One of the most common approaches to tracking customer groups over time is through the use of cohort analysis. The goal behind such analysis is to segment your customer base into time-based groups (e.g., cohorts based on when they became customers) and track how they spend, how loyal they are, and so on.
While there are numerous approaches to cohort analysis, the most common is to track average spend, revenue, and order size across cohorts. What is particularly beneficial about such an analysis is that every business has similar goals: if your business is growing, you are likely trying to increase your average spend over time, while also increasing total revenue.
One of the ways we help companies achieve their goals is through performing a cohort analysis that is broken down by both year and month. The benefit of such an approach is that you an see seasonal patterns at the monthly level, while also comparing annual numbers to the months that are most similar (e.g., easily view every metric for September across multiple years).
To illustrate this approach, see the cohort anlaysis below. This chart shows the number of customers that became customers in specific months and years for an engineering organization. Note that to maintain anonymity and data privacy, we index everything to 100, representing the customer base in January 2009.
Just with the chart above, we can already see that the number of new customers is growing over later years. Additionally, the company has greener cohorts near the start and end of the year, showing that the period between October and February is the best for customer acquisition work.
Now, let’s run a similar report on total revenue, which is shown below.
In this case, we see some parallels but also some differences. The total revenue generated by each cohort tends to be higher at the start and end of the year, as evidenced by the green coloring. However, the total revenue generated in the middle of the year tends to be relatively poor, even though the first chart shows that the number of new customers is increasing in those months, relative to earlier years.
Already, a basic analysis tells an interesting story: customers obtained between October and February tend to perform well, every year. However, even as the company has been getting new customers across all months of the year, many of them have not been generating as much revenue. Relative to earlier years, mid-year revenue is dropping.
Such an analysis is quick and extremely helpful, and can likely yield great results for your business. If you have any interesting experiences, let us know!
|Next Post:||Case Study: using images in emails to encourage clicks|
|Previous Post:||Buying self-serve analytics tools? Make sure they’re “end-to-end”, are scriptable, and provide support|
Click here to subscribe to Canopy Labs Insights!