Utilize SQL (prefer PostgreSQL) and some python if necessary to complete.
Please be thorough and keep confidential. Thanks.
Questions
1. Marketing Attribution
• a. What campaign was responsible for each user’s finding our app?
• b. Which campaign was responsible for each sale?
Tips:
• Think about assumptions or data cleanup steps related to data inconsistencies, many to many mappings, or broken database schemas in the data. These may cause SQL joins to return incorrect results. Thinking about timestamps is useful here
• Please describe how you decided what attribution logic to use. How much time would you allow between the attribution and the user creation, and why ? For example, you can compare simple vs complex methods and discuss the pros and cons
2. Low Sales
• It looks like sales have been a bit low the last couple of days of the sales data set. Is this something we should be worried about?
Tips:
From a statistical perspective two things to lookout for are:
• Historical trends and seasonality
• Distribution of sales, statistical tests, time series techniques, conditional probabilities
From a business perspective:
• The link between sales, users, device and attribution data – can you find a potential root cause ?
• Comparison with the recent past and recent trends
Make sure you provide a clear conclusion and next steps (if any)