Predictive analytics can be used to compare properties' levels of synthetic-fraud risk, among other factors.
Predictive analytics can be used to compare properties' levels of synthetic-fraud risk, among other factors.

Aside from the monetary losses involved, evictions can often be tough emotionally on property managers, apartment communities, and renters alike. Evicted residents, obviously, will be unhappy, and every veteran property manager can tell a story about an evicted tenant who retaliated by vandalizing the property or who became a nuisance to neighboring renters.

To better minimize evictions, property managers need better insight and information to help them identify renters who are the best match for their properties. Fortunately, big data–enabled custom analytics are now available to help property managers better understand their renter portfolios, even as they change over time.

Big data analytics that provide insight and intelligence about existing renters in addition to new applicants can help protect property managers and mitigate their risk.

Deep Data and Analytics
Property managers have traditionally done up-front applicant background checks yet have lacked access to deep data and predictive-analysis tools. Using both can help significantly reduce risk.

The process of assessing tenants using a basic background check and references provides only a limited view of the applicant. Credit-based data, conversely, can provide a score for an individual tenant that estimates how much of an eviction risk he or she carries. These data can also identify prospects who are potential diamonds in the rough, meaning they may appear unqualified at first but may actually turn out to be great residents.

Big data offer two kinds of opportunities to optimize renter portfolios: descriptive analyses and predictive analyses.

Descriptive data look at a rent roll and determine who’s late with their payments and whether there are any problem rental units. Big data can be deployed to compare eviction rates at different properties, or even to evaluate a website to see where applicants are quitting in the rental process.

Prescriptive analysis uses a combination of data to form models and predict future behaviors. Fraud and evictions are two of the biggest problems big data analytics can help solve. With the data at hand, predictive models can be created that make actionable recommendations regarding new applicants by evaluating criteria that may make them increased risks, such as a history of negative credit behavior weighed alongside other data points.

Case in point: A lower-end rental property could attract mostly people with spotty records and evictions. Property management might need to accept a larger amount of customers with question marks apparent in their past credit history. In this situation, going deeper than just a credit score is necessary in order to evaluate whether there’s a good risk versus a bad one.

If there’s a history of financial losses, the primary question is, can a property manager reduce bad debt to the tune of tens of thousands of dollars? Cost savings could be possible by moving from a rules-based system to one that uses models driven by big data. The rules-based approach was the traditional way to do things but, compared with big data, this method lacks a dynamic and in-depth means of making evaluations or predicting the future. Achieving just a 2.2% reduction in evictions could mean savings in the hundreds of thousands of dollars, if the rental operation is large enough.

Stephanie Brock, president of Dallas-based third-party property management firm US Residential, has personally experienced the value of big data at her company’s properties.

“Business intelligence can be particularly useful in identifying the causes and trends around bad debt,” says Brock. “Tools that compare year-over-year data allow you to see the big picture and dig deeper where needed.

“For example, we wanted to determine how many move-outs were due to financials, evictions, rent skipping, etcetera. Because we had data to compare, we were able to identify that people weren’t moving out because of those reasons, but because they were allowed to stay in their apartment longer. It was taking longer to get residents out of their unit, which made that bad debt number go up. Those two data pieces really can make a difference on our NOI.”

Stamping Out Fraud
In addition to the problem of evictions, there’s the risk that some renters arrive from day one with the intention of never paying. This often comes in the form of either identity theft or through the emerging trend of “synthetic fraud,” whereby a wholly fake identity has been created using a mix of existing and fictitious data elements and a newly minted credit history is used for an identity that doesn’t really exist. Using this scheme, the renter occupies a unit by entering with a false credit history and a fake name, never pays any rent, and then disappears to pursue their next victim after a few months of free rent.

To catch fraudsters early in the application process, you can measure certain risk factors to obtain a risk score. Identity verification tools, with cues from big data sources, can identify potentially fraudulent applicants who require additional review, especially considering they may be only making contact online rather than in person. A higher risk could be due to the origin of their location, the demographic data they provide, or even the device they use to make contact. If an applicant turns out to be highly risky, they can be asked for further proof of their identity.

By using big data to add critical intelligence to renter portfolios, property managers can decrease expenses associated with evictions and add favorable tenants to their revenue stream. With the use of big data, renter portfolio modeling can effectively help identify the right renter while also minimizing exposure to risk from both fraud and evictions, helping reduce the painful process of eviction.