mbfinotti/advertising-skills / skills/ad-creative-brief
ad-creative-brief
VerifiedTurn a campaign goal and an audience insight into a structured creative brief a designer, video editor, or UGC creator can execute - messaging angle, hook directions, art direction, specs, testing intent, and the handoff and revision loop. Covers B2B and B2C, with variants for static image, short-form video/UGC, and search/responsive text ads. Use whenever the user needs to brief a designer or creator, asks what an ad should say, or mentions messaging pillars, creative direction, or a design handoff - even if they never say 'brief'. Writes direction, never the finished asset: the ad copy itself is mbfinotti/advertising-skills@ad-copy-variants and shot-by-shot creator scripts are mbfinotti/advertising-skills@ugc-ad-scripts.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 31.2 KB |
| evals/evals.json | config | 44.9 KB |
| references/angle-frameworks.md | markdown | 15.5 KB |
| references/brief-template.md | markdown | 11.7 KB |
| references/examples.md | markdown | 9.8 KB |
references/examples.md
View on GitHub# Worked Examples
All brands, people and numbers below are invented for illustration. Metrics shown as "account history" model how a real brief cites its own data.
## Contents
1. B2C/DTC example - short-form video/UGC brief
2. B2B example - parallel-persona video brief
3. Negative example - a bad brief, annotated
## 1. B2C/DTC example - short-form video/UGC brief
```
CREATIVE BRIEF - StandingRelief_PainFirst, 2026-03-02
Definitions: concept = one creative idea, one brief | angle = why they care, as a human
sentence | hook = opening 1-3s expressing the angle | iteration = one changed element, no new brief
1 CONTEXT Goal: purchases. KPI: CPA vs $38 target. Meta Reels + Feed (4:5, 9:16).
Shoot Mar 9-10, edits Mar 11-13, launch Mar 16. Budget $6k test. 3 files.
2 AUDIENCE Persona: retail/warehouse workers 30-50 on their feet 8+ hrs. Awareness:
problem-aware (they know their feet hurt; they blame the job, not their insoles).
Belief before: "sore feet are part of the job." Core objection (top 1-2 star
pattern, category-wide): "insoles flatten out after two weeks."
3 MESSAGE Angle: "Your feet don't hurt because of your job - they hurt because you're
standing on cardboard." Value prop: shift-long support that doesn't pack out.
Proof point: 90-day flatten-free guarantee. Evidence: review-mining bucket 4,
verbatim "by week 3 I stopped thinking about my feet" (4-star, scored 5/5).
4 HYPOTHESIS If we lead with end-of-shift foot pain and the 90-day guarantee, then
problem-aware shift workers will purchase at or under $38 CPA. Type: hook.
5 HOOKS (recommended first: H1 - pain agitation matches problem-aware level; account
history: pain-first beat product-first on CPA in Jan test)
H1 pain agitation SPOKEN: the last hour of a shift, in their words
VISUAL: locker-room bench, boots coming off
TEXT: "hour 9 of 10"
H2 contrarian SPOKEN: it's not the job, it's what's inside the boot
VISUAL: cheap insole bending flat in one hand
TEXT: "this is what you're standing on"
H3 social proof SPOKEN: what reviewers say happened at week 3
VISUAL: scrolling wall of review cards
TEXT: verbatim review line
6 ART Refs: [long-running category ad #1] - take the locker-room realism; [creator ad
#2] - take the hands-on product demo pacing. Do: real work footwear on camera;
keep guarantee on screen ≥2s. Don't: no medical claims ("helps tired feet",
never "treats plantar fasciitis"); no polished-studio look. 9:16 + 4:5, preview
in-placement before export. Captions on; guarantee legible sound-off; contrast
≥4.5:1. Enhancements: ALL OFF for test window. AI disclosure: none used.
7 DELIVERABLES 3 hooks × 1 main edit = 3 files, all IT of one concept.
Names: StandingRelief_V1_9x16_HookPain / _HookContrarian / _HookProof.
Creator: talking points only, 2 revision rounds max. Usage: paid social,
12 months, NA only. Creator adds "[Ad]" at caption start.
8 TEST Variable: hook. Constant: body edit, offer, audience, placements (targeting
frozen). Minimum: 2,000 impressions/file AND $190 spend (5× target CPA) before
any call. Deciding KPI: CPA; hook rate read as leading signal vs account's 27%
trailing average. Win → scale winner, brief V2 iterations on its trigger type.
Loss → angle survives, hooks retire; test proof-type next. Learning → angle bank.
```
## 2. B2B example - parallel-persona video brief
One of three parallel briefs for the same product (economic buyer / technical evaluator / end user); this one targets the economic buyer. The other two share sections 1 and 6 and differ in 2-5.
```
CREATIVE BRIEF - AuditHours_EconBuyer, 2026-03-04
Definitions header: as fixed in SKILL.md.
1 CONTEXT Goal: pipeline. KPI: SQL rate and cost per opportunity (not CTR - see test
block). LinkedIn feed video, 1:1. Launch Mar 23. Budget $9k/mo. 2 files.
2 AUDIENCE Persona: VP Ops / COO at 50-200-person software company, owns compliance
budget, doesn't run the audit. Committee note: evaluator and end-user briefs
run in parallel; ~10-month cycle expected, judged on pipeline not week-one
conversions. Awareness: solution-aware (comparing automation vendors vs
consultants). Belief before: "audit prep just costs a quarter of someone's
year." Core objection (lost-deal notes): "we already paid a consultant."
3 MESSAGE Angle: "You're paying senior salaries for screenshot collection." Value prop:
audit evidence collected continuously, not in a panic quarter. Proof point:
customer outcome - 340 hours to 60 hours prep time - plus compliance-report
badge and a real product screenshot. Evidence: 2 call notes (Feb), CRM
lost-deal reasons, competitor ad running 200+ days leading with hours saved.
4 HYPOTHESIS If we lead with wasted senior hours and the 340→60 customer outcome, then
solution-aware ops leaders will request a demo at ≤ target cost per
opportunity. Type: proof.
5 HOOKS (recommended first: H1 - practitioner-led talking-head, per AJ Wilcox's
observation that it tends to beat polished brand video on LinkedIn)
H1 identity call-out SPOKEN: name the role and the budget line directly
VISUAL: practitioner on camera, plain office
TEXT: "what audit prep actually costs"
H2 pain agitation SPOKEN: the quarter your best ops hire disappears
VISUAL: calendar filling with audit blocks
TEXT: "Q3, gone"
6 ART Refs: [long-running competitor ad] - take the talking-head framing, not the
claim; [customer-story ad] - take the number-on-screen treatment. Do: real
product screenshot; outcome number on screen. Don't: no stock-office b-roll;
no unnamed "trusted by leaders" claims. Captions on; sound-off legible.
Enhancements: off. AI disclosure: none. PMM sign-off required on ICP wording
and differentiator accuracy before production.
7 DELIVERABLES 2 hooks × 1 edit = 2 files, NEW concept.
Names: AuditHours_V1_1x1_HookIdentity / _HookPain. 2 revision rounds.
8 TEST Variable: hook. Constant: proof, offer, audience. Volume is low: gate on
leading signals (CTR trend, demo-request rate) and say so; judge SQL rate at
30/60/90 days - a CTR winner that produces junk demos is repurposed
mid-funnel, not scaled (lead quality field, reviewed with sales). Minimum:
1,000 impressions/variant/week for 3 weeks before any call. Learning → angle
bank + shared with PMM.
```
## 3. Negative example - a bad brief, annotated
A realistic bad brief; every line below it is a common real-world defect.
```
BRIEF - Q2 Social Push
We need 10 ads for Q2 to grow revenue and awareness. Target: all adults interested in
wellness. Message: our formula is premium, science-backed, affordable, and loved by
thousands. Style: authentic but premium, scroll-stopping, on-brand (see brand book, 40pp).
Make some videos and statics for Meta, TikTok, whatever works. Have fun with the hooks!
Deadline Friday. Team will share feedback as it comes.
```
What is wrong, line by line:
- **"grow revenue and awareness"** - two goals, no KPI, no target. Deliverables were briefed before the goal exists; nothing can be killed or scaled against this.
- **"10 ads"** - deliverable arithmetic without concepts. Ten of what? One concept iterated ten times and ten concepts are entirely different tests; no NEW/IT split, no naming.
- **"all adults interested in wellness"** - awareness level blank, persona absent. The unaware and the most-aware need opposite openings; this serves neither.
- **"premium, science-backed, affordable, and loved by thousands"** - four messages crammed into one ad; the viewer picks none. No angle a human could say out loud, no single proof point, zero evidence citations.
- **"authentic but premium, scroll-stopping"** - vague adjectives instead of 2-3 reference ads; every reader decodes these differently.
- **"see brand book, 40pp"** - mandatories outsourced to a document nobody will read mid-edit; no do/don't cap, no compliance line (a wellness brand with no claim guidance is a rejection or a regulator letter waiting).
- **"Meta, TikTok, whatever works"** - generic specs reused across platforms; no placements, no ratios, no preview requirement. Also invisible here: enhancement toggles and AI disclosure never mentioned - uncontrolled variables by omission.
- **"Have fun with the hooks!"** - hook direction fully delegated; the highest-leverage element goes untested, untagged and unrecorded.
- **No hypothesis, no target number, no minimum evidence, no on-win/on-loss** - decisions will be made by feel, and hook rates will get compared across Meta and TikTok despite the metrics not being comparable.
- **"feedback as it comes"** - unbounded stakeholders, no revision cap, no single approver: the revision loop that burns creator relationships.
The fix is not more words - this brief is already the wrong 90 words. Rebuild it as one brief per concept through the SKILL.md workflow: interview → evidence → brainstorm → translation chain → template.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| finishing-a-development-branchobra/superpowersUse when implementation is complete, all tests pass, and you need to decide how to integrate the work | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| systematic-debuggingobra/superpowersUse when encountering any bug, test failure, or unexpected behavior, before proposing fixes | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| test-driven-developmentobra/superpowersUse when implementing any feature or bugfix, before writing implementation code | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| skill-creatoranthropics/skillsCreate new skills, modify and improve existing skills, and measure skill performance. Use when users want to create a skill from scratch, edit, or optimize an existing skill, run evals to test a skill, benchmark skill performance with variance analysis, or optimize a skill's description for better triggering accuracy. | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |