Home / General News / Webinar Recording: How Successful Agencies Build and Tweak Their Collections Model

Webinar Recording: How Successful Agencies Build and Tweak Their Collections Model

Collection agency executives already needed a law degree to help them navigate the tricky compliance waters of more rules and lawsuits. Now, those executives also need to be statisticians, able to apply regressive modeling and analysis when building models that help determine an individual’s likeliness of paying a debt. That is, unless agencies choose to outsource their account scoring efforts to a third-party company.

A trio of agency executives spoke during a webinar recently on the topic of account scoring and how to use it to build an effective collections model. The webinar was sponsored by Microbilt. A copy of the recording can be accessed here.

But scoring can only predict so much. The models aren’t yet sophisticated enough to take into account variables like the weather and the time of year. For example, more individuals are likely to pay off their delinquent utility bills as the temperature cools and they need to turn their heat on, said LaDonna Bohling, a vice president with Contract Callers, Inc. As well, Acer Capital Recovery works its higher-balance accounts during the early months of the year, when individuals are filing their income tax returns and might have more money in their pockets to use to pay off their debts, said Ed Saleh, the agency’s owner.

When applying an account score to a portfolio, Saleh says that he generates 80% of his liquidation rates from the top 20% of accounts, and those are the accounts that get the most attention.

Another variable, however, are the clients placing the accounts with the agencies. They may have their own ideas about how they want their accounts to be worked, regardless of whether it is effective or not. Sometimes, an agency needs to go back to a client and renegotiate an agreement so that it works for both sides, Bohling said.

“Show them the numbers,” she said during the webinar. “Show them how it should fit more in line with our work strategy.”

The larger the agency, the more sophisticated the scoring model appears to be a rule of thumb. At American Profit Recovery, for example, the agency uses both an external model from a third-party vendor as well as an internal scoring model, said Mike Hiller, the company’s vice president of collections.

“Our internal mode is better at predicting payments,”  Hiller said. “The vendor model is good for correlating.”

American Profit Recovery tweaks its internal model every quarter, Hiller said.

When assessing third-party vendors, each of the speakers urged their colleagues to test scoring models — against each other and against the expected results.

“We’ve tested some accounts with multiple vendors and the scores are very different,” Hiller said. “What matters more is the distribution of scores than the scores itself.”

For Bohling, what matters most is developing a strategy and sticking to it.

“You can score all the accounts you want, but if you don’t have an effective workflow strategy for what you’re going to do after the accounts are scored, then you’re going to learn an expensive lesson,” she said.

When it comes to scoring accounts, there are certain financial triggers that should send up the collection agency equivalent of the bat signal in that the account should be contacted immediately. Those triggers include events like new applications for bank cards, new auto loans or leases, or a new mortgage. Those triggers indicate that an individual has money to potentially make payments with, the speakers said. But, perhaps more importantly, those applications will also likely include updated contact information to help an agency get in touch with someone about paying their debts.

 

 

 

About mikegibb

Check Also

Large Banks Group To Form Separate Company To Vet Service Providers

Four of the nation’s largest financial services organizations have created a separate company that will …

Leave a Reply

Your email address will not be published.

Skip to toolbar