early_churn_analysis

Analysis of Users Who Added Verses but Never Completed 1st Session (Represents about 40% of Churn)
March 27th, 2013

Get all users who have added verses but never completed a session

stuck = User.all.select { |u| u.has_started? && u.completed_sessions == 0 }

Divide them into users who have added less than 5, exactly 5, more than 5 verses

stuck_lt_5 = stuck.select { |u| u.memverses.count < 5 }

irb(main):014:0> [stuck.count, stuck_lt_5.count, stuck_eq_5.count, stuck_gt_5.count]
=> [5373, 2393, 751, 2229]

Hmmm ... doesn't really show anything

How many users actually started a session but didn't complete it?

irb(main):025:0> [attempted.count, attempted_lt_5.count, attempted_eq_5.count, attempted_gt_5.count]
=> [1626, 286, 293, 1047]

So of the 5373 users who added verses but never completed a session, 1626 of them actually started a session but never finished it. 

Conclusion: the majority of users never actually start a session.

Is this true for more recent users?

stuck = User.all.select { |u| u.has_started? && u.completed_sessions == 0 && u.created_at > 1.year.ago }

irb(main):027:0> stuck.count
=> 2917

irb(main):029:0> attempted.count
=> 863

Conclusion: 70% of people never start the first session.

irb(main):035:0> stuck_gt_5.count
=> 1868 (64% of 2917)
irb(main):040:0> stuck_lt_5.count
=> 1049 (36% of 2917)


irb(main):037:0> attempted_gt_5.count
=> 770 (41% of 1868)
irb(main):042:0> attempted_lt_5.count
=> 93


Since quickstart was added, though, 64% of the users who failed to complete their first session did actually add five or more verses.