e-news

Analytics in collections: making best use of all of the available data

In the current economic climate collections is definitely a growth industry. As unemployment rises so delinquencies follow and the pressure on the collections function grows more intense. Rising volumes of collections cases is causing many organisations to pursue mitigating actions: some are re-deploying new business underwriters into collections roles and others are being forced to recruit more collections staff.

While these measures are effective they build extra costs into the business. A complementary approach is to look to technology to help with this problem and this is where Analytics comes in.

The first question to ask is ‘are we making best use of all of the available data?’ Many organisations are not capitalising on the data available to the collections process either by not storing data that they gather or by not using the data that they store.  

By the time a customer gets into collections there is often a rich array of data available: demographic details, account and customer behaviour, customer information provided during inbound or outbound contacts and credit bureau data. These data paint a vivid picture of the behaviour of the customer and all can be used to assess the risk of further deterioration or the likelihood of cure. 

Scorecards are a good way of summarising this complex information into a single number. Together with their ability to predict the future, they provide an excellent means of answering the questions at the front of every collector’s mind: ‘Who first?’ and ‘Who next’?Scorecards provide an excellent means of answering the questions at the front of every collector’s mind: ‘Who first?’ and ‘Who next’?

The traditional cycle for collections starts when a customer fails to make a payment, but taking remedial action on accounts showing financial stress can help prevent delinquency occurring. Here, behavioural scorecards can be used in conjunction with an event that signifies an early warning of impending difficulties, to trigger pre-emptive action. These events, such as a loss of deposits or excesses over limit, form the basis for a justifiable interaction with the customer. Typically it works best where an organisation has a ‘complete’ view of a customer’s finances (e.g. in a retail bank where the primary banking relationship is held). 

This can be the trigger to reduce or remove credit limits, reducing the exposure to the customer by “turning off the tap”, or to restructure debts to give the customer manageable payments. This type of customer treatment can be difficult, as the customer has essentially done nothing wrong, so it is very important that the data and models behind the strategy are accurate. 

Even though a high proportion of accounts may go into arrears at some point during the life of the account, typically the position is recoverable and the bulk of these cases will quickly return to order. These account holders are still regarded as valued customers by the business, and through their correct identification, it can be a profitable strategy to take no action at all, and to leave them to self cure.

Payment performance data is undoubtedly a very strong predictor in identifying deteriorating delinquency and future payment propensity during the early stages of collections. However, when an account falls further into arrears (typically around 5-6 payments in arrears), the risk profile of the account will look very similar to other accounts, and it becomes much harder to distinguish between them using this type of information.  

Demographic and external credit bureau data (where available) can provide the means to accurately profile customers, in terms of their ability and willingness to pay. This helps businesses determine the best strategy which will give them the best return, such as whether to pursue the debt further, move it to an ‘internal’ Debt Collections Agency (DCA) so the client thinks that a third party is chasing the debt or to sell it to an external DCA for a fee or a % of the debt recovered. 

It seems likely that the economic climate will get worse before it gets better. The middle of a crisis is not usually the place to start planning for the future but, if your organisation has access to the data then scoring may be able to deliver significant benefits at a much lower cost than employing another 20 people in collections.  If you do not have data stored then credit bureau data, if it is available, can represent a good first step in the process of increasing automation in the collections process.

Paul Russell
Director Analytical Solutions
Decision Analytics
Experian

Contact us for further discussions about this article