All posts by Veronica Herrera

Many times prescreen filtering stops after risk selection but that’s just one small piece of the puzzle. Experian has new tools that can help you pick out the most profitable consumers based on your business objectives.     Think about it - if you’re looking for consumers who will be profitable bankcard customers, wouldn’t you like to know what their total annual plastic spend is? Bankcard users come in all shape and sizes each with their own behavior patterns and preferences…so knowing if your prospects are revolvers or transactors, likely to balance transfer, or if they are rate sensitive, is all vital information to choosing who you want as your next customer.     At Experian, we have a suite of the most advanced analytical tools such as TAPS (Total Annual Plastic Spend), Trend View Solutions, Balance Transfer Index, Estimated Interest Rate calculators, and others.  We can help you sift through your list to find the most profitable consumers.t. In my next installment I’ll look at next generation prospecting models - beyond response.

Every prospecting list needs to be filtered by your organizations specific credit risk threshold.  Whether you’re developing a campaign targeting super-prime, sub-prime, or consumers who fall somewhere in between, an effective credit risk model needs to do two things: 1) accurately represent a consumer’s risk level and 2) expand the scoreable population. The newly redeveloped VantageScore 3.0 does both. With VantageScore 3.0 you get a scoring model that’s calibrated to post-recession consumer behavior, as well the ability to score nearly 35 million additional consumers - consumers who are typically excluded from most marketing lists because they are invisible to older legacy models. Nearly a third of those newly-scoreable consumers are near-prime and prime. However, if your market is emerging to sub-prime consumers - you’ve found the mother-load! Delinquency isn’t the only risk to contend with. Bankruptcies can mean high losses for your organization at any risk level.  Traditional credit risk models are not calibrated to specifically look for behavior that predicts future bankruptcies. Experian's Bankruptcy PLUS filters out high bankruptcy risk from your list.  Using Bankruptcy PLUS you’re able to bring down your overall risk while removing as few people as possible.   My next post looks into ways to identify profitable consumers in your list.   For more see: Four steps to creating the ideal prospecting list.

At Experian, we frequently get asked by clients how they can get bigger mailing list that open new markets and reach more people. But bigger isn’t necessarily better, and it doesn’t always translate to a higher return on your marketing investment. Instead of just increasing volume, let’s consider a different, more focused approach - using the latest in analytic tools and scores.  This approach relies on effective pre-screening to create the ideal prospecting lists based on your business objective. We’ve identified four key steps to building a prescreen list of your ideal prospects: Optimize risk selection Find the most profitable consumers Target customers who need or want your products Design the right offer In the next post, Optimal Risk Selection,  I’ll dig deeper into each step and present some tools and scores that can help meet the objective of each.      

Gone are the days when validating scoring models was a thing you did when you got around to it. Besides that fact that the OCC wants models validated at least once a year, it’s just good business sense to make sure your tools are working as expected. At a minimum, the OCC wants back testing, stress testing, benchmarking and sensitivity analysis, but there is another aspect to validations that needs to be taken into consideration. Most lenders do not rely exclusively on a scoring model of their decisioning (or at least they shouldn’t). Whether it’s a dual score strategy or attribute overlay, additional underwriting criteria is often used to help refine and optimize decision strategies. However, those same overlays need to be incorporated into the model validation process so that the results are not misleading. VantageScore Solutions, LLC has just published a concise white paper offering excellent examples of how to make sure your overlay criteria are an integral part of the overall validation process, ensuring your effort here are yielding the right results. And while on the topic of model validation, next time I’ll review what to do when you have no idea what to test for. Stay tuned!

As a scoring manager, this question has always stumped me because there was never a clear answer. It simply meant less than prime – but how much less? What does the term actually mean? How do you quantify something so subjective? Do you assign it a credit score? Which one? There were definitely more questions than answers. But a new proposed ruling from the FDIC could change all that – at least when it comes to large bank pricing assessments. The proposed ruling does a couple of things to bring clarity to the murky waters of the subprime definition. First, it replaces the term “subprime” with “high-risk consumer loans”. Then they go one better: they quantify high-risk as having a 20% probability of default or higher. Finally, something we can calculate! The arbitrary 3-digit credit score that has been used in the past to define the line between prime and subprime has several flaws. First of all, if a subprime loan is defined as having any particular credit score, it has to be for a specific version of a specific model at a specific time. That’s because the default rates associated to any given score is relative to the model used to calculate it. There are hundreds of custom-build and generic scoring models in use by lenders today – does that single score represent the same level of risk to all of them? Absolutely not. And even if all risk models were calibrated exactly the same, just assigning credit risk a number has no real meaning over time. We all know what scores shift, that consumer credit behavior is not the same today as it was just 6 years ago. In 2006, if a score of X represented a 15% likelihood of default, that same score today could represent 20% or more. It is far better to align a definition of risk with its probability of default to begin with! While it only currently applies to the large bank pricing assessments with the FDIC, this proposed ruling is a great step in the right direction. As this new approach catches on, we may see it start to move into other polices and adopted by various organizations as they assess risk throughout the lending cycle.

Previously, we looked at the various ways a dual score strategy could help you focus in on an appropriate lending population. Find your mail-to population with a prospecting score on top of a risk score; locate the riskiest of all consumers by layering a bankruptcy score with your risk model. But other than multiple scores, what other tools can be used to improve credit scoring effectiveness? Credit attributes add additional layers of insight from a risk perspective. Not everyone who scores an 850 represent the same level of risk once you start interrogating their broader profile. How much total debt are they carrying? What is the nature of it - is it mortgage or mostly revolving? A credit score may not fully articulate a consumer as high risk, but if their debt obligations are high, they may represent a very different type of risk than from another consumer with the same 850 score.  Think of attribute overlays in terms of tuning the final score valuation of an individual consumer by making the credit profile more transparent, allowing a lender to see more than just the risk odds associated with the initial score. Attributes can also help you refine offers. A consumer may be right for you in terms of risk, but are you right for them? If they have 4 credit cards with $20K limits each, they’re likely going to toss your $5K card offer in the trash. Attributes can tell us these things, and more. For example, while a risk score can tell us what the risk of a consumer is within a set window, certain credit attributes can tell us something about the stability of that consumer to remain within that risk band. Recent trends in score migration – the change in a level of creditworthiness of a consumer subsequent to generation of a current credit score – can undermine the most conservative of risk management policies. At the height of the recession, VantageScore LLC studied the migration of scores across all risk bands and was able to identify certain financial management behaviors found within their credit files. These behaviors (signaling, credit footprint, and utility) assess the consumer’s likelihood of improving, significantly deteriorating, or maintaining a stable score over the next 12 months.  Knowing which subgroup of your low-risk population is deteriorating, or which high risk groups are improving, can help you make better decision today.

Last month, I wrote about seeking ways to ensure growth without increasing risk.  This month, I’ll present a few approaches that use multiple scores to give a more complete view into a consumer’s true profile. Let’s start with bankruptcy scores. You use a risk score to capture traditional risk, but bankruptcy behavior is significantly different from a consumer profile perspective. We’ve seen a tremendous amount of bankruptcy activity in the market. Despite the fact that filings were slightly lower than 2010 volume, bankruptcies remain a serious threat with over 1.3 million consumer filings in 2011; a number that is projected for 2012.  Factoring in a bankruptcy score over a traditional risk score, allows better visibility into consumers who may be “balance loading”, but not necessarily going delinquent, on their accounts. By looking at both aspects of risk, layering scores can identify consumers who may look good from a traditional credit score, but are poised to file bankruptcy. This way, a lender can keep their approval rates up and lower risk of overall dollar losses. Layering scores can be used in other areas of the customer life cycle as well. For example, as new lending starts to heat up in markets like Auto and Bankcard, adding a next generation response score to a risk score in your prospecting campaigns, can translate into a very clear definition of the population you want to target. By combining a prospecting score with a risk score to find credit worthy consumers who are most likely to open, you help mitigate the traditional inverse relationship between open rates and credit worthiness. Target the population that is worth your precious prospecting resources. Next time, we’ll look at other analytics that help complete our view of consumer risk. In the meantime, let me know what scoring topics are on your mind.

For as long as there have been loans, there has been credit risk and risk management. In the early days of US banking, the difficulty in assessing risk meant that lending was severely limited, and many people were effectively locked out of the lending system. Individual review of loans gave way to numerical scoring systems used to make more consistent credit decisions, which later evolved into the statistically derived models we know today. Use of credit scores is an essential part of almost every credit decision made today. But what is the next evolution of credit risk assessment? Does that current look at a single number tell all we need to know before extending credit? As shown in a recent score stability study, VantageScoreSM remains very predictive even in highly volatile cycles. While generic risk scores remain the most cost-effective, expedient and compliant method of assessing risk, this last economic cycle clearly shows a need for the addition of other metrics (including other generic scores) to more fully illuminate the inherent risk of an individual from every angle. We’ve seen financial institutions tightening their lending policies in response to recent market conditions, sometimes to the point of hampering growth. But what if there was an opportunity to relook at this strategy with additional analytics to ensure continued growth without increasing risk?  We'll plan to explore that further over the coming weeks, so stick with me.  And if there is a specific question or idea on your mind, leave a comment and we'll cover that too.