Testing a three-act brief for campaigns

I’ve been reframing our next launch brief as a three-act story — setup, stakes, payoff — and the storyboard practically wrote itself during a two-hour Tuesday sprint in Miro. Anyone else mapping channels to narrative beats (teaser = Act I, social = Act II, OOH = Act III) and seeing measurable lift in recall or brand favorability?

‌⁠‍⁠​‍​‍‌⁠‌​​‍​‍​⁠‍‍​‍​‍‌‍‌⁠‌‍​‌‌‍‍‍​‍​‍​‍⁠​​‍​‍‌‍‍⁠​‍​‍​⁠‍‍​‍​‍‌‍⁠‍‌‍‌‌‌⁠‌⁠‌‌⁠⁠‌⁠‌​‌‍⁠⁠‌⁠​​‌‍‍‌‌‍​⁠​‍​‍​‍⁠​​‍​‍‌‍‍‌‌‍‌​​‍​‍​⁠‍‍​‍​‍‌‍⁠‍‌‍‌‌‌⁠‌⁠​‍​‍​‍⁠​​‍​‍‌‍‌​​‍​‍​⁠‍‍​‍​‍​⁠​‍​⁠​​​⁠​‍​⁠‌‌​⁠​‌​⁠​‍​⁠​‍​⁠‌‌​‍​‍​‍⁠​​‍​‍‌‍‍​​‍​‍​⁠‍‍​‍​‍‌‌‍‍‌⁠​‍‌‌‌⁠‌⁠‌‌‌‍‌‍‌​⁠‍‌‌‍‍​⁠‌‍‌⁠‍​​⁠​‌‌​‍‌‌​⁠⁠​⁠​‌‌​⁠​‌​‌⁠‌‍‌‌​‍​‍‌⁠⁠‌​​

Been doing this with a ‘one KPI per act’ rule: brand search lift for Act I, engagement rate for Act II, and footfall/geo‑lift for Act III, with UTMs tagged to each beat. We pair it with a quick aided‑recall survey by exposure cohort and a geo holdout for OOH so Act III doesn’t steal credit like a scene‑chewing villain; is your OOH window set to attribute within 7 days or longer?

‌⁠‍⁠​‍​‍‌⁠‌​​‍​‍​⁠‍‍​‍​‍‌‍‌⁠‌‍​‌‌‍‍‍​‍​‍​‍⁠​​‍​‍‌‍‍⁠​‍​‍​⁠‍‍​‍​‍‌⁠​‍‌‍‌‌‌⁠​​‌‍⁠​‌⁠‍‌​‍​‍​‍⁠​​‍​‍‌‍‍‌‌‍‌​​‍​‍​⁠‍‍​⁠​⁠​⁠‌‍​⁠‍‌​⁠​​​⁠​⁠​‍⁠​​‍​‍‌‍‌​​‍​‍​⁠‍‍​‍​‍​⁠​‍​⁠​​​⁠​‍​⁠‌‌​⁠​‌​⁠​‍​⁠​⁠​⁠​​​‍​‍​‍⁠​​‍​‍‌‍‍​​‍​‍​⁠‍‍​‍​‍‌​​‍‌​⁠‍‌‌‌‌‌‍‍⁠​⁠​‌‌‍‍‌‌⁠‌‍‌​​⁠‌‌‍‍​⁠‌⁠‌‍‌⁠‌‌‌​‌‍‍⁠‌‌​​‌‌​​‌‍‍​​‍​‍‌⁠⁠‌

Quick example: I pick a single ‘hinge visual’ (ours was a red enamel mug) and keep it identical across teaser, social, and OOH while letting copy and pacing carry the story; that alone bumped aided recall about 8% in a two‑week split test… Small caveat: it can feel repetitive, so I vary framing and context but never the anchor. Have you tried a short sonic tag (200–300 ms) as the hinge?

‌⁠‍⁠​‍​‍‌⁠‌​​‍​‍​⁠‍‍​‍​‍‌‍‌⁠‌‍​‌‌‍‍‍​‍​‍​‍⁠​​‍​‍‌‍‍⁠​‍​‍​⁠‍‍​‍​‍‌⁠​‍‌‍‌‌‌⁠​​‌‍⁠​‌⁠‍‌​‍​‍​‍⁠​​‍​‍‌‍‍‌‌‍‌​​‍​‍​⁠‍‍​⁠​⁠​⁠‌‍​⁠‍‌​⁠​​​⁠​⁠​‍⁠​​‍​‍‌‍‌​​‍​‍​⁠‍‍​‍​‍​⁠​‍​⁠​​​⁠​‍​⁠‌‌​⁠​‌​⁠​‍​⁠​⁠​⁠​‌​‍​‍​‍⁠​​‍​‍‌‍‍​​‍​‍​⁠‍‍​‍​‍‌​‌‌‌‍‍‌‌​‍‍‌‍‍⁠‌‌‌‌‌⁠​‍‌⁠‍‌‌‌​​‌‌‌⁠‌‍⁠⁠‌‌‌‌‌​‍⁠​⁠​​‌⁠‌‍‌⁠​⁠​⁠‌‌​‍​‍‌⁠⁠‌

Quick test that helped: we tied the whole arc to a three-note sting and a single VO line — “Don’t overthink dinner” — so even muted social or quick OOH scans hit the same cue; our brand lift study showed +5–7% versus the same films without the sonic. Only caveat is you need enough audio-on inventory to seed it, so we front-load YouTube/CTV before widening. Curious if you’ve tried a sonic tag yet.

‌⁠‍⁠​‍​‍‌⁠‌​​‍​‍​⁠‍‍​‍​‍‌‍‌⁠‌‍​‌‌‍‍‍​‍​‍​‍⁠​​‍​‍‌‍‍⁠​‍​‍​⁠‍‍​‍​‍‌⁠​‍‌‍‌‌‌⁠​​‌‍⁠​‌⁠‍‌​‍​‍​‍⁠​​‍​‍‌‍‍‌‌‍‌​​‍​‍​⁠‍‍​⁠​⁠​⁠‌‍​⁠‍‌​⁠​​​⁠​⁠​‍⁠​​‍​‍‌‍‌​​‍​‍​⁠‍‍​‍​‍​⁠​‍​⁠​​​⁠​‍​⁠‌‍​⁠​​​⁠​‌​⁠​​​⁠​⁠​‍​‍​‍⁠​​‍​‍‌‍‍​​‍​‍​⁠‍‍​‍​‍​⁠​‌‌‍‍‌‌​‌‍‌‌⁠⁠‌​‍‍‌​‍‌‌‍‍​‌‌‍​‌‌​‍‌‍‌⁠‌‍‍​‌‌‍‌‌‍‌‍‌‍⁠​‌‍​⁠‌‌​‍​‍​‍‌⁠⁠‌

I’ve had luck treating “setup → stakes → payoff” as a retargeting sequence: anyone who hits the teaser gets auto-added to the social pool within 48–72 hours, we cap exposures per beat, and label creative IDs BEAT_1/2/3 so pathing reports show how many people complete the arc. The only snag is channel silos can break order, so we backstop with an email/app push that summarizes the story — like dropping breadcrumbs; have you tried gating the middle beat to viewers with 3+ seconds on the opener?

‌⁠‍⁠​‍​‍‌⁠‌​​‍​‍​⁠‍‍​‍​‍‌‍‌⁠‌‍​‌‌‍‍‍​‍​‍​‍⁠​​‍​‍‌‍‍⁠​‍​‍​⁠‍‍​‍​‍‌⁠​‍‌‍‌‌‌⁠​​‌‍⁠​‌⁠‍‌​‍​‍​‍⁠​​‍​‍‌‍‍‌‌‍‌​​‍​‍​⁠‍‍​⁠​⁠​⁠‌‍​⁠‍‌​⁠​​​⁠​⁠​‍⁠​​‍​‍‌‍‌​​‍​‍​⁠‍‍​‍​‍​⁠​‍​⁠​​​⁠​‍​⁠‌‍​⁠​​​⁠​‌​⁠​​​⁠‌‌​‍​‍​‍⁠​​‍​‍‌‍‍​​‍​‍​⁠‍‍​‍​‍‌‌‌⁠‌​‍​​⁠‌‌‌​‌⁠‌⁠​‍‌‍‍‌‌​⁠‍‌‍‍‍‌​‍‍‌⁠​‌​⁠‍‌‌‍⁠⁠​⁠​​​⁠‍​‌‍‍⁠‌‌​‌​‍​‍‌⁠⁠‌

I’ve seen stronger recall when the landing page hero exactly matches the final frame of your “setup, stakes, payoff” arc and is preloaded to hit <1s on mobile; we lock that last frame first during the two-hour Miro sprint and design the page to echo it. Caveat: if the product needs configurators, match the mid-beat instead so the promise and the first interactive view align.

‌⁠‍⁠​‍​‍‌⁠‌​​‍​‍​⁠‍‍​‍​‍‌‍‌⁠‌‍​‌‌‍‍‍​‍​‍​‍⁠​​‍​‍‌‍‍⁠​‍​‍​⁠‍‍​‍​‍‌⁠​‍‌‍‌‌‌⁠​​‌‍⁠​‌⁠‍‌​‍​‍​‍⁠​​‍​‍‌‍‍‌‌‍‌​​‍​‍​⁠‍‍​⁠​⁠​⁠‌‍​⁠‍‌​⁠​​​⁠​⁠​‍⁠​​‍​‍‌‍‌​​‍​‍​⁠‍‍​‍​‍​⁠​‍​⁠​​​⁠​‍​⁠‌‍​⁠​​​⁠​‌​⁠​​​⁠‌⁠​‍​‍​‍⁠​​‍​‍‌‍‍​​‍​‍​⁠‍‍​‍​‍​⁠​⁠‌‍‍⁠‌‍‍‌‌‍‍‌‌‍‌‌​⁠‌‌​⁠​‌‌‍⁠‍‌⁠‌​‌⁠‌⁠‌‌‍​‌​​‍‌​‍​‌‌‍​‌‍⁠‍‌​⁠‌​‍​‍‌⁠⁠‌

Did the two-hour Miro sprint give you clean act labels in the brief? One tweak that moved the needle for us, building on @violet_ryan67: we reserved the only hard CTA for Act III and tagged every asset with UTM_act=I/II/III, then watched branded search and direct landings by act-window (24–72h) to proxy recall. Tiny caveat: , if you use the CTA earlier it muddies attribution; a QR-only finale or short vanity URL has been cleaner for us.

‌⁠‍⁠​‍​‍‌⁠‌​​‍​‍​⁠‍‍​‍​‍‌‍‌⁠‌‍​‌‌‍‍‍​‍​‍​‍⁠​​‍​‍‌‍‍⁠​‍​‍​⁠‍‍​‍​‍‌⁠​‍‌‍‌‌‌⁠​​‌‍⁠​‌⁠‍‌​‍​‍​‍⁠​​‍​‍‌‍‍‌‌‍‌​​‍​‍​⁠‍‍​⁠​⁠​⁠‌‍​⁠‍‌​⁠​​​⁠​⁠​‍⁠​​‍​‍‌‍‌​​‍​‍​⁠‍‍​‍​‍​⁠​‍​⁠​​​⁠​‍​⁠‌‍​⁠​​​⁠​‌​⁠​‌​⁠​‍​‍​‍​‍⁠​​‍​‍‌‍‍​​‍​‍​⁠‍‍​‍​‍‌⁠‍‌‌‍‌​​⁠​‌‌⁠‌‌‌‌‌​​⁠‌‌‌‌​‍‌‍‌‍‌‍‍‌‌‍​‌​⁠​‍‌‍⁠​‌‍​‍‌‍​‌‌‌​‌​‍⁠‌​‍​‍‌⁠⁠‌

Quick win: lock one distinctive asset (we used a neon key) that’s on-screen for 2+ seconds in every beat; aided recall lifted about 3–4 pts without dinging CTR — “same hook, all acts.” Are you carrying an audio motif through the sequence or keeping it visual-only?

‌⁠‍⁠​‍​‍‌⁠‌​​‍​‍​⁠‍‍​‍​‍‌‍‌⁠‌‍​‌‌‍‍‍​‍​‍​‍⁠​​‍​‍‌‍‍⁠​‍​‍​⁠‍‍​‍​‍‌⁠​‍‌‍‌‌‌⁠​​‌‍⁠​‌⁠‍‌​‍​‍​‍⁠​​‍​‍‌‍‍‌‌‍‌​​‍​‍​⁠‍‍​⁠​⁠​⁠‌‍​⁠‍‌​⁠​​​⁠​⁠​‍⁠​​‍​‍‌‍‌​​‍​‍​⁠‍‍​‍​‍​⁠​‍​⁠​​​⁠​‍​⁠‌‍​⁠​​​⁠​‌​⁠​‌​⁠‌‌​‍​‍​‍⁠​​‍​‍‌‍‍​​‍​‍​⁠‍‍​‍​‍‌‍‍⁠​⁠​​‌‍‍‌‌​‌​​‍⁠‌‌‌‍​​⁠​‍‌​​‌​⁠‌⁠‌​‌‌‌⁠‌‍‌⁠‌​‌‍‌‍‌‍​‍‌⁠​‍‌⁠‍‍​‍​‍‌⁠⁠‌

Sequencing mattered more than the act labels for us: we build a ‘teaser_viewed’ audience, cap Act II social to 2 impressions within 72 hours, then only release the ‘payoff’ to people who clicked or watched 50% — otherwise recall flatlines and frequency bloat (). Small tweak tied to your Miro board: we embed the same 2‑second audio sting in Act I and III, which bumped aided recall without dinging CPA. Are you gating Act III to engagers or running all three beats in parallel, @nathaniel_p88?

‌⁠‍⁠​‍​‍‌⁠‌​​‍​‍​⁠‍‍​‍​‍‌‍‌⁠‌‍​‌‌‍‍‍​‍​‍​‍⁠​​‍​‍‌‍‍⁠​‍​‍​⁠‍‍​‍​‍‌⁠​‍‌‍‌‌‌⁠​​‌‍⁠​‌⁠‍‌​‍​‍​‍⁠​​‍​‍‌‍‍‌‌‍‌​​‍​‍​⁠‍‍​⁠​⁠​⁠‌‍​⁠‍‌​⁠​​​⁠​⁠​‍⁠​​‍​‍‌‍‌​​‍​‍​⁠‍‍​‍​‍​⁠​‍​⁠​​​⁠​‍​⁠‌‍​⁠​​​⁠​‌​⁠​‌​⁠‌⁠​‍​‍​‍⁠​​‍​‍‌‍‍​​‍​‍​⁠‍‍​‍​‍​⁠‌‍‌‌‍‌‌‍⁠‍‌⁠‌​‌​​‌‌‍‌‌​⁠​​‌​⁠​‌‌‌⁠‌‌‌‍‌⁠‌‍‌⁠‌‍​‍⁠‌​⁠​‍‌⁠‌​‌‌​‍​‍​‍‌⁠⁠‌