CASE 01 · LENDING STRATEGY
Pricing Experiments & Rates
An attractive offer.
A healthy portfolio.
For the same loan amount and term, how might pricing change offer take-up—and who accepts? Explore the balance across partner channels and seasonal conditions.
Offer take-up at two prices
Accepted / offeredScenario assumptions & experiment design +
This sandbox uses invented response curves to explain a decision framework. It is not fitted to lending data, a pricing recommendation, or evidence of a causal effect. APR is treated as a nominal annual rate with monthly amortization and no fees.
Within the selected channel and season, both offers use the same $10,000 principal, 36-month term, eligible applicant pool, and credit policy. Approval is held at an illustrative 60% of completed applications. Offer take-up is accepted offers divided by offers made; it is not approval or funding. We assume every approved applicant receives an offer. In practice, pricing may also affect affordability eligibility; that requires a separate measured approval outcome.
At 16% APR, assumed take-up is 42% for the rate-comparison partner and 50% for the embedded partner. Each 1 percentage point APR increase reduces take-up by 2.2 or 1.2 percentage points respectively. Competitive season subtracts 5 points and multiplies the price sensitivity by 1.25. These choices are illustrative, not estimated elasticities.
The assumed accepted-cohort default rate starts at 4.0% or 3.5% by channel; competitive season adds 0.3 percentage points. Each APR point above or below 16% changes this proxy by 0.12 percentage points. This intentionally hypothetical selection relationship is not a claim that lower prices universally improve credit. The example uses a 4.5% risk guardrail and a common 12-month outcome window. A real analysis needs mature outcomes or validated early indicators.
To estimate a causal pricing effect, compare concurrent randomized offers within pre-specified channel, risk, amount, and term strata, with persistent assignment per applicant. Define denominators, primary metrics, guardrails, sample size, and stopping rules before each test wave. Check assignment balance and uncertainty; account for repeated looks and multiple comparisons. Block or stratify by time to avoid confusing seasonality with price effects. Observational adjustments alone do not establish causality.
Track offer acceptance per eligible randomized applicant as well as conditional take-up, approval, funding, margin after funding costs and losses, and credit outcomes. Accepted-cohort risk can change through selection; comparing only acceptors does not identify the direct causal effect on borrower credit risk. A launch decision needs those trade-offs, not just the lowest rate or highest take-up.