Direct mail has always been one of the most effective channels for marketing credit products. Credit cards, personal loans, home equity lines, student refinancing — the mailbox has historically delivered some of the highest response rates in financial services marketing, and the numbers back that up. Credit card mail alone accounts for close to 20% of total direct mail volume in the U.S. Mortgage and loan products add another 15%.
But for lenders and financial services marketers, running sophisticated direct mail campaigns has come with a persistent and serious constraint: the Equal Credit Opportunity Act.
Today, Postie changes that.
We’re launching purpose-built, ECOA-compliant models — making Postie the first direct mail platform to offer machine-learning-powered audience targeting designed to operate within ECOA’s regulatory framework. For financial services marketers, this isn’t a feature update. It’s a capability that didn’t exist before.
Understanding the Compliance Problem
The Equal Credit Opportunity Act establishes the federal framework for extending and marketing credit to consumers in the United States. The law prohibits discrimination based on protected characteristics, including race, sex, age, national origin, religion, and marital status. That prohibition extends not just to credit decisions, but to how credit products are marketed in the first place.
This creates a meaningful challenge for direct mail targeting. Most lookalike and machine learning models used in marketing, including models on major digital platforms, incorporate features that either directly encode protected characteristics or derive from them. A model that uses age as a signal, or that learns patterns correlated with national origin, cannot be used to market credit products without creating legal exposure.
The compliance reality is stark: lenders insisting on ECOA compliance have had almost no machine learning targeting options. Major digital platforms have not been able to represent that their lookalike models are ECOA-compliant. So these marketers have been left with manual feature selection — not targeting, just selection — which produces flat, unoptimizable performance.
When you select a feature filter from a data provider (say, an income bracket or a geographic credit profile), you aren’t targeting. You’re getting everyone in that category, with no ranking, no prediction, no optimization. You’ll get the same flat performance on campaign 100 as you did on campaign 1.
That’s the gap Postie closes.
What We Built and Why It’s Different
Postie’s ECOA-compliant models are not a workaround. They aren’t a stripped-down version of our standard modeling with a compliance label bolted on. They are purpose-built models, trained on a carefully defined subset of data that removes protected attributes and, critically, removes features that could be derived from or upstream of protected attributes.
That second part matters enormously. A model that doesn’t directly use age as an input can still encode age-correlated patterns through proxies. This can include certain purchase behaviors, geographic signals, or lifestyle indicators that track closely with protected characteristics. ECOA compliance requires eliminating that upstream influence, not just removing the obvious protected fields.
Our ECOA-compliant models are built on data from Epsilon and Experian — two of the most robust and widely used marketing data sets in the industry. Both random forest and deep learning model architectures are supported, giving you four net-new modeling options accessible directly within Postie’s existing model-building interface.
These models operate within the regulatory framework of ECOA as designed, not as an add-on compliance check, but as the foundation on which the entire model is built.
Performance: What You Can Expect
Until now, financial services marketers have been largely forced into manual feature selection when it comes to targeting. Feature selection produces a completely flat performance curve: no ranking, no optimization, no ability to improve over time. Postie’s ECOA-compliant models produce the same characteristic lift curve as our standard models — better predictors at the top, natural falloff at the bottom, and real optimization leverage throughout.
For a marketer who has been forced to use manual selection, or who has avoided machine learning targeting entirely because of compliance concerns, moving to Postie’s ECOA-compliant models is an unmistakable upgrade.
Where These Models Fit in the Credit Marketing Ecosystem
To better understand these models, it helps to understand ITA marketing offers.
Invitations to apply (ITAs) are lender-initiated outreach seeking to identify consumers likely to respond to and be approved for credit, without using consumer-level credit bureau data. This is where Postie’s ECOA-compliant models operate, and where the opportunity is substantial.
ITAs are a large addressable segment in the financial services space. They use non-credit marketing data, and because they don’t rely on regulated credit bureau data, consumers who have opted out of pre-screen offers can still receive ITAs. That means an even larger addressable universe compared to pre-screen campaigns.
The Bigger Picture for Financial Services
Financial services is the largest category of direct mail spending in the United States. It’s large because direct mail works exceptionally well for credit marketing. The combination of physical permanence, household-level targeting, and high-consideration purchase cycles makes the mailbox a uniquely effective channel for reaching prospective borrowers and account holders.
Until now, the most sophisticated targeting tools available to these marketers have been off-limits due to compliance requirements. That’s changed.
Postie’s ECOA-compliant models open the door for banks, credit unions, fintech companies, and co-branded card programs to run machine learning-powered direct mail campaigns with confidence. The compliance teams that have previously said no to lookalike modeling — on direct mail and across digital platforms — now have a targeting option purpose-built for their requirements.
This is a meaningful unlock for an entire category of marketers who have wanted the performance benefits of programmatic direct mail but haven’t had a compliant path to get there. Postie is that path — and right now, we’re the only ones.
Getting Started
ECOA-compliant models are available now within Postie’s existing model-building interface. The workflow is the same as building any model in the platform. It’s easy to get started.
If you’re marketing a credit product and want to understand which modeling approach is right for your compliance posture, our team is ready to walk through the options. Just get in touch.
