Interview: Ankit Gupta, Vice President Citibank Australia


What has been your journey in Analytics so far?

My Analytics journey began in May 2006 with one of the biggest global banks. With the breadth of product suite and the geographical reach of the bank, I had the privilege of working across 13 different APAC geographies and for almost every consumer bank product that exists in the market. Each market had its unique features for the same product and yet experiences from one market could be modified to be made applicable to others. It has been close to 11 years now and I still learn every day at work. No single day is like the previous day because the problems we are solving are different, the methods to solve them are evolving as well. 

What kind of work do you do for your current organization?

At present I am working with the Australian market, mainly in the credit cards space. Majority of my work revolves around ensuring that all offers given to new customers of the bank are attractive, yet profitable to the bank. The current competition in the market makes that incredibly hard with a very small margin of error in getting things right. Apart from that, I look at the financial modelling for new business development and third party data to enrich the already existing vast consumer data with the bank.

What kind of technology/ platform/ software do you use?

At present SAS/ Teradata is the major technology platform that our financial data resides in. Apart from that we use SPSS/R from time to time for specific applications. But Excel/Access still remain as the backbone/final interface since the business stakeholders find it to be the easiest tool to get their answers from.

What do you think are the three keys to a successful career in analytics?

  1. Have a hypothesis but not a bias - We all get into solving a problem with some hypothesis in our minds. But as long as there are no strong biases in our minds, we will be able to go through without overlooking the information the data is providing. Once a bias sets in, an analyst only looks for data to prove that and ignores any contradictory results as noise. That could prove harmful
  2. Learn to speak the language of the audience - One mistake all analysts, including myself, have made in our analytics life is not being able to communicate what we wanted to, because we were not connected with the audience in terms of their knowledge of the subject. We tend to be so immersed in the data elements that we miss the bigger picture and therefore lose the audience's attention.
  3. Correlation is not causation - This, according to me, is the single largest pitfall in any analyst's life. We tend to confuse correlation to causation and make some very wrong recommendations based on that. One of the classic examples from Wikipedia is "The faster windmills are observed to rotate; the more wind is observed to be. Therefore, wind is caused by the rotation of windmills." So to have a great career in analytics, one needs to be wary of these pitfalls and steer away from them.

What kind of capabilities do you think an Analytics professional needs to build going forward?

According to me any analytics professional needs to have at least these 3 capabilities

  1. A way to talk to data - Data resides in datawarehouses/ datamarts and any tool/software that can be used to extract that into a meaningful format is a good starting point for the analytics journey
  2. A way to link that extracted data to the audience - Whether it is the old school Excel/Access or one of the new age tools like Qlikview/Tableau, an analyst needs to be able to get the information extracted from the datawarehouse to be made understandable to the end user in a format he/she is familiar in.
  3. Specialized fields - Every analyst needs to develop at least one area of specialization over a period of time. That skill/ area is something that will enable them to move from one sector to another, one geography to another. For e.g. if one was to specialize in the area of retaining customers, the skills/tools learnt in the process could be used for financial services industry or the telecom sector or even at a Casino :)