Watch a Senior Data Analyst Ace Google Case Study Interview Question: Ads' Revenue vs. User Searches

[@michelledo3898]

I'm sorry but how is this a good example interview? The interviewee went off track the whole time by proposing an irrelevant AB test that hardly does any good to the business and the interviewer ironically went along with this direction. Why bother doing that AB test when it's already stated in the question that revenue is up and user searches is down? On top of that she failed to clarify some critical parts of the case - context around "increased the # of ads shown after a subset of searches", why user searches went down (since this is not a metric diagnosis case the interviewer should at least give some details on whether it's a decrease in searches per user or user count, etc.), info on ad CTR, business goals/priorities, etc. Please give us more carefully selected content next time, thanks!!

[@CibhinBharathi]

Exactly!

[@kaustubhursekar9914]

This is exactly I was thinking. She could have simply ignored the first part of question and focussed “searches down” drilled the examiner on what searches? “Is it total users total searches counting multiple searches per user” this way, she could then jump to; - whether active users are down - is it that users are searching targeted products and session times is less - is there a disaster, internal bug, etc. keeping the searches number down. So many facets here which then can be tied to click thru rate and supporting RCA to why the ads revenue was up.

[@trez6465]

The interviewer is a senior Data Scientist from Google? she doesn’t seem to even understand what a null hypothesis is versus What is a data investigation hypothesis

[@AnuM-vw7ze]

Bad example ! The interviewee is not on track

[@in_ashish]

0:32 ehehe...!