Where We Were

For decades, rental screening was built around a simple assumption: applicants have a job, and that job produces a pay stub or a W-2. Verify the document, check it against an income requirement, move on.

That assumption made sense when it was made. The American workforce was more homogeneous – more people held single, full-time, salaried positions with employers who issued standardized payroll documents. Screening systems were designed to match that reality, and for a long time, they did.

But the workforce those systems were built for doesn't exist anymore.

Where We Are

The way people earn income has diversified. The gig economy, remote work, and digital commerce created an entirely new landscape of how Americans get paid. Someone running a successful Etsy shop. A W-2 employee who drives for Uber on weekends. A software engineer who rents out their Tesla during the workday through Turo. A 58-year-old who took early retirement and lives off a trust and investment dividends. These aren't one off cases – they're the modern workforce. Income now comes in layers, and the layers don't look like a pay stub.

More people are renting – and renting longer. Homeownership is less affordable than at any point in modern history. People who would have bought a home at 28 are now renting at 35, 40, and beyond. That shift doesn't just add volume to the applicant pool – it changes its composition. When people rent deeper into their lives, the income types in your pipeline start to reflect that: Business Owner, Trust, Self-employment, Social Security, retirement income, veteran benefits. These aren't signs of a weaker applicant – they're signs of a broader renter population. That volume can be an advantage, but you need a system that’s capable of capturing it.

Two Dots has processed millions of rental applications. The data confirms what the labor market already tells us: only 58% of applicants have clean, employment-only income.

The other 42% either have no traditional employment at all or have employment layered with a second source – Social Security, gig income, various benefits, business revenue – that your system also needs to verify.

This is the modern American renter. A system that can only reliably process a clean W-2 pay stub can only evaluate 58% of the people applying.

What This Means for Your Operations

Traditional screening systems look at pay stubs, W-2s, or direct deposit data from a bank link. When an applicant doesn't fit that workflow – 42% of them – the system either kicks them to your team or denies them. Both paths create serious problems.

Problem 1: Your team becomes the screening system and creates costly delays. When automation can't handle an applicant's income, your teams are left manually collecting documents, calculating income from multiple sources, and trying to reach a decision. This takes days. Over 66% of renters apply to multiple properties simultaneously – if you're adding days of paperwork, the best applicants sign somewhere else.

Problem 2: You're denying people you shouldn't be. If your system can't process complex income, you're effectively denying 20–30% of applicants whose income is legitimate but doesn't fit your system's narrow verification methods. 

For example, maybe it's a consultant who left a corporate VP role to go independent – same clients, higher income, no pay stub. Or, it’s a single mom with a steady job and court-ordered child support as supplemental income – her paystubs verify 2x rent, but the letter from Ashland, Kentucky showing 4x rent with child support is “unverifiable”. Both applicants get kicked to manual review and are eventually denied – not because they don’t meet the requirements, but because many teams simply don't have the tools to make a sound decision.

Problem 3: That denial pattern is a legal problem. For the independent consultant, the single mom, or any applicant with “unverifiable” income – the default is denial. This default establishes a systematic denial pattern, triggering source-of-income discrimination laws – which now exist in at least 21 states, the District of Columbia, and over 100 municipalities, with more added every year. California, New York, Colorado, Illinois, Maryland, Virginia, and Washington all enacted statewide protections in just the last few years. In 2024, fair housing organizations received over 2,189 source-of-income discrimination complaints, and a single case against Parkchester Preservation Management settled for $1.05 million. 

Problem 4: This is exactly where fraud hides. The same income categories your system can't handle are the ones most likely to involve fraud – and that's not a coincidence.

Applicants who submit offer letters instead of pay stubs are nearly four times more likely to be committing fraud. Fraudsters choose these income categories precisely because traditional systems can't automatically verify them. An offer letter can be fabricated in minutes. Professional fraud services sell complete employer verification packages – fake references, live phone representatives, forged verification letters – for as little as $5. And because complex income applicants often have multiple sources, a fabricated component can be layered on top of a legitimate one, making it even harder to catch.

When your system punts these to a human reviewer, the fraud lands with someone who is incentivized to approve, not trained in fraud detection, and processing high volumes under time pressure. The cost of each fraudulent move-in – eviction, lost rent, unit damage, legal fees – runs $7,000 to $15,000.

Each of these problems costs you individually. Together, they compound.

The NOI math makes all of this concrete. Consider a 10,000-unit portfolio with roughly 50% annual turnover and $2,000 average monthly rent. At an 80% approval rate, you need about 6,250 applicants per year to fill 5,000 units. Drop the effective approval rate to 70% because your system can't handle complex income, and you now need over 7,100 applicants – nearly 900 more to source, screen, and process. The extra marketing spend, extended vacancy days, and additional processing labor add up to a $400,000–$600,000 annual NOI drag. At 50,000 units, that's $2–3 million – not from bad debt or maintenance, but from a verification system that can't keep up with who's actually applying.

That doesn't account for the downstream effect of filling units with weaker applicants because the qualified ones got tired of waiting.

The Applicant Experience No One Talks About

There's a cost that doesn't show up in any NOI model: what this process feels like.

Applicants with complex income experience something like this:

“We’re sorry, but we were unable to verify your income” - Sincerely [Xsystemsupport@donotreply.com or propertyname@managementcompany.com]

OR

Someone from the property emails asking for additional documents. You send them. They ask for more. Days pass. 

Applicant: Any updates? 

Team without the tools to help: … 

Best-case scenario: you applied to more than one property, and the one down the street approved you in a couple hours.

The Tradeoff That Shouldn't Exist

 Nearly every large operator is living in one of these two scenarios:

1. Don't accept complex income, and watch occupancy drop while risking discrimination lawsuits across an expanding patchwork of state and local laws. 
2. Accept it and route it to your teams, and they become the last line of defense against sophisticated fraud targeting the exact income categories they're least equipped to evaluate.

Two Dots was built to eliminate this tradeoff. 

Our AI screening agent, Eve, makes a decision on 100% of applications (including complex income types), is available 24/7, and recommendations carry the same legal weight as a credit score – as an FCRA-regulated Consumer Reporting Agency. Eve handles the deep employer research on offer letters, the multi-source income calculations, and the fraud detection in a single automated workflow – delivering fast approvals and a seamless experience for every applicant. 

Our partners like BH Management and Weidner saw approval rates go up, bad debt fall by 60%, and turn times drop by more than 6 days.

But regardless of how you solve it, the question is the same: if a plaintiff's attorney pulled your last 10,000 denials and filtered for income type, what pattern would they find?