Compare Cohorts View: Life-month VS. Calendar Month

• Cohorts are sub-groups of Customers based on the month they became active. Analyzing Cohorts provides rich insights that can transform your understanding of Customer behavior and its impact on Revenue. With Cohort Analysis you can delve into granular specifics while gaining a panoramic view of overall business performance.
  
  
• Cohorts Analysis can provide a visual representation of trends to help you identify what's working and what's not. These insights help connect the effect of decisions and their outcomes, giving you valuable opportunities to make rapid adjustments and spur innovation. This lends support of long-term growth by increasing new and existing customers as well as enhancing impact on revenue trends.
  
  
• We can evaluate Cohorts by Calendar Month but stacking them by Life-month allows you to layer them so we can see how Customer behavior tracks over time and the impact it has on revenue as Customers are retained.

Cohort Analysis is a technique that's widely used by private equity, venture capital and software companies to analyze customer behavior and financial performance of a business when evaluating investment prospects. 

Despite how critical it is, there aren't any university classes, online courses, or public resources available to master this powerful technique. Instead, knowledge about Cohorts Analysis remains primarily enclosed in small circles of Silicon Valley investment funds and tech firms while it makes its way through expanding global tech and the venture ecosystem.

We, at Headline, believe that knowledge should be democratized to facilitate accelerated innovation in an ecosystem which ultimately yields outsized benefits to all participants, including the entity that shares the knowledge. As a result, we publish a monthly series that reveals the fundamental techniques of Cohort Analysis to empower business leaders and investors by more effectively analyzing business performance and gleaning the powerful insights.

When evaluating Cohorts we use a few approaches. We'll start by looking at Cohorts on a month-to-month basis. In this graph, columns represent the months during which the Cohorts became active. As a new Cohort is added each month, the Cohorts from previous months are carried forward as well:

The following graph shows the classic month-to-month stacking of the number of Customers in each Cohort by Calendar Month starting in July 2019 through December 2020:

Each row represents a Cohort of Customers that became active during a given month. For example, the top row represents customers that became active during July 2019 and the 15th row represents the customers who became active during the month of August 2020.

To illustrate how this works, below are specific examples from the above Figure:

- Cell B2 shows that there were 64 customers who became active during July, 2019. This sub-group of customers constitutes the July Cohort.

- Cell C2 shows that 43 customers of that first Cohort (i.e. July Cohort) remained active the following month (i.e. August 2019). This shows that of the initial 64 customers, 21 were lost during the first month (64 initial customers - 43 remaining customers).

- Cell C3 shows 151 new customers became active in August 2019 (i.e. August 2019 Cohort). As a result, the total number of active customers in August 2019 is 194 (43 remaining customers from the July Cohort plus 151 new customers in the August Cohort).

- Cell D3 shows that 100 customers of the August 2019 Cohort remained active the following month of September 2019.

As you can see, each month a new Cohort is created comprising the new Customers who became active during that month, plus existing Cohorts carried forward from previous months. As a result, the data on this graph forms a triangular shape.

We also analyze revenue generated by each Cohort. This is shown below in Figure 2 by Calendar Month starting in July 2019 through December 2020:

- In Cell C22, we see that the 43 customers in the July 2019 Cohort generated $4,868.03 in revenue during August 2019.

- In Cell S39, we see that the December 2020 Cohort generated $24,134.01 in revenue from those 349 new customers it acquired in December 2020. For the number of customers who remained active in this Cohort during December 2020 see cell S19 in Figure 1.  

The Cohorts in Figure 1 and 2 above are displayed by Calendar Month. Another way to evaluate Cohorts is lining them up by Life-month. 

This type of analysis provides valuable insights into Customer Behavior. Below are two examples to illustrate how Cohort Analysis can be used.

The following graph shows Cohorts by Life-month:

- Life-month 1 is the initial month all Cohorts became active (i.e. Column V)

- Life-month 2 shows the Cohorts when they are 30 days old (i.e. Column W)

- Life-month 13 is when the Cohorts are a year old (i.e. Column AH)

Below is a subsection of the above graph to illustrate the stacking of Cohorts to evaluate data by the length of time Customers have been active:

You'll notice that Life-month-to-Life-month analysis also has a triangular shape but it's inverted in comparison to the Calendar Month-to-Calendar Month analysis (i.e. Figure 1 and Figure 2).  Although the rows in this chart represent Cohorts, the columns are lined up by Life-month (not Calendar Month). 

July 2019 Cohort - Life-month 1: Cell V2 shows that in July 2019 there were 64 new customers who became active. These are the same 64 customers in Cell B2 of Figure 1.

July 2019 Cohort - Life-month 2: Cell W2 shows that 43 customers of the July Cohort remained active in the second Life-monts. These are the same 43 customers in Cell C2 of Figure 1.

August 2019 Cohort - Life-month 1:  Cell V3 of Figure 3 shows that 151 new customers became active in August 2019. These are the same 151 customers in Cell C3 in Figure 1.

We can also use the Life-month-to-Life-month approach when evaluating Revenue by Cohort as shown below.

July Cohort: Cell V22 shows the 64 new customers of this Cohort spent $6,722.79 during their first Life-month.

August Cohort: Cell W23 shows this Cohort spent $5,893.76 during their second Life-month.

Dissecting Cohorts by Life-month, provides insights into Customer Behavior based on the length of time they've been active, rather than looking at overall growth or decline of Revenue with no correlation to the length of time Customers have been active.

Cohort Analysis provides a granular way to illustrate what's happening with the number of active customers and how revenue is derived by month. This provides valuable insights that we can’t get when we look only at total numbers in aggregate since the company did not acquire all of its Customers and Revenue anew each month. Instead, these totals are actually functions of the new customers added during the current month, plus customers retained from previous months.

Layering Cohorts: We find that the best way to represent this impact is using the "Monthly Active Customer" and "Revenue-by-Cohort" Charts (Figures 5 and 6 below). These charts are often called “layer cakes" because each layer represents a Cohort.  The Cohort Layers show the number of customers acquired (or the amount of revenue received) during the month as well as the customers (or revenue) from that Cohort retained. 

Layering Cohorts in this way, differentiates the breakdown of Total Number of Active Customers or Revenue over time.

Cohort Analysis unlocks a key insight into behavioral trends. It allows us to understand how Customer Behavior may change as the business matures and/or new Cohorts are acquired over time. This approach also provides a comparison of behavioral trends of recent Cohorts with older Cohorts to determine whether Customer Behavior changes over time.  

Additionally, looking at Cohorts by Life-month allows us to normalize Customer Behavior and identify how Customers perform as they age with a business. 

By evaluating Cohorts in this way, we can effectively compare how many customers remained active in one Cohort from Life-month 1 to Life-month 2 compared to another Cohort from Life-month 1 to Life-month 2.  We'll get more in-depth into this in our next article.

As a result of these advantages, we use Cohorts when evaluating business performance because they are the most effective way to structure data to understand a Customer population and extract insights into Customer Behavior and determine how it impacts financial performance of the business. This will become clearer as we dive deeper into Cohort-based Analysis in our article regarding retention.