Clair is a digital banking platform that reaches hourly workers by being embedded in the software they use to clock-in and clock-out. Clair’s flagship offering is a free earned-wage advance product to grant workers early access to money they've already earned.
Workers sign up for Clair by tapping a “get paid early” button within the app they use to clock-in and clock-out. This opens an embedding onboarding experience in which a worker (1) creates a bank account, (2) switches their paycheck to the account, and (3) takes their first wage advance.
Event tracking has been implemented for each onboarding step. Events are captured in the events
table – each row represents a unique action taken by a user:
user_id
unique identifier for each userevent_name
specifies the action taken by the user; possible events are ordered below:start_onboarding
tracks when the user opens the onboarding experienceverify_phone_number
tracks when the user enters the verification code to verify their phone numbercreate_account
tracks when the user has successfully created an accountattempt_direct_deposit_switch
tracks when the user attempts to move their paycheck to Clairswitch_direct_deposit
tracks when the user successfully moves their paycheck to Clairtake_wage_advance
tracks when the user takes a wage advance
event_timestamp
specifies when the event occurred
General information about each user is captured in the users
table:
user_id
unique identifier for each userpartner_name
identifies the app where the worker clocks in and out
These tables can be found in the data directory within this repo.
Using the tools of your choice, analyze the data to identify where drop-off is occurring during onboarding. The Product team intends to prioritize the feature backlog based on your recommendations. There is no “right” answer; the emphasis is creativity and clearly communicating your approach and findings. Please limit time spent on the assessment to 2 hours.
- 2-4 slides describing your approach, findings, and recommendations; include visuals to support your conclusions
- A copy of the underlying analysis (e.g., Tableau workbook, Jupyter Notebook, etc.)
Note: While this dataset is structured to imitate real user activity, the event names, timestamps, user identifiers, and partner names are synthetically generated for privacy and security reasons. Given the randomness infused into the exercise, this is not an effort to outsource work to interviewing candidates, but rather, an opportunity to share a flavor of the data you’ll interact with working at Clair.