Kotak Mahindra Bank Uses Analytics Insight to Revive Dormant Accounts

India’s home to a lot of big stuff. It has the world’s largest youth population. It has the record for the most bank accounts opened in a week. And it has one of the largest deposits of thorium anywhere on the planet. India is also home to over 40 percent of all the dormant bank accounts—accounts that haven’t seen a transaction for a specific amount of time–in the world. Of the 460 million adults with dormant accounts across the world, 195 million are found in India, according to the World Bank’s The Global Findex Database 2014. In April 2014, Kotak Mahindra Bank had about 0.4 million of those dormant accounts, an accumulation of over a few years. Dormant or not, those accounts cost money to maintain. It was a fair amount of dead weight to carry and the bank decided to figure a way to revive them. Talk about raising the dead. Deciding Who to ReviveReaching out to customers and asking them whether they wanted to revive their accounts isn’t the way all banks go. In 2011, for example, Punjab National Bank gave its customers a deadline, beyond which it would simply shut down all dormant accounts. It is one way to deal with dormant accounts—but a potentially wasteful one. Kotak Mahindra Bank knew that the cost of reviving an existing customer was much lower than going out and finding one.  “Roughly, the cost of going out and getting a new customer is 7x of reviving an existing account,” says Puneet Kapoor, Sr. EVP, Kotak Mahindra Bank, who is also in charge of the bank’s analytics wing. Also Read:Eveready Powers Sales with Visual AnalyticsBanking Big on IT: Arundhati Bhattacharya, SBIAshit Panjwani, SAS, on the Analytics Advantage for OrganizationsDetecting advanced threats with user behavior analytics  April of 2014, executives at Kotak Mahindra Bank asked themselves a hard question: “We had two choices: Shut down the accounts and not carry deadweight in our books, or weigh the opportunity to retrieve some value from these accounts,” says Kapoor. The bank decided to revive the accounts—but not all of them. It would have been wasteful to try to reach out to all four lakh account owners. Kotak Mahindra Bank would have to choose which accounts would be worth following up on—and that’s where analytics came in.  “Obviously, if I have to make an effort to reach out to a customer and revive that account, I need some insight on which people we should target and what we should say to them,” says Kapoor. The analytics department divided the four lakh dormant accounts into three buckets: Business accounts, salary accounts, and general saving accounts.  “We then started applying intelligence on these, in term of who we should reach out to,” says Kapoor. The analytics team first looked through general saving account customers, which constituted 2.6 lakh of the total four lakh dormant accounts, which were divided these into high variant and low variant accounts. Low variant accounts are those which required monthly minimum balances of Rs 10,000; high variant ones required either Rs 20,000 or Rs 50,000.  Then the team used analytics to figure out which of the high variant accounts, when they were active days, ever had an over Rs 1 lakh average monthly balance, and had managed that for any five months.  A second bucket of customers was created, but set the threshold at between Rs 50,000 and Rs 1 lakh. Further, a third group was also made, and set that filter at between Rs 25,000 to Rs 50,000.  “These filters are interesting. Why? Because they give you an insight into the wallet-size of the customer. Often customers bite off more than they can chew. When they open accounts they opt for products with a higher cut-off, and are then unable to service the account,” says Kapoor. This insight also gave the bank an important action point. It could now go back to customers and point out that they had probably opted for the wrong account type. By offering some account holders the choice to use a product of a lower variant, it could get them to restart their accounts.  “Instead of going elsewhere with a lower price point, they can do that with me. It’s a win-win for both parties,” Kapoor. Using analytics, the bank could hone their target list further using other clues including the pin code of the account holders’ home, or whether that person had taken a loan for an expensive car. That revealed that the customer could be offered an account with a higher monthly balance.  “If the pin code is Bombay 400006, which is Malabar Hill, you can conclude that the customer is wealthy. By applying more filters, you get your waterfall, which you can then get your channel to go after,” says Kapoor. The bank used a SAS analytics platform to run statistical procedures like correlation, association, segmentation, geo-clustering to arrive at the target segment. A Success by All AccountsThe group applied the same intelligence on low variant accounts as well. Out of the four lakh dormant accounts the group started with, their analysis revealed that 90,000 accounts were worth pursuing.  “The idea was to gauge the exact potential of the customer,” says Kapoor. These names were given to the bank’s channel to follow up with. At the time of this story, the channel had reached out to 36,000 customers and had revived thousands of accounts. Soon money started flowing back into those accounts and the bank managed to call back upwards of Rs 10 crore of stable float. What this means is that with these accounts are turning active, the accounts are seeing funds movement upward of Rs 10 crore – the float, part of which the bank in turn gets to use for onward lending. That’s called banking on analytics. 

If we have to make an effort to reach out to a customer and revive an account, we need some insight on which people we should target.