KDnuggets Home » News » 2015 » Apr » Opinions, Interviews, Reports » Interview: Mario Vinasco, Facebook on Advancing Marketing Analytics through Rigorous Experimentation ( 15:n13 )

Interview: Mario Vinasco, Facebook on Advancing Marketing Analytics through Rigorous Experimentation


We discuss marketing analytics at Facebook, multi-channel performance assessment, success factors, lessons from Look Back feature, advice, and more.



AR: Q7. How and when did you get the motivation to write a book on Marketing Analytics? What is your favorite chapter in your book "Marketing Analytics - using predictive models in creative ways"? marketing-analytics-book

MV: I have been presenting at different conferences and events and the book was a way to document the case studies, techniques and learnings beyond presentation slides.

My favorite chapter is the case study where I used simulation; this date back to my college years.

AR: Q8. What trends do you foresee dominating the Marketing Analytics landscape over the next 2-3 years?

MV: At least at Facebook we will be doing more network experimentation, in other words experiments that take into account network effects; typically A/B test involve 2 independent customer segments, but this is usually not true when one group can share to the other. We will also start incorporating more of the machine learning models that are becoming very affordable and easy to run at scale.
network-experiments
 
AR: Q9. What is the best advice you have got in your career?

MV: To be grateful and humble and if the worst happen, have the confidence that I can always get up again.

AR: Q10. What do you think are the most effective ways for marketers to learn Big Data skills?

start-smallMV: Marketers can start by working with smaller data sets in MS Excel; this will allow them the opportunity to understand key concepts such as non-unique records (equivalent to count distincts), Vlookups equivalent to joins, and Pivot tables equivalent to window functions (more or less). Learning the basics of data management will take them very far.

Then, progression to real queries in Hive becomes easier.

AR: Q11. What key qualities do you look for when interviewing for Data Science related positions on your team?

interviewMV: We evaluate candidates across many dimensions; my particular focus is on technical skills and hypothesis formulation; a solid candidate needs to show strong SQL skills, ability to explain statistical confidence intervals and conceptualize distributions of events in social networks.

These answers usually prompt for follow ups, and we assess how well the candidate can ask clarifying questions, do estimates, and create different scenarios.

fooled-by-randomnessAR: Q12. Which book (or article) did you read recently and liked? What keeps you busy when you are away from work?

MV: I recently read again the book “fooled by randomness” which is always a brain teaser and fun to read.

Away from work, I play musical instruments, play futbol and do yard work at home.

Related:

Sign Up