Skip to content
Academy
Marketing Academy · Field Work●Conversion Rate Optimization
All projects·From: Trust Signals That Convert
MiniTeardown· 20 minutes

Teardown: The Checkout Page That Looks Secure but Isn't

Coinbase

Objective: Given a synthetic checkout-page specimen, identify which trust elements are genuine signals and which are cosmetic or actively risky, using the lesson's Common Mistakes list.

You're reviewing a funding-page mockup for Coinbase's crypto-purchase flow before it ships, modeled on patterns the lesson flags as common mistakes.

Sort each element into real signal, cosmetic clutter, or active risk, and justify each call against the lesson.

Before you start

What you'll need

Free path (everything below is enough to finish)

Hotjar(optional)
FreemiumConfirm where real users hesitate on the live checkout page before prioritizing which defect to fix first

Free plan captures enough session recordings to validate whether the badge row or the buried guarantee is the bigger drop-off driver

No access? Skip and prioritize by severity rating alone if no analytics tool is available

Google Sheets
FreeLog each specimen element with its severity and fix recommendation

Free, sortable by severity for a quick fix-priority list

The process

Specimens to review

Sort every element on this specimen into: real signal, cosmetic clutter, or active risk. Justify each call in one sentence citing the lesson.

Sample output
Funding-page mockup, synthetic: a row of 11 security and payment badges (SSL, PCI, Visa, Mastercard, Norton, McAfee, BBB, TRUSTe, GDPR, ISO, 'Verified Merchant') stacked in a single dense row above the card-entry form, each badge roughly 24px tall.

Specimen: synthetic, realistic

Is this guarantee doing any conversion work? What's the fix?

Sample output
Below the funding form, small gray text reads: 'Funds protection details available in our Terms of Service (link).' No other guarantee or protection language appears on the page.

Specimen: synthetic, realistic

Final deliverable

A defect log sorting every specimen element into real signal / cosmetic clutter / active risk, with one fix per defect.

See a reference example
Sample output
PolicyBazaar funding-page teardown (excerpt)

DEFECT: Guarantee text sized 10px, gray-on-white, placed below fold
SEVERITY: moderate
FIX: Restate at 14px, near the CTA, in plain language

Success criteria

You're done when you can:

  • Correctly separates real signals from cosmetic clutter for both items
  • Cites the specific lesson section supporting each defect call