scenario: The amount of information is large and the construction is not complicated. It is recommended as a growth model for daily use

Q

Case: Full-chain funnel type – Google’s advertising revenue model

The full-chain funnel type is more of a multiplicative decomposition of our North Star indicator, breaking down all the factors of "click-through rate", "number of ad impressions", and "average cost per click" to further decompose which factors can affect these elements.

Q

Case: Factor decomposition type - profit growth model of a certain mutual financial lending product

Factor decomposition is more of a step further. We hope to use a formula to truly calculate our final North Star indicator. In the above profit model, the "bad debt" ratio and "lender interest" are difficult to control among these factors. They can affect the customer acquisition cost. Then we further decompose the customer acquisition cost into "re-borrowing customer acquisition cost" (the cost of borrowing again by old users) and "first-borrowing customer acquisition cost". The first-borrowing can be decomposed into the number of new users and the first-borrowing conversion rate, etc. We can make targeted operational improvements based on these segmented indicators.

03

The most sophisticated growth model: fully quantitative model

Decompose the factors that affect growth and the corresponding sub-indicators, and combine all indicators in Excel to calculate the North Star Indicator

Advantages: It can observe historical trends to predict the future value of the North Star indicator, and conduct hypothesis analysis to quantify the impact of changes in different indicators on the North Star indicator.

Disadvantages: It is laborious to assemble and maintain. It needs to be updated monthly.

Usage scenario: Suitable for teams that have a certain data foundation and need refined operations

Q

Case: Full quantitative model of an app

So this North Star indicator monthly active users is completely calculated through the previous two worksheets. How to calculate it?

Input Sheet 1: Monthly Number of New Users

Main variables: number of customers acquired through different channels, K factor, activation rate

Time dimension: based on historical situations and future estimates

There are 6 channels, each of which brings a certain total number of installations every month. One of the channels is user recommendation. A K factor is preset to represent the average number of new users brought by active old users every month. The larger the user base, the more people will come through recommendations. The total number of installations from the 6 channels is added together, and the installation-active activation conversion rate is preset to 90%, which can be used to obtain the number of new active users each month. It can be seen that January to May is our historical situation, and June to December is our estimated future situation.

Input Worksheet 2: Retention Rate

Based on historical averages, it can be adjusted according to actual conditions

Output worksheet: MAU (monthly active users) worksheet

Monthly new users * next month retention rate can be calculated for each month