Conversation with Ravi Santhanam, Chief Marketing Officer at HDFC Bank
00:07 | Debjani:
Ravi, you’ve had a very storied career to date. For our viewers, if you would take a minute or two to help us understand what you have accomplished, where you are, and what you’re working on.
00:22 | Ravi:
Sure, it’s a pleasure to be talking to you again, Debjani. After my masters in business administration, I was a big finance person. I used to work for a bank on the finance side — on the corporate finance side — and from there, I moved to a telecom, on a strategy route, and from there to a business school. That’s how it all started off, and big data analytics also came into my fold vis-à-vis part of the heading the data of business because we were actually doing a lot of CVM, customer value management.
What are your short and long term key initiatives?
00:51 | Debjani:
Today, you are heading up marketing at HDFC Bank, as CMO of the bank. Talk to us a little bit about what it means to be at the helm of marketing at one of the largest banks in India. What are your short-term and long-term imperatives, and what are the things you are looking at?
01:14 | Ravi:
In the HDFC bank’s case, the marketing is organized with an in-depth big data science team; the consumer analytics team is a part of the marketing organization. I have a thirty-five member data science team, which looks in-depth into the data that we have as a bank, and tries to predict consumer behavior, consumer journeys, and all of this stuff. That’s the first part of it. We have to spend a lot of time understanding the consumer using the data that we have. The data is available with us, the research team is also available with us, and the insights team is also available with us. Then the campaign still takes it and creates various cohorts of customers, various segments of customers, and then puts it into omnichannel marketing, as omnichannel campaign management, as we call it. We make sure, across our entire suite of channels for push and pull, a similar set of customers has consistent campaign offers.
Over the last three years, what we have achieved in the organization is heavy investing in marketing technology, to automate the way in which we reach out to our customers. Backed by data, backed by those consumer insights, and backed by a creative thought process which actually brings the human side of it and the machine side of it together, to actually go ahead and say, “What is the consumer problem that I can solve with this communication?” If there is no problem to be solved, don’t communicate. Because we are a service industry, and as a service industry, mobile is not a marketing device for us. Mobile is actually a fulfillment device for us because we can actually complete the transaction for the consumer in the mobile phone, in the computer. Today, for example, 90% of the transactions that HDFC Bank does are digital. So that is the way consumers are moving. The last bit is all about the acquisition, the mortgage transactions, and the purchase of a car, and all of that. That is the last bit of transactions that are not moving digital as of now, and we do expect that also to move into the digital world very, very quickly.
How are you injecting AI and ML into your day to day operations?
03:21 | Debjani:
That number is an amazing number — that 90% of the transactions with HDFC Bank are happening digitally. What are the next areas that you are looking at, in regards to: digital that you need to adopt and enable, how you are thinking about AI and ML, and how you are thinking of injecting those into the day-to-day digital operations?
03:51 | Ravi:
You have to get a lot of data, which we need to understand, unless I need to talk to the consumer before I take it to the consumer. After that, across all the channels, start the same thing. And make sure that the consumer is looking at it, feels that it is relevant for them, and once they apply, take it directly through digital, including disbursement of money and retainment instructions also stitched up. That’s the last piece that we are putting together. So four years, four and a half years back, we had started off this wherein, for a certain set of customers, because of the fact that they had been banking with us for a long period of time, we do know a lot more about those customers. We have actually given them, on a three-click basis, a certain amount of money that is pre-approved, so they can just come login to our mobile banking or our banking site, click on those buttons, and then take their money and go.
How do I extend that to most of my customers? And that’s the challenge, and that’s the opportunity also for us, and that’s the biggest thing that we do. The first and foremost thing for you is to figure out how do I reach, and how do I reach the relevant customer? And then I reach the relevant customer—what should we talk about, and how should we talk about those kinds of things to those customers? And then, what should be the nudge for those customers to actually consider, and from there, move onto the fulfillment journey? We have just got three buckets. There is a huge audience in the market. How should I convert that audience into traffic onto my website, into my properties? How do I convert that traffic into leads? And how do I convert the leads into fulfillment, so that final conversion happens?
3 steps to move a lead to a conversion
05:34 | Ravi:
We look at three buckets of what we have to do. Converting audience to traffic is primarily: understanding a huge amount of your existing customers, reaching out to them wherever they are, and also having a lot of understanding of who actually you will be able to provide money. You don’t ask someone to come into your home, and then shut the door on them. As a bank, you cannot invite a person to say, “Hey, I will give you money,” and afterward, you find out that the credit score is just not acceptable to you, and push them away. So you need to be extremely clear that the kind of people you attract to come out to your properties are the people that you will actually take through the journey. If you don’t, please stop, don’t talk to them. That requires a lot of understanding of data science. If you do not understand data science, you do not understand look-alike modeling and other stuff, how are you going to go ahead and do it? And not everybody goes through the credit process, not everybody goes through the fulfillment process. Then you know who has dropped off, and that’s the response data available. Are you going to use the response data to go back and redo your work, in terms of identifying the right set of customers? This looks very simple in English, and we call it machine learning. So that’s machine learning for us, and that’s how we keep churning.
Importance of data science for personalization
06:53 | Ravi:
Once you bring them onto your website, traffic has come onto your website. Now, there is no more a cohort of customers; the person who has come onto your website is a single individual consumer. Who has walked into my home? Debjani has walked into my home. Now, I need to actually become personalized to Debjani. If I’m not going to be personalized, in terms of understanding and greeting her with whatever I know about her, and I don’t personalize, how will she actually go ahead and be a lead? If it is not going to talk to her, while I might have attracted her to come into my property, but when she comes into my property and feels, “This is not for me,” then we have lost the game. So the next phase of the entire data science work is to personalize the communication, personalize the messaging, personalize the landing page, personalize everything, to personalize the experience, once we have an outcome. You need to invest heavily in data and science, the data science, to get to that personalization.
Then the third part is that, finally we are managing to talk to them in a very personalized way. They get impressed, they understand it, and they are saying, “Now I want to go through the journey with you and complete the fulfillment process.” At this point of time, I need to guide them through the fulfillment process. So that’s what I see as a third step from a lead to a conversion. We need to understand every drop-off stage, and what could have been the reasons for the drop-offs to happen. And we need to understand how to do re-marketing for people who drop off and go away. We need to understand at what stage people have dropped off, so what could have been the reason, so where should I go back and talk to them, and till how long I should talk to them, which channel to use to talk to them, and how do I create an omnichannel experience with respect to whichever channel that they may have come in? Will you go and talk to my branch, will you go and talk to my call center, will you go and talk to mobile banking? Then we’ve got everything done, finally the money is there, when the customer actually converts. And how are we going to do all of this stuff? That’s why we call it a journey. How to design the consumer journeys, how can you personalize the consumer journeys, and how can you make sure that there is enough information that is available to you to make sure it is omnichannel?
How are you using automation on customer journeys?
09:01 | Debjani:
That brings two more questions to mind; I think our viewers would be very interested because you are sort of the gold standard in what you are doing here and accomplishing here. How large a data science team does this require, internally, for you to do this? And the second question is a little more philosophical in nature in the sense that, as you bring in automation on these consumer journeys, across those channels, how much automation is good, and how much are you retaining for human beings to do?
09:37 | Ravi:
I think Artificial Intelligence works only when there are human beings available. If you see it as a data point of 1’s and 0’s, I don’t think you are going to get there. So it’s a mix of both the things, so we need to be very clear about how you bring the human intelligence. And automation of a repeating task? I’m all for it, so that people do not waste time on what could have been done by the machines. So, wherever there is a value addition that is possible, don’t automate. Wherever there is a value addition that is not possible, and a machine can do a better job? Obviously, the crunching of data — machines will do a far better job; moving data — machines will do a far better job. So, those things cannot be done manually, so you need to do a lot more automation.
The Data Science team at HDFC Bank
10:16 | Ravi:
In terms of the Data Science team, we’ve been lucky at HDFC Bank that, from 2009, we started this Data Science team, slowly and steadily. Half of the people are working on the credit data science side, and half of the people are working on the consumer marketing data science side. Continuously keep challenging using hackathons, continuously challenging saying, “This is the problem that we all have as a bank; who is going to come and help us solve this problem?”
Predictive and Prescriptive analytics
10:43 | Ravi:
Tomorrow, what I see as going to happen — because tomorrow, for me, is all about journey analytics. It’s all about prescriptive analytics, in terms of figuring out what kind of a customer journey the person had before completing the transaction. So, how do I actually nudge the next set of customers in that journey part, rather than try to predict that this person will take this journey? I think that’s passive; that’s over and done now. So that predictive analytics are over and done. We are now going to move to prescriptive and journey analytics, and that requires a lot of difference in set — a different kind of response and understanding of journey parts. And these are new things that we are all learning on the fly, and you need to be open to bringing as much talent from the outside, as you have developed talent inside. And it needs to be a kind of mix and match, which will work.
11:32 | Debjani:
That’s awesome, and of course, we’ve been very much a part of your journey, and we’ve seen it from the inside. One of the things you talked about is the growth in the user set.
Mobile banking growth
11:44 | Debjani:
If it’s okay, I just wanted to give our viewers the level of growth your mobile banking app has gone through—like 100% growth year over year, over a period of time. Which is mind-boggling in itself: the adoption of digital technology, and the adoption from your consumer base of that adoption has been just wonderful to see. Where do you see the next set of innovation happening in the next two years?
What innovations do you see in the next 2 years?
12:14 | Ravi:
My personal choice would be: how can I make it more personal, for personalization? Because I see journeys and the nudging in the journeys, because of the fact that, very specifically, we have taken a conscious decision that we want a zero touch experience for our customers: completely end-to-end, starting from the time the customer thinks of our product, to the time the customer actually has the product. Because we are one of those industries that can actually deliver that experience. From a dataset of hundreds and thousands, how can I get to an individual consumer, and personalize their experience, and make it personal? I keep telling everybody, at the end of the day, the difference between the best salesman and the next best salesman, is all about the relationship they would have built with the customer. Product wise, knowledge wise, everybody will be almost similar. The best salesman is somebody who understands the customer, and has an emotional connection with the customer. That is the personal touch, and that’s the reason for a salesman to actually be a very good salesman.
We need to be a digital salesman, with a very physical presence. How are you going to build empathy—when you are going to get into that personal space, which is the second phase of how to convert traffic into a lead, into a final completion of the journey? That depends heavily on your ability to get that emotional connection. There are enough cues available to figure out: what is the mood of the customer? This is the time span; there are a lot of things outside of the company and inside of the company. How to bring all the stuff to make that experience so personal that they believe that they are actually talking to somebody in the bank, in the physical presence, and there is a friend in that site—in that property—who’s helping them complete the transaction? And that’s where I think I’ll see a lot more work being done.
14:12 | Debjani:
I love your answer, but the unexpected word there, Ravi, that you used was: how do you bring that empathy to the digital world? Which is quite a high bar, I think, but it’s awesome. Thank you very much for spending the time with me today, Ravi.
Predictive analytics are over and done. We are now going to move to prescriptive and journey analytics, and that requires a lot of differences in set—a different kind of response and understanding of journey parts.
— Ravi Santhanam, Chief Marketing Officer at HDFC Bank
You might also be interested in
Thought Leadership Interview
Sarah Stalnecker of New Balance: How to Deliver on Consumer Expectations Using Data
Thought Leadership Interview