Skip to content
Academy
Marketing Academy · Field WorkCopywriting
MiniTeardown· 20 minutes

Would It Win the Test? Score Three Headline Specimens

Drunk Elephant

Objective: Given three synthetic-realistic headline candidates drafted for the same blog brief, decide whether each would win an A/B test, name every real defect using the lesson's own vocabulary, and correctly recognize which near-defect is actually a deliberate, effective choice.

You're doing a one-day trial task for a clean-beauty content team modeled on Drunk Elephant's ingredient-education positioning. They've drafted three headline candidates for a blog post about their 'Suspicious 6' ingredient-exclusion list and want a scored teardown before anything reaches the editorial calendar.

Three specimens, one prompt each: read the headline, decide whether it would win an A/B test, and name every real defect, tying each one to a specific Step, Mistake, or Trigger from the lesson (not a vague opinion). Rate each defect's severity. For the third specimen, do the harder part too: name at least one thing you almost flagged as a defect but correctly recognized was not one.

Before you start

What you'll need

Free path (everything below is enough to finish)

FreeBuild the 3-specimen defect table with severity and lesson reference columns

Free spreadsheet tool, more than enough for a 3-row scoring table.

FreemiumOptional character-count cross-check on specimen 3's borderline length

Free web tool, useful only to confirm the exact character count, not to find the defects themselves.

The process

Specimens to review

Headline candidate #1 for the Suspicious 6 blog post. Would this win the test? Name every real defect.

Sample output
Why Ingredients Matter

Specimen: synthetic, realistic

Headline candidate #2 for the same post. Would this win the test? Name every real defect.

Sample output
This One Ingredient Swap Will Change Your Skin Forever

Specimen: synthetic, realistic

Headline candidate #3 for the same post. Would this win the test? Name every real defect, then name at least one thing you almost flagged that is not actually a defect.

Sample output
The 6 Ingredients We Will Never Put In Your Skincare (And Why)

Specimen: synthetic, realistic

Final deliverable

A completed defect list for all three specimens (defect, severity, lesson reference for each) plus a short written note on the specimen-3 issue you correctly did not flag as a defect.

See a reference example
Sample output
Sample scoring format from a different brief, not the Suspicious 6 post, to show what 'done' looks like. For a men's grooming brand's candidate 'The Beard Oil Ingredient Most Brands Skip (And Why It Matters)': no critical defect found, one cosmetic note that at 61 characters it sits one over the sweet spot, and the parenthetical correctly left unflagged since it extends the curiosity gap rather than clutters it.

Success criteria

You're done when you can:

  • Correctly identifies the critical defect in specimen 1 (topic-only opener, no benefit, no formula)
  • Correctly identifies both the over-promising defect and the specificity defect in specimen 2, as two separate issues
  • Does not mark the parenthetical or the word 'Never' in specimen 3 as defects
  • Every defect named is tied to a specific lesson Step, Mistake, or Trigger, not a personal opinion