Search
  • Vikrant Mehra

Approaching problems with data

Something I noticed at college, which doesn't fly at work. People avoid numbers like the plague.


No one's asking you to run regressions or find the degree of correlation for every set of numbers you come across. At least, not immediately.


Instead, start and get comfortable with the basic graphs and measures of central tendency.


Good old mean, median and mode.


They're the first tool in your toolbox when confronted with ANY kind of problem. And while they may not give you the full story, they'll definitely tell you where to look for answers.


Case in point - one problem we tackled at work was understanding people who added their bank details, but never actually deposited money.


Looking at the numbers showed that most users waited exactly 30 minutes before making a deposit.


The same 30 minutes that a bank will make you wait as they verify a new payee's details before you can transfer money.


We realized people would forget to come back and actually complete their deposit, so we added a reminder. A simple notification later and problem solved!

2 views0 comments

Recent Posts

See All

My biggest takeaway from Shobith John and Yohan Sudheer's branding workshop yesterday is something I remind myself about constantly as I work: "Branding is always about the audience. Your customers, t

I came across an interesting paper on Bitcoin investing while looking through my Notion today. It's a 2019 study, looking deeper into the differences in perspective between male and female investors.