Published by Mike Finegold, Chief Data Scientist, Fulcrum Analytics on June 15th, 2016
Many critical financial services functions have been historically dependent on high-touch interactions with experienced personnel. Retail bank wealth managers make product recommendations to their customers based on intuition, experience, and personal interaction. Investment banking salespeople are on the phone with multiple institutional investors every day trying to understand client interest and trade existing bank positions. Underwriters at large insurance companies go through a time-intensive process to obtain renewals, respond to RFQs, and make pricing decisions. These employee-dependent practices have permeated the financial services sales process for decades without much fundamental change.
But the rapidly changing technological environment is creating an enormous opportunity for forward-thinking companies to generate efficiencies out of these critical processes. At the same time, it presents a potentially existential threat to companies that cling to the status quo—some competitors are going to take advantage, and those that don’t adapt simply won’t be able to compete.
FinTech companies have made inroads into just about every sector of financial services through a combination of lean infrastructure and data-driven processes. Although their future success isn’t guaranteed, they continue to gain market share. But years of data on the transactions, assets, claims, and behavior of their clients give legacy institutions an inherent advantage. Traditional institutions can leverage this advantage by embracing modern data analytics and rethinking the sales process.
While big data may have been overhyped in some cases, in financial services the opportunity is clear. There are billions of dollars to be gained from making client interactions more effective by utilizing existing data assets. Enormous pressure throughout the industry to reduce costs provides the perfect opportunity to make each step of the sales process more productive. We know this because we’ve implemented solutions already and are developing others throughout insurance, retail banking and investment banking.
Consider the traditional workflow for a bond salesperson. She makes and receives phone calls throughout the day from institutional clients, gets IOIs and RFQs, checks client portfolios and bank positions, and interacts with her sales and trading colleagues. She takes notes on some of these calls and has a variety of software tools to give her up-to-date info on client and market activity. A combination of intuition, recent history, and internal demand lead her to choose which clients to call about which trading opportunities.
But with tens of thousands of bonds in the market and a dozen or more clients, how can she be sure that her limited time is spent on calls that are most likely to serve her clients? For larger clients, in addition to the bonds they trade regularly—which opportunities are they most likely to be interested in? For smaller clients, which ones are likely to be active in the market at a given time? On which days are they worth a phone call? Given her limited time, there is great opportunity cost to any wasted effort.
Modern machine learning methods can leverage the vast available data to help prioritize her sales calls. Her knowledge, combined with a prioritized list of opportunities updated in real-time, can ensure that she wastes as little time as possible—creating better client service and more efficient turnover of the bank’s portfolio. The successful salespeople will be the ones who harness these new tools and make themselves more productive.
Now consider a very different workflow at a large insurance carrier. Client interactions there involve multiple departments, but there is a similar fundamental challenge. With hundreds of clients up for renewal and thousands of RFQs to respond to, the underwriting team must decide where to focus their efforts. Gathering data to assess which policies to pursue is highly dependent on the work of other internal groups—pricing, transactional actuaries, claims, finance, and reserving—and while it is important to bring a competitive price to the broker, it is also important that the price accurately reflects the risk. As with bond sales, we need to identify which combination of client and action are most important to focus on at any given moment.
Which clients are at-risk when up for renewal? When and how should the company act to retain top clients? What early triggers can help identify priority areas? For RFQs, which ones are worth the time of the underwriter and which are better off being ignored altogether? In short, which activity is most likely to lead to incremental profit?
Access to data on millions of claims gives incumbents an enormous advantage—they are far better equipped to predict future risk, work with clients to reduce incidence of future claims, and mitigate costs of existing ones. Insurance companies can utilize big data to identify trends and large-scale loss patterns in real time. This kind of agility would allow an insurance company to change ahead of their competitors, deal with issues more strategically, and gain market advantage.
Big data analytics can help an insurance company provide value-added services to clients and identify trends to predict future profitability of clients and prospects. Combining internal portfolio data with market and competitor data allows for better identification of prospects that are likely to generate the most profit. Focusing underwriting efforts on these promising clients will lead to a healthier portfolio in the long run.
We have only scratched the surface here. Everywhere there is a client interaction, there is an opportunity to be more efficient through the effective use of big data and modern methods. In financial services, a more efficient process with a focus on the greatest profit potential can mean hundreds of millions from cost savings, increased sales, and more profitable client relationships.
For more information, contact us today.