-
Amid the debate over whether or not banks should mine social media data for their credit-risk analyses, some financial services providers have quietly begun doing so.
February 1
How important is it to use expanded behavior-based data such as social media to vet borrowers? Considering
But embracing broader data sources, or using social media activity as part of vetting a borrower's personal financial responsibility, can require a difficult IT and strategic change for lenders. Internal systems have to be upgraded, new scoring analytics have to be migrated across departments, and, most importantly, bank users who aren't IT experts have to be able to use the new tools with minimal training.
"My example is if Facebook had been around 20 years ago, would we have been able to identify people who would have eventually walked away from their mortgage? As we account for the behavioral factors, banks have to figure out how to actually deliver better decision-making," says Rod Nelsestuen, senior research director, TowerGroup.
In an effort to allow lenders to upgrade scoring systems faster to accommodate different and expanded data sourcing, often called "big data" by tech firms, credit scoring and business intelligence providers that target financial institutions, such as Experian (EXPN), Equifax (EFX) and SAS, are develop new scoring and modeling systems. These systems work across departments, are designed to deploy and deliver their reports quickly, and include interfaces that are accessible to end users on the sales and marketing side of the business. "It's not about the models or a raw set of information …the institutions want to ensure that users that don't have to understand [the complex scoring techniques] to be able to use them," Nelsestuen says.
Experian last week showcased its new PowerCurve decision management suite at an internal conference in Scottsdale, Ariz. PowerCurve includes two software products — PowerCurve Strategy Management and PowerCurve Customer Management. Strategy Management is a credit risk product that allows staff to design, test, execute and continuously update decisioning strategies by adding new data and analytics. Customer Management is the marketing piece, providing a cross-organization view of a consumer's financial relationships that crunches credit data to quantify each customer's potential lifetime value, which is then matched to cross-sell and up-sell offers.
"What's changing is there is more data available, and our clients have to make decisions about how to put that data to work," says David Proctor, senior director of product management, Experian Decision Analytics.
In the traditional credit score expansion process, Proctor says that as new data gets put into models, lenders have to test how the new data impacts functions in different departments. He says underwriting and other staff have to learn how to use the new information to make decisions on credit worthiness, pricing and cross-selling. Proctor says these cycles are usually 12 to 18 months, though Experian is hoping to compress that time with the new products. "It takes time to understand the new data and what it means, whether it's loan data, demographic information, or portfolio data. So with all of this new data being used as part of scoring, how do we make that [12 to 18 month] time frame smaller?"
Experian believes the path to a shorter time frame will come from deploying new scoring and modeling changes directly to the risk department. Risk departments will use Experian's software to update and change the underlying systems that run the desktop work stations at different departments at the financial institution, so an expanded customer profile that includes deeper credit scoring can be accessed quickly through an interface that staff are accustomed to using, rather than having to install a new scoring system for each department. "Powercurve isn't designed to provide a tool for each department, but to make the IT job easier by helping them make changes by using software," Proctor says.
A Crowded Field
Experian's competitors include other scoring firms like Equifax and FICO (FICO), and business intelligence firms such as SAS, all of whom are also grappling with painlessly integrating social media and other new sources of data into existing workflows to add heft and agility to decision-making metrics.
Daryl Toor, an Equifax spokesperson, says the firm has updated its systems to include a debt monitoring solution that notifies lenders of new loan applicant activity during the pre-funding periods. To speed information delivery, it's also established a partnership with Interthinx that enables lenders to access The Work Number's automated solution for income and employment verification, which provides digital access to capacity to pay.
A FICO spokesperson said the firm was working on new credit scoring systems, but would not comment further. On the marketing side, FICO just acquired Entiera, which multi-channel campaign management delivered as a Software as a Service (Saas). Entiera Insight combines structured and unstructured customer data, and enables marketers to design and execute large-scale marketing campaigns across multiple channels. It will be combined with products that use FICO's predictive analytics marketing products such as FICO Precision Marketing Manager and FICO Retail Action Manager.
"Historically these kinds of [predictive modeling marketing] solutions have been customized and one-off with the client," says Sally Taylor-Schoff, vice president of product management for marketing products at FICO. "SaaS makes these applications more readily available. By using banks' ability to pull data together into a customer database, we can get clients using the systems much faster."
"We're beyond the days of being able to simply process credit card transactions, which really haven't changed in the past 20 years," Taylor-Schoff says. "As more marketers want to use other forms of data, they need flexibility …they need to be able to analyze data of any kind."
There are also regulatory pressures that encourage faster updating of scoring models. David Wallace, global financial services marketing manager for SAS, says the increased level of regulatory mandates has driven operational costs up significantly while at the same time revenue opportunities are challenged by the slow economic recovery. "One area in which technology can play a significant role is to better leverage silos of 'big data' that exist throughout business processes and use predictive analytics for rapid actions driven from fact-based decisions," Wallace says.
SAS' entry into the race includes SAS Credit Scoring for Banking and SAS Real Time Decision Manager. Scoring for Banking gathers and organizes data, develops, deploys and validates scorecards, performs data mining and model development, creates and deploys Basel II pooling, performs regulatory model validation per Basel II pooling and produces basis reports for risk managers. Decision Manager creates, tests and deploys credit decisioning in batch and real time mode for approval/decline, pricing, cross sell/upsell offers, next best action, credit limit management, authorizations and collections.
"As volume and variety of data increases, the integrated analytics architecture incorporates and uses the new data seamlessly, and then updates scorecards internally without sending data to external vendors. This results in faster time to market for scorecards," says Naeem Siddiqi, global product manager for banking analytics solutions at SAS.
To further speed updates and expansion of data used in scoring, SAS' High Performance Analytics helps lenders react quickly to changing market dynamics, and derive insights at faster speeds. SAS searches data sets with billions of rows to spot correlations that can accelerate the development of new predictive models. Wallace says these models are tested on full data sets instead of samples along with iterative exploration of all available techniques.
