Conservative Financial Industry Progressing with Machine Learning

1550

Savvy marketers in the financial services industry are in the vanguard of early adopters using machine learning (ML) to streamline operations and optimize business outcomes.

A survey of 1,419 customers, including more than 150 in the financial services sector, conducted by MIT Technology Review Insights in association with Google, found that marketers in the industry are among the most progressive in deploying ML. The research shows that 41 percent of financial services marketers currently use ML. The survey data also indicates that ML adoption will continue to crest: another 30 percent of financial services marketers plan to deploy the technology within the year. 

This is encouraging but hardly surprising. Leading financial services marketers indicate their ML usage is fueled by early success. The survey data found that 66 percent of business leaders agree ML enables their teams to focus on strategic marketing.

“ML is having a profound and transformational impact across every function in financial services, and marketing is one of the areas leading the way,” says Ulku Rowe, technical director for financial services at Google Cloud and former CTO at JPMorgan Chase. “ML is helping financial services marketers to keep up with constantly evolving consumer behavior and to ensure that they get the best value out of every marketing dollar spent.”

Cutting-edge—but conservative

The financial services sector is an amalgamation of cutting-edge initiatives and conservative practices.

On the one hand, financial services marketers must be at the forefront of new technology deployments. They must deliver offerings that serve current and future customer needs. And they are under intense pressure to carry out accurate marketing campaigns and specialized promotions that help drive revenue and earnings, observes Christian Renaud, a research vice president at 451 Research. They are constantly tasked with beating their competitors and their own performance every fiscal quarter. Meeting such demands requires a comprehensive knowledge of the customer base and specific market segments.

ML can deliver deep insights into customer behavior by analyzing specific pieces of data. Renaud calls ML an “increasingly necessary tool” that can help financial services marketers sift through reams of data, determine what worked in past marketing campaigns, and make informed choices. “Marketers can also utilize ML in order to see which campaigns resonated with specific geographic and demographic customer segments.” They can predict future industry trends and customer buying patterns, he says, pinpointing what time of year is best to launch a specific marketing promotion, for example.

At the same time, the financial services industry is by necessity among the most conservative because it must strictly adhere to a wide array of complex compliance regulations. Strong security is also necessary to safeguard corporate and consumer customer data assets and mitigate security risks to the institutions. ML can bolster security and mitigate risk, Rowe says.

“Customers expect financial institutions to keep their money and information secure. They expect the bad actors to be kept at bay. They expect compliance with regulations,” she says. “We see ML helping across all those dimensions.”

Renaud says the technology can be deployed throughout an institution to lift information security to a new level. When the need arises, ML can assist compliance and accounting teams with forensic analysis in following the money trail and spotting any anomalies. “This crucial capability of ML cannot be overstated, considering that security vulnerabilities and fraud are at an all-time high,” he says.

ML at work in financial services

The survey results indicate that 53 percent of marketers from the financial services industry say using ML allows their companies to gain a competitive advantage.

Financial services customers using Google Cloud Platform use ML to improve customer service, Rowe says, deploying chatbots and analyzing data collected across their organizations to create personalized offers. And they look to technologies like natural language processing for client onboarding and loan approvals.

“With reduced lending costs, and better credit-decision models, they are able to expand their reach and offer better products to more underserved clients in faster, more cost-efficient and engaging ways,” Rowe says.

AXA, a multinational insurance company based in Paris, uses ML to optimize pricing by predicting “large-loss” traffic accidents with 78 percent accuracy, notes Stephen Arthur, Google’s managing director for finance partnerships. Large-loss cases significantly affect the bottom line because they require payouts of more than $10,000. AXA’s research-and-development team in Japan used ML to predict whether a driver might cause a large-loss case while insured. The company analyzed 70 variables, including the driver’s age and address, the annual insurance premium, and the age of the car. AXA could then use the data derived from the ML analytics algorithm to predict losses and align its insurance premiums accordingly.

“The most compelling reason for ML and analytics in financial services marketing is the step change capability to deliver a superior, contextually relevant end-to-end experience that anticipates customer need based on signals of intent,” Arthur says. “The level of personalization” he adds, “is unsurpassed, and in many cases, not entirely intuitive to how we traditionally segment audiences.”

South African insurer Santam turned to machine learning to turbocharge its display advertising campaigns. Partnering with marketing firm iProspect, Santam launched Google’s Smart Display product, allowing marketers to automate ad generation across websites and mobile devices and match ads with customer profiles. The result was 75 percent more customer conversions than with traditional display ads.

Anticipating customer needs

In the survey, 60 percent of financial services marketers said they believe ML can capture intent throughout the entire customer journey. The ability to contextually examine connections and relationships enables marketers to analyze the customer journey, from initial contact with a brand to conversion and ongoing engagement. They can use ML to trace the first sales or marketing pitch to the actions that led to the buying decision and use the insights they glean to improve areas such as customer service and support.

“Life events are a critical trigger for opening new financial products and services,” says Arthur. That said, it’s often difficult for marketers to pick up on the signals of consumers experiencing life events in real time. “At Google, we’ve been able to use machine learning to leverage our data to understand and predict when someone is experiencing a significant life event and help marketers reach them more effectively during those moments,” Arthur says.

The survey shows that 44 percent of financial services marketers use customer value-based segmentation. ML helps marketers assess the value of their future customers, for example. Appropriately valuing prospects is “particularly compelling” to financial services companies, where understanding the varying lifetime value potential of what would traditionally appear to be similar prospects has huge implications, Arthur says.

“There is no reason to pay the same to acquire two people, one of whom funds a brokerage account with $10,000, while the other moves $500,000 into a managed investment service,” Arthur explains. “Each customer is valuable,” he adds, “but in very different ways, and knowing this intent or propensity to act effectively changes media investment.” And each customer can have wildly different needs, Rowe adds. She contrasted the needs of a millennial in the early stages of a career, living in a big city, with those of a baby boomer in the suburbs. “One may prefer a credit card that gives them higher points for restaurants and travel, whereas the other one may prefer cash back. One may be getting ready for starting a family, while the other may be more focused on leisure activitiies,” Rowe says.

Read the source post in MIT Technology Review.