The architecture of American commerce has historically excluded entire categories of legitimate businesses from financial infrastructure—not due to actual fraud or insolvency, but due to categorical risk aversion embedded in traditional underwriting models. CBD retailers with compliant operations, subscription services with sustainable customer bases, telemedicine providers serving genuine medical needs, and dozens of other verticals face systematic payment processing rejection despite operating lawful, viable businesses. This exclusion creates a shadow economy where merchants cycle through processors, experience unexpected account terminations, and operate with constant financial infrastructure instability. Today we examine this challenge with Evan Albert, CEO of SeamlessChex, whose company processes over $2 billion annually with the highest approval rates for high-risk merchants. His firm has built specialized infrastructure that fundamentally reimagines how payment processors evaluate, approve, and support merchants in complex risk environments.
At Zarin Fabrics, our family has operated since 1936 by understanding a fundamental truth: perceived risk often reflects market misunderstanding rather than actual business viability. We began as a pushcart on the Lower East Side serving customers mainstream retailers ignored, and grew into a three-floor warehouse by recognizing opportunity where others saw only obstacles. Payment infrastructure that enables legitimate commerce across all sectors—not just the statistically safest—strengthens the entire economic ecosystem. As someone who's navigated market transformations across nine decades of family business operations, I wanted to understand the technical and strategic mechanisms Evan has developed to solve what amounts to a market failure in payment processing.
Q: Evan, let's establish the technical foundation. When mainstream processors decline high-risk merchants, what specific underwriting mechanisms drive those rejections, and what alternative evaluation frameworks have you developed at SeamlessChex?
A: The rejection mechanism is fundamentally algorithmic and categorical rather than analytical. Mainstream processors use merchant category code (MCC) filtering as a first-pass screen—before any human review, applications containing MCCs like 5912 (drug stores and pharmacies, which includes CBD), 5967 (direct marketing subscriptions), 7995 (betting and gambling), or 5962 (direct marketing travel) trigger automatic declines. This happens because their acquiring bank relationships explicitly prohibit these categories, and because their risk models are calibrated for false-negative optimization—meaning they'd rather incorrectly reject viable merchants than incorrectly approve risky ones. The economics make sense for their business model: a mainstream processor handling mostly low-risk retail doesn't want their overall merchant portfolio risk profile contaminated by high-chargeback categories, even if specific merchants in those categories would perform well.
At SeamlessChex, we've built what I call probabilistic merchant evaluation rather than categorical exclusion. Our underwriting framework analyzes seven distinct risk vectors: industry baseline risk (unavoidable factors inherent to the vertical), operational risk mitigation (what controls the merchant has implemented), compliance infrastructure (licensing, legal structure, regulatory adherence), processing history (past performance with other processors), business model mechanics (how revenue is generated and fulfilled), financial stability (capitalization, cash flow, business maturity), and reputational signals (online reviews, BBB ratings, media presence). Each vector gets scored independently by analysts with vertical expertise—we have underwriters who specialize exclusively in CBD, others in subscription services, others in telemedicine. This specialization matters enormously because risk factors are highly contextual.
For example, a CBD merchant's risk profile depends critically on whether they're selling full-spectrum products or CBD isolate, whether they're sourcing from licensed cultivators or grey-market suppliers, whether they conduct third-party lab testing and publish certificates of analysis, how they handle marketing claims around medical benefits, and what states they ship to. Generic underwriters evaluating CBD merchants categorically miss these distinctions entirely. Our CBD-specialist underwriters can identify a low-risk CBD operation that should be approved versus a high-risk one that should be declined, even though both fall under the same MCC.
This granular evaluation framework is why we maintain the highest approval rate for high-risk merchants—we're approving the substantial percentage of merchants in "high-risk" categories who actually operate low-risk businesses. Our decline rate within high-risk applications is still significant because we're genuinely evaluating risk rather than approving indiscriminately, but we're approving perhaps 60-70% of high-risk applications versus the mainstream processor rate of nearly 0%. That difference represents thousands of viable businesses gaining payment infrastructure access.
Q: You mentioned processing over $2 billion annually across 20+ acquiring bank relationships. Walk me through the technical architecture that enables that multi-bank infrastructure and how it differs from traditional processor-bank relationships.
A: This is where the infrastructure complexity becomes apparent. Traditional processors maintain correspondent relationships with one, maybe two acquiring banks. When a merchant applies, the processor submits their application to that single bank for approval. If the bank declines—which they will for any high-risk category—the merchant is rejected. The processor has no alternative pathway because they've optimized their entire technology stack around a single banking relationship that offers the best economics for their target merchant profile.
SeamlessChex operates what's essentially a multi-bank routing and relationship management platform. We've built direct integrations with 25+ acquiring banks that each specialize in different risk profiles, geographic markets, and industry verticals. These aren't casual partnerships—each bank relationship required months of negotiation, legal structuring, compliance alignment, and technical integration. Each bank has different underwriting requirements, risk appetites, settlement terms, chargeback thresholds, and pricing structures. Our platform intelligently routes merchant applications to appropriate banking partners based on industry, processing volume, geography, and risk profile.
Here's a concrete example: A CBD merchant applies to process $300,000 monthly. Our system identifies three potential banking partners in our network that underwrite CBD merchants at that volume threshold. Bank A specializes in CBD but requires 6 months of processing history and clean compliance documentation. Bank B accepts newer CBD merchants but has higher reserve requirements. Bank C works with CBD merchants specifically in the wellness/supplement positioning versus recreational. Our underwriter evaluates which bank best fits this specific merchant's profile and submits accordingly. If Bank A declines due to the merchant being too new, we can immediately route to Bank B rather than rejecting the merchant entirely.
This routing capability is why we process over $2 billion annually—we're not constrained by a single bank's capacity, risk appetite, or category restrictions. As we scale, we add banking relationships rather than hitting volume ceilings. The technical infrastructure required to manage this is substantial: we maintain separate processor integrations, reconciliation systems, settlement protocols, and compliance reporting for each banking partner. We've built middleware that normalizes transaction data, routing logic, and reporting across heterogeneous banking systems. This isn't just payment processing—it's financial infrastructure orchestration.
The strategic value extends beyond approval rates. Multi-bank infrastructure provides merchant stability. If a banking partner changes their risk appetite or exits a vertical, we can migrate merchants to alternative banks without business disruption. Merchants aren't vulnerable to single-bank policy changes. It also provides negotiating leverage—we can direct volume to banking partners offering optimal terms, which creates competitive dynamics that benefit our merchants.
Q: Let's discuss chargeback management specifically, since that's often cited as the core risk factor for high-risk merchants. What technical and operational mechanisms have you developed to manage chargebacks at scale?
A: Chargebacks represent the asymmetric risk that makes processors avoid high-risk categories. In traditional card processing, merchants bear 100% of chargeback liability plus fees, but excessive chargeback rates (above 1% of transactions) can result in losing your merchant account entirely. For subscription services, certain verticals naturally generate 2-3% chargeback rates simply due to business model dynamics—customers forgetting about renewals, disputing charges rather than requesting cancellation, or experiencing buyer's remorse. Mainstream processors can't accommodate 2-3% chargeback rates even if they're predictable and manageable, so they categorically decline these merchants.
We've developed a multi-layer chargeback management system that treats disputes as operational challenges to be mitigated rather than reasons for merchant rejection. Layer one is preventative: we analyze each merchant's transaction data to identify chargeback pattern precursors. For subscription merchants, we monitor things like time-to-first-chargeback after signup, correlation between promotional discounts and dispute rates, and seasonal variation. We provide merchants with specific operational recommendations: optimize your billing descriptor so charges are recognizable, implement pre-charge email notifications, create obvious cancellation flows, improve customer service responsiveness. These operational changes can reduce chargeback rates by 30-50% without any payment infrastructure changes.
Layer two is automated representment. When a chargeback occurs, our system immediately categorizes it: friendly fraud (customer purchased but claims they didn't), subscription amnesia (customer forgot about recurring charge), product/service dispute (legitimate quality complaint), or fraud (stolen card). For friendly fraud and subscription amnesia cases—which comprise perhaps 60% of high-risk chargebacks—we automatically compile evidence and submit representment: transaction receipts, IP address logs, email confirmations, terms of service acceptance, delivery confirmation. Our representment win rate is approximately 40%, which means we're recovering significant value that merchants would otherwise lose.
Layer three is forensic analysis for patterns indicating systematic problems versus random variation. A sudden chargeback spike might indicate a fraud attack, a customer service failure, a product quality issue, or a seasonal anomaly. Our system flags anomalies and our account managers work with merchants to diagnose root causes. Perhaps the merchant's website was hacked and customer data compromised—we'd implement additional fraud screening. Perhaps a bad product batch led to quality complaints—we'd work with the merchant on customer remediation. This analytical approach treats chargebacks as business intelligence rather than simply processing costs.
Layer four is selective merchant exits. Despite sophisticated management, some merchants genuinely operate problematic businesses—deceptive marketing, poor product quality, intentionally difficult cancellation processes. When we identify merchants with consistently high chargeback rates despite mitigation efforts, we exit those relationships. This selective approach is how we maintain the highest approval rate for high-risk merchants while keeping our overall portfolio risk sustainable. We're approving based on sophisticated evaluation and managing based on performance, not excluding categorically.
The technical infrastructure supporting this includes real-time transaction monitoring, automated alert systems, evidence compilation engines, representment automation, and predictive analytics. We've essentially built a chargeback management platform as sophisticated as our payment processing platform because for high-risk merchants, dispute management is as critical as transaction processing.
Q: How specifically does your fraud detection differ from mainstream processors when dealing with high-risk transaction patterns that would trigger false positives in standard systems?
A: Fraud detection in high-risk environments requires fundamentally different calibration than traditional retail. Mainstream processors use fraud models trained predominantly on low-risk transaction patterns—average ticket values of $50-$200, geographic concentration in domestic markets, velocity patterns consistent with individual consumer behavior. When you apply those models to high-risk merchants, you generate enormous false-positive rates that damage conversion and customer experience.
Consider a telemedicine merchant. Legitimate transaction patterns include: high average ticket values ($200-$500 for consultations and prescriptions), frequent international transactions (patients accessing US-licensed physicians from abroad), rapid transaction velocity (patient surges during health events), and recurring patterns (ongoing prescriptions). Every one of these characteristics would trigger fraud alerts in standard systems calibrated for retail. If we used mainstream fraud detection, we'd decline perhaps 20-30% of legitimate transactions, which would make the merchant's business model unviable.
We've developed vertical-specific fraud models trained exclusively on transaction patterns within particular industries. Our telemedicine fraud model knows that $400 transactions from international IP addresses aren't inherently suspicious—they're normal. What IS suspicious in telemedicine is things like: multiple transactions from the same IP using different patient identities, prescription patterns inconsistent with medical protocols, velocity spikes uncorrelated with marketing activities, or geographic clustering suggesting organized fraud operations. These nuanced indicators only become visible when you have deep data sets within specific verticals.
We use machine learning models that continuously adapt to each merchant's specific transaction patterns. During the first 30-60 days of processing, our systems learn what's normal for this particular merchant—their customer acquisition channels, geographic distribution, ticket value patterns, time-of-day transaction clustering. After this learning period, we're detecting anomalies specific to this merchant rather than applying generic fraud rules. A transaction that would be suspicious for Merchant A might be completely normal for Merchant B, even within the same industry.
We also implement multi-factor fraud scoring rather than binary approve/decline decisions. Transactions receive risk scores (0-100) based on dozens of signals: device fingerprinting, IP geolocation, email domain age, billing-shipping address correlation, velocity patterns, AVS/CVV verification, historical customer behavior. Low-risk transactions (score 0-20) approve automatically. Medium-risk transactions (20-60) might require additional verification—3D Secure authentication, phone verification, email confirmation. High-risk transactions (60-100) decline or flag for manual review. This graduated approach maximizes approval rates while maintaining security.
The technical infrastructure includes real-time decisioning engines capable of scoring transactions in under 100 milliseconds, machine learning pipelines that retrain models daily with new transaction data, and human review queues for ambiguous cases. We process over $2 billion annually through this system, which means we're making millions of fraud decisions monthly. The scale provides enormous data advantages—our models see fraud patterns across thousands of merchants and can identify emerging attack vectors quickly.
For high-risk merchants who've experienced either excessive fraud losses with insufficient protection or excessive false declines with overly aggressive protection, our calibrated approach represents a dramatic improvement. We're optimizing for the right balance: maximize legitimate transaction approval while minimizing fraud exposure to sustainable levels.
Q: You've built sophisticated infrastructure, but that requires significant technical investment. How do the economics work—how does serving high-risk merchants remain financially viable given the operational complexity you've described?
A: This is the fundamental economic question that explains why mainstream processors avoid high-risk merchants despite the obvious market opportunity. The traditional payment processing business model is volume-based and margin-thin: you process enormous transaction volumes at small percentage margins (maybe 0.3-0.5% net margin after card network fees, bank fees, and operational costs) and compensate for thin margins with massive scale and operational efficiency through automation.
High-risk processing economics work differently. Our rates are higher—where a mainstream processor might charge 2.5% + $0.10 per transaction for low-risk retail, we might charge 3.5-5% + $0.30 for high-risk categories. This rate premium reflects actual costs: specialized underwriting, dedicated account management, sophisticated fraud detection, chargeback management, reserve requirements, and higher acquiring bank fees. The common criticism is that high-risk processors are exploitative, but the economics are legitimately different. Our operational costs per merchant are 3-5x higher than mainstream processors due to relationship-based service versus automated self-service.
The key to viability is selective approval and performance management. We're not approving every high-risk application—we're approving merchants we've evaluated as genuinely viable. Our approval rate is the highest for high-risk merchants, but it's still probably 60-70% of applications, not 100%. The merchants we decline would likely generate unsustainable chargeback rates or fraud losses that would destroy economics. The merchants we approve perform sustainably because we've properly evaluated risk and implemented appropriate mitigation.
Additionally, our multi-bank infrastructure creates negotiating leverage that improves economics. We can direct volume to banking partners offering optimal interchange rates, favorable reserve terms, or reduced chargeback fees. As we've scaled to over $2 billion in annual processing, we've become a significant volume partner for our acquiring banks, which provides negotiating power that smaller high-risk processors don't have. We can pass some of these economic improvements to merchants through competitive rates while maintaining our margins.
The strategic insight is that high-risk processing is an expertise business, not a commodity business. Merchants pay for specialized knowledge, infrastructure, and relationship management that enables their business model. For a CBD merchant processing $500,000 monthly who's been declined by every mainstream processor, paying an extra 1% in processing fees to gain stable payment infrastructure is obvious value. That $5,000 monthly cost enables $500,000 in revenue—it's not exploitative, it's appropriate pricing for specialized service.
Our economics improve as merchants scale with us. A new CBD merchant might start at $50,000 monthly processing volume. As they grow to $200,000, then $500,000, then $1 million monthly, our operational costs per transaction decrease (we've already done the underwriting, built the relationship, integrated their systems) but we're processing more volume at similar rates. This creates a virtuous cycle where successful merchant growth drives our growth while improving our margins. We're incentivized to help merchants succeed rather than maximize short-term extraction.
Q: Looking strategically, where is the high-risk payment processing landscape evolving, and what infrastructural or regulatory changes could fundamentally reshape this market?
A: We're at an inflection point driven by several converging forces. First, alternative payment rails are maturing beyond traditional card networks. ACH processing, particularly with real-time verification through platforms like Plaid, offers dramatically lower costs (maybe 0.5-1% versus 3-4% for card processing) and lower chargeback risk (ACH disputes are less common and easier to contest). For many high-risk merchants, particularly those with higher ticket values or subscription models, ACH is economically superior. We're seeing increasing merchant preference for offering ACH as a primary payment method with cards as backup. Cryptocurrency rails are similarly evolving—not for speculative investment but as actual payment infrastructure with near-zero chargeback risk (transactions are irreversible) and dramatically lower fees. For international high-risk merchants, cryptocurrency eliminates currency conversion costs and cross-border processing barriers.
Second, regulatory frameworks are maturing in previously grey-area industries. The 2018 Farm Bill federally legalized hemp-derived CBD, which should eventually lead to mainstream processor acceptance—but we're still seeing categorical CBD declines four years later because processors move conservatively. As state-level cannabis legalization expands and federal reform progresses, we'll see massive payment infrastructure needs in that sector. Telemedicine received enormous regulatory clarity during COVID as states implemented interstate licensure compacts and CMS expanded reimbursement. These regulatory developments should reduce risk classifications, but processor behavior lags regulatory reality by years. This creates opportunity for specialized processors who move quickly while mainstream processors remain cautious.
Third, AI and machine learning are transforming what's possible in risk assessment. The underwriting and fraud detection I've described—vertical-specific models, continuous learning, nuanced pattern recognition—will become exponentially more sophisticated. We'll move from categorical risk assessment to genuinely individualized merchant evaluation. Instead of "all CBD merchants are high-risk," systems will evaluate "this specific CBD merchant's compliance infrastructure, operational controls, and business model produce a 0.8% predicted chargeback rate, which is acceptable." This granularity will expand access for legitimate businesses while improving risk management.
Fourth, embedded finance and vertical SaaS platforms are creating industry-specific payment infrastructure. Rather than general-purpose processors serving all merchants, we're seeing specialized platforms built for specific verticals—telemedicine practice management software with integrated payment processing, CBD e-commerce platforms with compliant processing, subscription management platforms with optimized recurring billing. At SeamlessChex, we're evaluating how to participate in this embedded finance evolution while maintaining our multi-vertical expertise.
The strategic question for businesses in high-risk categories is how to position for this evolution. Invest heavily in compliance infrastructure now—proper licensing, transparent operations, documented policies—because that's increasingly differentiating in risk evaluation. Evaluate alternative payment rails beyond cards to reduce costs and risk exposure. Build relationships with specialized processors rather than constantly chasing rate promotions from processors who don't understand your business. And most importantly, treat payment infrastructure as a strategic business foundation rather than a commodity cost center.
For processors, the winners will be those building genuinely sophisticated risk evaluation infrastructure rather than relying on categorical exclusions. The market opportunity is enormous—thousands of legitimate businesses systematically excluded from financial infrastructure—but capturing it requires technical sophistication, banking relationship development, and operational expertise that takes years to build. That's why SeamlessChex's position processing over $2 billion annually with the highest approval rate for high-risk merchants represents a genuine competitive moat. This isn't infrastructure you can replicate quickly, which protects our market position even as the industry evolves.
For more information about payment processing infrastructure for high-risk merchants, visit seamlesschex.com.