Case: Switching from flat to tiered referral rewards lifted active-referrer rate 3.4x

Case: Switching from flat to tiered referral rewards lifted active-referrer rate 3.4x

A B2B fintech with a customer-referral program found that 92% of referrals came from 6% of eligible customers — a textbook power-law, but an extreme one that left the program dependent on a handful of advocates.

The original incentive was flat: $200 account credit per closed referral. In 2023 they restructured to a tiered model — $150 for the first, $250 for the third, $500 for the fifth-plus within a year — plus a non-cash status tier (early feature access) at four referrals.

Over three quarters:

— Active-referrer rate (customers making ≥1 referral) rose from 6% to ~20%, a 3.4x increase.
— Average referrals per active referrer climbed from 1.3 to 2.1.
— Cost per acquired customer via referral rose 18% — the tiers were genuinely more expensive at the top.

The net was positive because referred customers retained materially better: their 12-month logo churn ran ~40% below non-referred, consistent with the long-standing Wharton finding (Schmitt et al., 2011) that referred customers carry higher lifetime value and lower churn.

A confound worth naming: the restructure coincided with a referral-UX overhaul (one-click sharing inside the product). Some of the activation lift almost certainly came from reduced friction, not the incentive curve itself. The company ran them together and cannot cleanly separate the two.

The trade-off: tiered rewards raise top-end cost and can feel gameable. They added a fraud check (referred account must reach 90-day activation) that voided ~7% of payouts.

Open questions: How much of the 3.4x is incentive design versus friction removal? And does the new active-referrer base persist once the novelty of status tiers fades?