The future of mortgage lead generation will be using big data techniques in order to better pinpoint which consumers are more likely to purchase a home or refinance their current loan.
One of the tenets of big data is synthesizing a large amount of information in order to make more intelligent decisions.
Using big data techniques could allow for the development of a "consumer score" to use in order to improve the results of lenders' marketing and lead conversion efforts, says Tim Savage, principal of Berkeley Research Group, a consulting firm headquartered in Emeryville, Calif.
The consumer score would measure the likelihood that a potential customer will buy a particular product. Right now, a consumer scoring model does not exist. "But there are data aggregators out there who are trying to pull together data on individuals from very disparate areas," he says.
This includes public data on social media sites as well as proprietary data that vendors sell. The aggregators are working on linking up these various data sources to create a data set that would allow them to do a better job modeling consumer behavior.
"It is an attempt to extend the credit score model to a broader issue regarding consumption," Savage explains. For the mortgage business, this means going one step back in the loan marketing process and pinpointing those people who are most likely to be purchasing a home or refinancing their existing mortgage.
For example, there are originators who send direct-mail materials to apartment renters in the spring months because that is the time when people typically start the home buying process.
Savage says that by using a consumer score, mortgage originators would be able to specifically target which apartment renters are more likely to actually start the home buying process.
"Instead of being a mass mailing, such as sending a flyer to everyone in an apartment complex, let's instead come up with a predictive model that says this person is more likely to buy this spring than someone else.
"Therefore, they can focus and target their marketing effort," Savage says. It might cost more to do this, but the conversion rate, as well as the return on investment, is likely to be higher as well.
Because the marketing piece is focused on the particular target, rather than just a generic postcard or flier, it has a higher cost to produce. But the consumer is more likely to respond to the offer.
While there are no technology barriers to the creation of a consumer score, it remains to be seen if this approach would have real value for mortgage industry participants, especially in terms of cost and ROI, he says.
Marketers are reluctant to change practices that they consider to have worked in the past. Yet the mortgage industry for years has complained about the very low conversion rate for their marketing efforts.
They are reluctant to invest in untried methods. "But the advantage of this approach is that it can be done in a scalable manner. You can start small, get a sense of whether it does yield any returns and if it doesn't, you can pull the plug," Savage states.
But even without a consumer score model, the big data resources are available for mortgage markets to better target their lead creation efforts.
Brian Fitzpatrick, president and CEO of Trevose, Pa.-based mortgage technology provider LoanLogics, notes the most amount of data already exists for consumer s who already have mortgages.
It is not just conversion rates of leads involving potential homebuyers that are very low. Retention rates among servicers, their existing customers, the lead they already have in-house, "are abysmal, they are single digits," he points out. This is true for both refinancings and for the purchase of another home.
"All of the data exists to tell when the consumer is likely to refi; and despite the fact that all of this data exists, lenders do a horrible job using that data to determine how to keep the customer," Fitzpatrick says.
So he is wary of those who are using data to try to attract new customers when it is much cheaper and effective to use existing sources of information to keep the customers that they already have.
For example, LoanLogic's LoanHD loan quality platform has data that servicers could extract to enhance their retention efforts. It can be used to see if the consumer is ready to refinance based on their current interest rate. The platform has information on other debts; this information can be used to determine if the consumer is a candidate for debt consolidation loan.
There is also data available that shows that consumers are loading up on other forms of debt; that could make them a poor candidate for another mortgage loan.
Even voice response units used as the first consumer contact with the lender are sources of data to be mined to find refi candidates, even if they aren't calling about getting a new loan.
Fitzpatrick has never understood that since this data is already available why the industry is not already using it better.
He answers his own question by noting that many servicers are reluctant to give up the immediate dollar made on the servicing fee. A new loan application removes that income stream for the time being. But that income goes away if the customer goes elsewhere for the refi.
"Big data can and should be used more effectively used by existing lenders to help solve the issue of low retention rates in the industry," Fitzpatrick declares.