Skip to content
Academy
Marketing Academy · Field Work●Paid Ads
CoreTeardown· 40 minutes

The Authenticity Teardown: Spotting What Makes UGC Read as an Ad

Duolingo

Objective: Given 3 synthetic UGC ad scripts (one Hook, one Product Demo, one Testimonial), identify the specific lines and structural choices that break the authenticity illusion, versus surface-level flags (a phone camera, a plain background) that don't actually matter.

You're reviewing a batch of 3 UGC ad scripts submitted by a creator platform for Duolingo's language-learning app before greenlighting production, checking each against the lesson's authenticity criteria.

Score each script's defects by severity, distinguish real red flags from things that merely look unpolished but aren't actually the problem.

Before you start

What you'll need

Free path (everything below is enough to finish)

FreeScore and annotate each submitted script before greenlighting production

Free, comment threads track defect notes creator-facing feedback can reference directly

The process

Specimens to review

This is a Hook-format script meant for cold traffic. Identify what breaks the authenticity illusion, and note anything that looks unpolished but isn't actually a defect.

Sample output
SCRIPT A (Hook format, submitted for cold-traffic testing):
[Creator on camera, holding phone] "Hey guys! I'm SO excited to share Duolingo with you today, it's a language-learning app that has completely transformed the way I study! With Duolingo, you can learn over 40 languages using bite-sized, gamified lessons that fit into your busy schedule. I've used the app to learn Spanish for 3 months and I've seen incredible results! Duolingo makes learning fun and effective, unlike boring textbooks!"
[Cut to app screen recording]

Specimen: synthetic, realistic

This is a Testimonial-format script for bottom-funnel retargeting. Identify any real defects versus things that might look like defects but actually support the format's goal.

Sample output
SCRIPT B (Testimonial format, submitted for bottom-funnel retargeting):
[Creator sitting casually at home] "Okay so I have to talk about this. I was honestly skeptical about language apps, I'd tried like 3 before and gave up on all of them within a week. But Duolingo's streak thing genuinely got me, I'm on day 94 now and I actually understand basic conversations in French, which felt impossible six months ago. It's not perfect, some of the later lessons feel repetitive, but for actually sticking with it? Nothing else worked for me like this did."

Specimen: synthetic, realistic

Final deliverable

A scored teardown of all 3 submitted scripts with defects ranked by severity and a go/no-go/revise verdict per script.

See a reference example
Sample output
ThredUp, UGC script review (excerpt)

SCRIPT A (Hook): REVISE, critical, opens with product name instead of a relatable problem
SCRIPT B (Testimonial): APPROVE, mild built-in criticism reads as authentic, no scripted-language flags
SCRIPT C (Demo): REVISE, moderate, framed as an opinion piece instead of a step-by-step walkthrough

Success criteria

You're done when you can:

  • Correctly separates real authenticity defects from cosmetic non-issues
  • Flags scripted/marketing-copy language specifically, not just general tone
  • Does not penalize legitimate testimonial criticism as a defect