Home / Social proof / Demo spec
Download raw .md

Demo spec — Product page before/after ("Marlow" DTC skincare)

Single self-contained interactive HTML file: demo.html in this folder. On-screen artifact for Act Four of the social-proof script. WEB pages: two browser-window mockups (traffic-light chrome, URL bar "marlow.co/serum"), side-by-side ≥1200px, stacked below. Labels: "BEFORE · 9 claims, 0 audited" / "AFTER · 7 claims, 7 audited".

The brand

"Marlow" — DTC skincare, $38 "Renewal Serum". Editorial-minimal aesthetic: warm ivory background, deep forest-green text/accents (#1E3B32-ish), a soft terracotta secondary, serif wordmark (Georgia stack), product shown as a simple inline-SVG amber bottle illustration. Distinct from all other demos (navy sleep app, coral running app, teal docs SaaS, green budgeting app).

BEFORE page (sins from the audit + Mathur dark-pattern crawl — believable, not cartoon)

  • Product block: bottle, name, $38
  • Rating: five perfect gold stars — NO count, NO decimal
  • Banner row: " Only 2 left!" · " 12 people viewing" · countdown "Sale ends 04:59" (ticks and RESETS — the series' recurring lie-detector easter egg)
  • Self-composed laurel: "★★★★★ BEST SERUM — 2026"
  • Three interchangeable rave testimonials, first-name-only: "Amazing!! Life changing!! — Jessica" / "Obsessed. — Mia" / "Everyone needs this. — Sarah"
  • "Everyone loves Marlow ★★★★★" strip
  • Buy box with a soup of six generic security seals (draw as generic gray shield/lock SVG chips)
  • Add to cart CTA Interactions: countdown ticks/resets; "12 people viewing" number randomly wobbles (11→13→12) — annotate that this is how the fakes actually work (random number generators).

AFTER page

  • Product block identical (same product, same price)
  • Rating: 4.6 ★ · 1,284 reviews (decimal + count) — clicking scrolls to reviews
  • Full rating histogram (SVG bars): 5★ 812 / 4★ 289 / 3★ 96 / 2★ 41 / 1★ 46 — one-star bar clickable → reveals a real critical review
  • Review area, two slots side by side: "Most helpful" (detailed positive, verified-buyer badge, user photo placeholder, specific: "Texture changes were visible around week 5. Pump broke once — support replaced it in 2 days.") and "Most critical" (2★, verified: "Did nothing for me after 3 weeks, and the scent is strong. Returned it — refund was fast.")
  • CONS chips row (Ulta pattern): "Strong scent (214)" · "Takes weeks (187)" · "Pricey (166)" — clickable-looking filters
  • One hero testimonial with full attribution: "My dermatologist asked what I switched to — six weeks in." — Dana R., verified buyer (avatar, date)
  • True-demand line replacing scarcity: "Restocked monthly — the last three restocks sold out." (calm, factual)
  • Buy box: payment marks only (Apple Pay, PayPal, Visa/MC as inline SVG wordmark-style chips) + one plain guarantee line: "30-day returns, no questions." Interactions: histogram 1★ click reveal; CONS chips toggle a small filtered-review snippet; hero rating scroll.

Annotation layer

"Show annotations" toggle (default ON). Numbered pins matching Act Four order: 1 the score (5.0-no-count → 4.6+1,284), 2 the histogram (audited record), 3 the blemish (most critical beside most helpful + CONS), 4 specificity/attribution (vague raves → Dana R.), 5 the deleted theater (fake scarcity → true demand line), 6 the buy box (seal soup → payment marks + guarantee). Pin cards cite ONLY these stats, verbatim: - Purchase likelihood peaks around 4.2–4.5 stars and declines toward 5.0 (Spiegel Research Center 2017, analyzing PowerReviews data) - Going from 0 to 5 reviews raised purchase likelihood ~270% (Spiegel Research Center 2017) - A small negative after positives increased purchase likelihood — the blemishing effect (Ein-Gar, Shiv & Tormala 2012) - One random fake upvote inflated final scores 25%; fake downvotes were corrected by the crowd (Muchnik, Aral & Taylor 2013) - Researchers found resetting countdowns and random-number stock counters on 183 sites, sold by 22 third-party vendors (Mathur et al. 2019, 53k pages) - Fake reviews and review suppression now carry FTC penalties up to $51,744 per violation (16 CFR Part 465, effective Oct 2024) - Our audit (n=72 top buy surfaces, Aug 2026): 71% of retailer product pages expose negative info vs 0 of 24 seller-authored surfaces; 79% of product pages pair the rating with a review count; zero "only X left" or "people viewing" cues appeared anywhere in the sample

Footer

"Built for the Build With Kris video 'Perfection Reads as Fraud' — sources in the video description." Scoreboard: BEFORE — claims 9 · audited 0 · FTC-risk elements 3 | AFTER — claims 7 · audited 7 · FTC-risk elements 0.

Craft bar

Fully self-contained, offline via file://, system fonts, inline SVG only, responsive, premium editorial feel. Page chrome adapts to prefers-color-scheme; mockup pages stay light. All displayed math consistent (histogram sums to 1,284).