<select> with three labels that read like SKU codes — “Cover Stock Standard 12pt”, “Extra Thick Cover Stock 17 pt”, “Pearlescent Cover Stock 14 pt”.<select id="product_attribute_g238_21" data-attr-name="Paper Weight"> with three <option> children. Verified at funeralprints.com/p/101 on 2026-05-12.
<select> with three inline radio pills. Same option names as today, no descriptive copy. Lightest-touch redesign.<select> is hidden but kept in the DOM (form-submission unchanged). Three radio pills are injected via PostHog web experiment in its place.<select> and in the public-facing Paper Types & Finishes Guide. No new words introduced.
By replacing the Paper Weight dropdown with a radio-button selector that explains what each thickness feels like, paper-thickness selection will become an in-context decision instead of a decision-fatigue tax — lifting overall conversion (matching the direction of the currently-running FP-32 “remove dropdown” test) while preserving the option to upgrade and lifting the share of customers who choose 14pt Pearlescent or 17pt Extra-Thick (higher AOV).
Three variants test increasing levels of explanation: V1 = format change only (pills, no description), V2 = pills + verbatim descriptions from the Paper Types guide, V3 = pills + descriptions + visual paper swatch. Every word and number on every variant is grounded in copy already published on funeralprints.com.
Primary metric: purchase rate (sessions exposed to the variant that complete checkout). Sales is always primary — CLAUDE.md hard rule.
Secondary metrics: revenue per session, paper-weight mix at order (% choosing 14pt Pearlescent or 17pt Extra-Thick), AOV, PDP → designer-start rate (clicks on Click to design)
Guardrails: bounce rate on PDP must not increase, designer start → cart rate must not drop, time-on-PDP must not balloon (no analysis paralysis)
Sample size: ~36k sessions/month sitewide, PDP sessions ~60% of that ≈ 21k/month. PostHog ships only the V2 cell (control vs variant, 50/50) ≈ 10.5k sessions/cell/month. Purchase rate is a smaller-effect metric than designer-start, so plan for **6–10 weeks** to detect a meaningful lift at 95% confidence. Confirm the actual MDE once baseline purchase rate is set up in PostHog.
Audience: URL matches /p/*, desktop + mobile, all geos. The <select> with data-attr-name="Paper Weight" must exist on the page (skips products without paper-weight options).
Selector for injection: hide .attribute.name_Paper_Weight, inject variant UI after it, and on change set document.getElementById('product_attribute_g238_21').value = <optionValue> followed by a change event so nopCommerce’s attribute-change handler fires.
Source page audited: /p/101 and the Paper Types guide at /t/33/paper-options-and-finishes.
Pre-launch checks:
data-attr-name="Paper Weight" should be safe but worth verifying on memorial posters, bookmarks, and programs.<select> value — specifically that the price line and any inline-stock display reflect the chosen paper weight.posthog.featureFlags.override({'fp-34-paper-thickness-radio':'v1'|'v2'|'v3'}), screenshot, send to Valentin before launch.