Above and Beyond Credit Bureau Data: A/B Test Leads to Fruitful Outcomes

Snapshot

When clients place bad debt accounts with us, we don’t simply take the information provided and immediately begin trying to collect on the account. Knowledge is key to making successful contact with the patient, which increases the probability of payment. But what information is needed to reach the right patients at the right time to increase the likelihood of a payment? It is more than basic contact information. Credit bureau data offers additional information such as other past due debts and credit lines the patient may have, but that information doesn’t offer the full picture of your patients’ ability to pay.

The Problem

While the information learned from the credit bureaus can help focus our calls, the credit bureau data provided is limited. They don’t allow access to how they collect and score the data, only delivering a score designed to predict the probability of payment. This is well and good, but here at Americollect we realized we could use our data from past calls to be more precise with future calls. As we saw it, there was much more data available internally to help create a patient a score of their ability to pay. The questions we encountered was how do we best use this additional data and how do we know if it makes a difference? Our data scientist was tasked with discovering the answer to these questions, and to see if this additional data would increase the likelihood of future payments when we made collections calls.

The Solution

The way to answer this problem came in the form of an A/B test. Our data scientist divided patients into two groups. First was the control group, which used the data provided by the credit bureau. Second was the treatment group, which utilized both the credit bureau data along with our custom score, which takes 100-plus data points not provided by the credit bureau. Within those two groups, we broke them down further into three subgroups that were divided by the number of calls they received. The patients with higher scores received more phone contacts, while the middle score group received less calls. The third group had the lowest scores and were considered least likely to pay so less calls were placed towards patients in this group.

The Results

After implementing the A/B test, we found that by using the credit bureau score along with our 100-plus data points, we saw a lift of over five-percent on the total amount of payments. This means that the payments made by the treatment group were higher than the control group. We were able to verify that this was a repeatable result, with a T-Test (a test to compare the average results of two groups) showing a less than four percent variation overall. If we were to apply the same predictive model to your collection efforts, there would be a nearly 96-percent chance that we would collect over $20,000 more when working 10,000 accounts on average.

Conclusion

For our clients, they saw an increase in the amount collection totals when we utilized our 100-plus data points along with the credit bureau data. Being able to take advantage of the knowledge of our data scientist led to this increase, which was then implemented across all collection activities. Our data scientist is constantly analyzing data and trends to ensure you are always receiving the most optimal method that boosts your revenue recovery. If you’re interested in learning how Americollect can do this for your facility, contact one of our Ridiculously Nice sales teammates today!