Why you should care about?
If you are a data scientist who has put ML model in production, you probably heard about app stickiness. The point is, anybody who is involved in product launching should want to do a better job at tracking our user behavior, especially in regard to user stickiness and retention. Try to see if you can prove your model can increase those metrics.
Daily Active User (DAU) / Monthly Active User (MAU)
DAU to MAU ratio measures the stickiness of your product - that is, how often people engage with your product. This metric became popular because Facebook uses it.
DAU - Number of unique users who engage with your product in a one day window
MAU - Number of unique users who engage with your product over a 30-day window
Pros & Cons of stickiness includes:
Particularly helpful for understanding how valuable your product is to users
Provides a snapshot of user retention. For early stage startups, this is a helpful metric for evaluating traction and potential revenue
Does not give you why users are being retained and which users are churning. Cohort retention analysis may be more useful in this case.
Is not useful for apps that are not daily: Not everything has to be daily use to be valuable. On the other side of the spectrum are products where the usage is episodic but each interaction is high value
Getting the user back to revisit. Retention and engagement is a separated concept as:
Google is a high retention, low engagement site
MySpace is a high retention, high engagement site
News sites are often medium/high retention, low engagement sites (like checking a headline)
Another important metric to view (that is different from retention)
Pageviews are coming ONLY from new users
Pageviews are coming ONLY from one generation of users (like early adopters)
Pageviews are coming ONLY from retained users
Pageviews are coming from new users and retained users
Total number of customers churned this time period / Total number of customers at the start of this time period * 100
The percentage of customers lost during a given period of time. For ecommerce, this means customers who fail to make a repeat purchase within an average timeframe for the business (could be 90 days, 120 days, or some other length of time. It can be calculated on a cohort basis instead of monthly
Pros and Cons include the followings:
Helps you see trends in product satisfaction (or dissatisfaction)
Analyzing customer churn based on cohorts can be particularly insightful for determining why or what other factors may be influencing customer decisions (e.g. pricing updates, new feature rollout, changes in messaging, etc.)
While customer churn is a helpful metric for detecting a ‘leaky bucket,’ it varies from revenue and doesn’t indicate which customers you’re losing (i.e. high-value customers, low-value customers, or perhaps customers who would be better served with another product)
It’s best to track this startup metric along with other key metrics such as Revenue Growth Rate, Net Promoter Score, and DAU/MAU Ratio.