Risk Tiering Merchants Is a Product Decision

Most platforms ship merchant risk tiers as a risk-team artifact: a spreadsheet of MCC codes, ownership flags, and country lists. That model is the cheapest version of a decision that should sit at the centre of the product.

Tiering is the surface where onboarding, limits, settlement, fees, support, and disputes all converge. If product does not own it, the platform optimises for whoever does, usually risk, sometimes finance, almost never the merchant.

What a tier actually controls

A merchant tier should control, at minimum:

• Onboarding depth, what documents, screens, and approvals are required
• Activation speed, how long from application to first transaction
• Transaction limits, per-txn, daily, monthly, by rail
• Settlement window, T+0 to T+7
• Rolling reserve, 0% to 20%
• Pricing tier, fee schedule and any promotional rates
• Support SLA, response and resolution targets
• Dispute handling, auto-accept thresholds, evidence requirements
• Monitoring intensity, review frequency, alert thresholds

All of those are product surfaces. Each one ships differently per tier.

A workable tier model

A four-tier model holds across most payments platforms:

1. T1, Verified self-serve. Low-risk vertical, clean KYB, small initial limits, T+3 settlement, no reserve. Activates in minutes.
2. T2, Standard. Mainstream vertical, full KYB cleared, T+1 settlement, no reserve, standard limits. Activates within a day.
3. T3, Enhanced. Higher-risk vertical, larger limits, T+1 with small reserve, manual onboarding review.
4. T4, Enterprise. Custom commercial terms, T+0 available, dedicated support, custom dispute handling.

A separate T0, Restricted holds merchants under investigation, with reduced limits and extended settlement.

Tiers are not static

The single biggest design mistake is treating the onboarding tier as the lifetime tier. Production-grade tiering is dynamic:

• Upgrade triggers, 90 days clean, volume thresholds, low chargeback rate, no compliance findings
• Downgrade triggers, chargeback spike, fraud event, ownership change, regulatory finding, sudden volume change
• Auto-review cadence, quarterly for T1/T2, monthly for T3, weekly for T4

Communicate tier changes to merchants. A silent downgrade is the fastest way to destroy trust.

Pricing should follow tier honestly

Most platforms tier pricing by negotiation, not by risk. That is fine commercially, but the tier model should still be honest: every promotional rate carries a documented expected loss assumption, and that assumption is reviewed quarterly against actuals.

When the actuals diverge, either the rate moves, the tier moves, or the loss budget moves. Pretending nothing has changed is how platforms accumulate quiet losses.

Vertical-specific tiers

A small number of verticals justify their own tier dimensions: gambling, crypto, adult content, FX, high-ticket travel, supplements, debt collection. These typically need:

• Stricter onboarding evidence
• Lower per-txn and aggregate limits
• Higher rolling reserves
• Specialised dispute handling
• Periodic external audits

Model these as overlays on the base tier, not as separate tiers. Otherwise the tier matrix explodes.

What to instrument

• Tier distribution by acquisition month
• Mean revenue per tier, net of losses
• Chargeback rate by tier, trending
• Tier upgrade and downgrade rates
• Average time to upgrade
• Tier-pricing actuals vs assumptions

Operator lens

The platforms that win in payments are the ones whose tier model is a product roadmap, not a risk policy. When tiers are clear, merchants know what to do to earn better terms. When they are opaque, the only signal merchants get is denial, and they go elsewhere.

Related: KYB Automation Without Blowing Up Risk · Layered Fraud Controls in the Payments Stack