Test Base Prices, Not Just Promo Framing (Without Training Shoppers to Wait)

Test Base Prices, Not Just Promo Framing (Without Training Shoppers to Wait)

Almost all "price testing" in ecommerce is cosmetic: $48 vs. $52 as the compare-at, or 10% vs. 15% on the banner. The base price never moves. That's promo-framing optimization dressed up as pricing work.

Real price testing moves what the customer actually pays — persistently, observably — and learns willingness to pay. You can do it without teaching the market to wait.

Promo tests and price tests are different animals

A promo framing test (10% vs. 15%, this week vs. last) measures how you present value. It trains shoppers, contaminates fast (codes leak, expectations form), and answers "which discount gets clicked."

A base-price test (the product costs $48 for this cohort and $52 for that one — no banner, no code, no announcement) measures what the product is worth. No training, because there's nothing to learn: the price simply is what it is, where the shopper sees it.

Conflating them is the standard failure: stores run a "price test" as a promoted discount, watch the control move too, and learn nothing.

The design

1. Persist the price per cohort. Geo- or cohort-split traffic, so the same shopper consistently sees their assigned price. Rotating daily prices everyone sees isn't a test — it's yo-yo pricing that teaches waiting.
2. Never run promo framing concurrently. If one arm has a banner and the other doesn't, you tested the banner. Test price alone, then promo presentation separately.
3. Sequence: price → promo → clearance. Establish the base-price curve first; then test promotional presentation around it; then handle clearance as inventory recovery, not data.
4. Measure contribution, not conversion. A lower price that converts more can still lose if margin and returns move against you. Contribution per eligible session, refund-adjusted, is the readout.

What you're actually estimating

Willingness-to-pay curves are local, not global. You're learning:

• Where charm-price decay sets in for your brand — does $49 vs. $50 still work at your price points, or are you past where round numbers matter?
• How much a subscription or bundle price can rise before churn moves.
• Which SKUs are price-elastic (small moves, big volume) vs. inelastic (price is not the lever — fit, trust, or shipping is).

None of that is learnable from discount-banner tests.

Guardrails

• No yo-yo. A shopper who sees $48 Monday and $55 Thursday learns to wait. Cohort-persistent prices avoid teaching delay.
• Mind maps and fairness exposure. Test within legal comfort — and never personalize the base price by identity. Vary presentation and segment, not who-gets-what.
• Watch returns. Price moves shift who buys, which shifts the return mix. A price test without refund tracking is half a test.

The takeaway

Promo framing tells you how to present value. Base-price testing tells you what the value is. Most stores only ever do the first — then wonder why every lesson they learn expires with the sale.

Move the base price for a randomized slice, measure contribution per session, and learn what your product is actually worth — before you decide what discount to hang off it.