3 key reasons to why User retention is your most important metric

Cohort-Created-To-Signup-Problem_thumb

What’s Your most important Key Performance Indicator (KPI)? New users? Sales per day? Website visitors? I would argue that they are all secondary to User Retention, i.e. retaining users and make to come back after one day, one week and one month. Here’s why:

  1. User retention is the most accurate metrics to show that people like your product or service. As Nir Eyal puts it in his great book “Hooked”; You want to help your users develop a habit of using your product.  The higher retention you have in the long run, the happier your users are and the more likely they are to recommend your product to a friend.
  2. Forget about new users for a second. It’s nice to see growth, but it’s worthless without people coming back. If less than 30% of your users are coming back one week after signing up or less than 20% of them are coming back after one month, then stop spending money on marketing and social media posts. Your funnel is leaking and you need to fix it. These numbers are of course different if you are selling cars on a website, then if you are trying to get people on board a new social network, but it’s a good ballpark figure.
  3. Sales and user retention are closely correlated. The more people are getting involved with your brand and your product, the more likely they are to buy something. If you constantly get them to come back to experience something new, reward them for their progress and create a habit, your sales will come.

So how do you do it? Make sure to set up proper cohort analysis of daily new users and track their 2nd day, 2nd week and 2nd month retention. How you define retention is up to you. In it’s basic form it’s just people coming back and opening an app or visiting a webpage. You could also tie it to some key product actions like logins or posting some user generated content.

Google Analytics can’t do cohort analysis (as far as I know), so I recommend you use a service like Mixpanel or Localytics.