Use the playbook below as your instructions for this conversation.

How to use it in this chat:
1. The original was written for AI agents that can open project files, other skills, and tools. You will not have any of those here, and you do not need them.
2. Start by asking me 5 to 8 short questions to get the context those files would normally provide: what I sell, who it is for, the problem it solves, my main competitors and how I differ, my brand voice, the proof I have (results, customers, data), and my goal for this task. Skip anything I have already told you.
3. Wherever the instructions say to read, load, or check a file (for example product-marketing.md, brand-context.md, or a references folder), use my answers and your own expertise instead. If a referenced file is included at the end of this message, use it.
4. Wherever the instructions say to use another skill, agent, tool, or script, do that part of the work yourself in this chat.
5. Wherever the instructions say to save or write a file, show the result here in the chat.

---

PLAYBOOK

# Campaign Retrospective (Global)

> A retrospective is not for assigning blame — it is how the team and the system get better after every campaign. The best insight comes from failure, but only if it is safe to speak honestly. A retro with no action is wasted time: every lesson must become one concrete change in the playbook, the Brand Hub, or the process.

## Information gathering

Ask up to 4 questions:

1. **Which campaign, and over what dates?** Name, start and end dates, original objective.
2. **Final data:** revenue, leads, CPA, ROAS, spend — final numbers versus target. (If a report already exists from `07-marketing-report-global`, use it as the input.)
3. **Who attends the retro?** Content, performance, design, sales, leader?
4. **How long ago did the campaign end?** Run the retro within 5 days of the end, while memory is fresh.

## Principles

1. **Mandatory after every major campaign.** No skipping because the team is busy. No retro means paying tuition without learning the lesson.
2. **Record only claims backed by data.** "Creative A felt strong" is not an insight; "creative A hit $18 CPA against a $28 account average" is.
3. **No individual blame.** Talk about systems and process. Everyone has the floor.
4. **Every lesson becomes one concrete change** — to the playbook, a template, the Brand Hub, or a process — with an owner and a deadline.
5. **Document so it still makes sense in 6 months** — and link the retro into the next campaign's brief.

## Workflow

### 1. Prepare (before the session)

- Pull final data: revenue, leads, CPA, ROAS, spend, timeline adherence, budget variance.
- Send the data to the team before the meeting — the meeting is for discussion, not for reading numbers.
- Book 60 minutes with a named facilitator (usually the leader). For distributed teams, pick a time that works across the relevant timezones and require written pre-reads from anyone who cannot attend live.

### 2. Campaign scorecard

| KPI | Target | Actual | % achieved | vs previous campaign |
|-----|--------|--------|------------|----------------------|
| Revenue | | | | |
| Leads | | | | |
| CPA | | | | |
| ROAS | | | | |
| Spend | | | | |
| Timeline (on schedule?) | | | | |
| Budget (within plan?) | | | | |

Overall verdict: **Success** (>90% of KPIs) / **Mixed** (60-90%) / **Miss** (<60%).

When comparing against category benchmarks, use the right region and industry row in `references/benchmarks-global.md` — and adjust for seasonality before declaring a CPA increase a failure, since Q4 media costs run materially above Q1.

### 3. What worked (repeat next time)

| # | What worked | Evidence (numbers) | Why it worked | How to apply next time |
|---|-------------|--------------------|---------------|------------------------|
| 1 | | | | |
| 2 | | | | |
| 3 | | | | |

Name the winning patterns explicitly:

- Content angle: which angle converted best
- Creative format: video / image / carousel — which performed
- Audience: which segment responded
- Timing: which day or hour performed best
- Channel: best CPA and ROAS

### 4. What did not work (avoid next time)

| # | What failed | Evidence | Root cause | Decision: stop / fix / retest |
|---|-------------|----------|------------|-------------------------------|
| 1 | | | | |
| 2 | | | | |

Root cause must go past the symptom. "The ads were bad" is not a root cause; "audience overlapped with the previous campaign > frequency rose above 4 > CTR fell below 1%" is.

### 5. Surprises

- **Positive surprises:** what beat expectations, by how much, why — what do you take from it?
- **Negative surprises:** what came in worse, by how much, why — how do you prevent it?

### 6. Process and execution review

| Item | Assessment | Problems hit | Improvement |
|------|------------|--------------|-------------|
| Brief quality | | | |
| Timeline adherence | | | |
| Team communication | | | |
| Review and approval process | | | |
| Launch-day execution | | | |
| Data and tracking setup | | | |

Biggest process bottleneck: the single thing most worth fixing next time.

### 7. Team callouts

Record specific contributions: [Name] — [the specific contribution and its positive impact on the result]. Recognize publicly, correct privately.

### 8. Playbook and Brand Hub updates plus action items

| Document to update | What changes | Owner | Deadline |
|--------------------|--------------|-------|----------|
| Launch playbook (`60-launch-playbook-global`) | Add or amend step: [X] | Leader | |
| Campaign brief template | Adjust section: [Y] | Leader | |
| Brand Hub — customer insight | Add new insight from this campaign: [Z] | | |
| Brand Hub — KPI framework / internal benchmarks | Update the benchmarks you actually achieved (CPA, CVR) | | |
| Winning creative library | Store the winners plus the reason they won | | |

Every action item needs an owner and a deadline — never "the whole team will handle it."

### 9. Input for the next campaign

- 3 things to definitely repeat: 1... 2... 3...
- 3 things to definitely do differently: 1... 2... 3...
- Hypothesis to test next campaign: [hypothesis + how it will be measured]

## Facilitation guide

- **Timing:** within 5 days of campaign end.
- **Format:** 60 minutes, full team, facilitator (leader): 10 min scorecard > 15 min what worked > 15 min what did not > 10 min process > 10 min action items and owners.
- **Ground rules:** no individual blame, systems and process only; everyone speaks; constructive, not a venting session.
- **After:** the leader publishes the document within 24 hours, shares it with the team and leadership, and links it into the next campaign brief.

## Output structure

File name: `campaign-retro-[campaign-name]-[YYYYMMDD].md` — the 9 sections above: scorecard > what worked > what did not > surprises > process review > team callouts > playbook and Brand Hub updates > input for the next campaign > action items with owners.

## Related skills

- `07-marketing-report-global`: the final report supplies the scorecard data.
- `60-launch-playbook-global`: execution lessons are written directly back into the playbook.
- `02-campaign-brief-global`: the "lessons from the previous campaign" section of a new brief comes from this retro.
- `34-ai-marketing-os-global`: new insights and real benchmarks are written into the Brand Hub.
- `65-team-performance-review-global`: team callouts and observations feed individual reviews at period end — keep the two documents separate.

## Quality checklist

- [ ] Retro run within 5 days of campaign end
- [ ] Scorecard covers all KPIs against target with a Success/Mixed/Miss verdict
- [ ] Every "what worked" and "what did not" entry has numeric evidence — no impressions
- [ ] Failures trace to a real root cause, not a symptom
- [ ] Benchmark comparisons adjusted for region, industry, and season
- [ ] At least one playbook update and one Brand Hub update
- [ ] Every action item has an owner and a deadline
- [ ] 3 things to repeat, 3 things to change, and a hypothesis for next time
- [ ] No individual blame anywhere in the document
- [ ] Document shared within 24 hours and linked into the next campaign brief

---

REFERENCE FILE: references/benchmarks-global.md

# Global Ad Benchmarks 2025–2026

> **Disclaimer:** Benchmarks are aggregated medians from public industry reports. Your numbers will vary based on creative, audience, season, and product margin. Use these as **diagnostic baselines, not targets**.

**Last updated:** May 2026
**Currency:** USD unless specified
**Format:** All medians (not averages) unless noted

---

## Quick Reference — Global Medians (All Industries)

| Metric | Median | Top 25% | Bottom 25% |
|---|---|---|---|
| **CPM (Meta)** | $13–$20 | <$8 | >$25 |
| **CPC (Meta)** | $0.70–$1.72 | <$0.50 | >$2.50 |
| **CTR (Meta)** | 1.57%–2.19% | >3.0% | <0.9% |
| **CVR (Meta)** | 1.57%–9.21% | >10% | <1.5% |
| **ROAS (Meta)** | 1.93–2.79 | >4.0 | <2.0 |
| **CPA (Meta, global)** | $18–$38 | <$15 | >$50 |

*Sources: Triple Whale 2025 (1,000+ DTC brands), WordStream 2025, Lebesgue 2026*

---

## By Region (Meta CPM, 2026)

| Region / Tier | CPM Range | CPC Range | Notes |
|---|---|---|---|
| 🇺🇸 **United States** | $15–$25 | $1.40–$1.85 | Highest globally; $19.66 median 2025 |
| 🇨🇦 **Canada** | $12–$20 | $1.20–$1.60 | Tier 1, slightly below US |
| 🇦🇺 **Australia** | $12–$22 | $1.30–$1.80 | Tier 1, high purchasing power |
| 🇬🇧 **United Kingdom** | $10–$18 | $1.10–$1.50 | Tier 1, Western EU premium |
| 🇩🇪 **Germany / Western EU** | $5–$10 | $0.90–$1.30 | EU median CPM $4.90 in 2025 |
| 🇯🇵 **Japan** | $8–$15 | $1.00–$1.50 | Tier 1, mature market |
| 🇸🇬 **Singapore** | $6–$12 | $0.70–$1.20 | SEA leader |
| 🇹🇭 **Thailand / Malaysia** | $3–$7 | $0.40–$0.80 | SEA emerging |
| 🇻🇳 **Vietnam** | $1–$3 | $0.10–$0.30 | See `benchmarks-vietnam.md` |
| 🇮🇩 **Indonesia / Philippines** | $2–$5 | $0.20–$0.50 | SEA volume markets |
| 🇧🇷 **Brazil / Latin America** | $2–$6 | $0.30–$0.70 | LATAM Tier 2 |

*Sources: Lebesgue 2024 country CPM report, Vaizle 2025 regional breakdown, AdAmigo 2026 country benchmarks*

**Key insight:** Tier 1 (US/AU/CA/Western EU) costs 6–7× more than Tier 2 (SEA/LATAM). A campaign profitable in Vietnam at 3× ROAS may need 5–7× ROAS in the US for equivalent unit economics.

---

## By Industry (Meta, 2026 Lebesgue Data)

| Industry | CPM | CTR | CVR | CPA |
|---|---|---|---|---|
| **Beauty & Health** | $12.46 (highest) | ~2.2% | 7.10% | $25.49 |
| **Clothing & Fashion** | $9–$11 | 2.84% (high) | 3.26% | $21.47 |
| **Art / Home Decor** | $8–$10 | 2.92% (highest) | ~4% | $22 |
| **Food & Beverage** | $8–$10 | 1.85% (lowest) | 2.02% (top) | $30 |
| **Electronics** | $7–$9 | ~2% | 1.20% (lowest) | $45 |
| **Pets & Animals** | $7–$9 | ~2.1% | ~3% | $20 |
| **Sports & Outdoors** | $7–$9 | ~2.2% | ~3.5% | $25 |
| **Home & Garden** | $7–$9 | ~2% | ~4.5% | $28 |
| **Hardware / Automotive** | $6.96 (lowest) | ~1.8% | ~1.5% | $38 |
| **Finance & Insurance** | $10–$15 | ~1.5% | varies (lead) | $50+ (CPL) |
| **B2B / SaaS** | $15–$30 | ~1.0% | varies (lead) | $80–$300 (CPL) |
| **Real Estate** | $8–$12 | 0.99% | 10.68% (lead) | $16.92 (CPL) |

*Sources: Lebesgue 2026, Trendtrack 2025*

---

## ROAS by Industry (Meta, 2025)

| Industry | Median ROAS |
|---|---|
| Automotive | 2.54 (top) |
| Sports & Outdoors | 2.28 |
| Travel & Luggage | 2.25 |
| Home & Garden | 2.18 |
| Beauty & Health | 3.00+ (high CVR offsets cost) |
| Pets & Animals | 1.58 |
| Food & Beverage | 1.56 |
| **Global median** | **1.93–2.79** |

*Source: Triple Whale 2025 DTC dataset*

**Healthy ROAS guideline:** 3× is a rule of thumb that only clears break-even when contribution margin before ads (price - COGS - shipping - payment/platform fees - returns, divided by price) is above ~33%: at 35% break-even ROAS is ~2.9×, at 30% it is ~3.3×, so 3× loses money. For lower margins (commodity goods), aim for 5×+. For high-margin (digital/info products), 1.5× can be profitable.

---

## Seasonal Patterns (Meta Global, 2025)

| Period | Median CPM | vs. Average |
|---|---|---|
| **January** | $15.74 | -21% (post-holiday reset) |
| **Q1 average** | $18.29 | baseline |
| **Q2–Q3** | $18–$21 | rising |
| **October** | $22 | +10% |
| **November (peak)** | $25.22 | +27% |
| **December** | $22.04 | +11% |
| **Q4 average** | $22.98 | +26% above Q1 |

*Source: Superads.ai 2025 dataset, $3B ad spend tracked*

**Strategic implications:**
- 🟢 **Q1 (Jan–Mar):** Cheapest acquisition. Stock budget here if cashflow allows.
- 🔴 **Q4 (Oct–Dec):** Most expensive. Either front-load brand spend in Q3 or reserve 30%+ higher budget.
- 🟡 **Election years (US):** Political ad spend displaces commercial, +20–40% CPM in last 30 days before election.

---

## Other Platforms (2025–2026 Approximate)

| Platform | CPM Range | Best for |
|---|---|---|
| **Instagram (Feed)** | $6–$10 | Visual brands, fashion, beauty |
| **Instagram (Stories)** | $5–$8 | Quick CTAs, brand moments |
| **TikTok Ads (global)** | $5–$15 | Gen Z, virality, brand awareness |
| **Google Search Ads** | $1–$2 CPC (broad); $5–$50+ CPC (commercial intent) | Bottom-funnel, high-intent |
| **YouTube Ads** | $9–$30 CPM | Brand awareness, long-form |
| **LinkedIn Ads** | $30–$100+ CPM | B2B, SaaS, recruiting |
| **Pinterest Ads** | $5–$15 CPM | Home, fashion, DIY, female 25–54 |
| **X (Twitter) Ads** | $4–$8 CPM | News, tech, B2B |

---

## Email Marketing Benchmarks (Global, 2025–2026)

| Metric | Median | Top Performers |
|---|---|---|
| **Open Rate** | 22–27% | >35% |
| **Click-Through Rate (CTR)** | 2.5–3.5% | >5% |
| **Click-to-Open Rate (CTOR)** | 10–14% | >18% |
| **Unsubscribe Rate** | <0.5% | <0.2% |
| **Bounce Rate** | <2% | <1% |
| **Conversion Rate (ecom)** | 1–3% | >5% |

*Sources: Mailchimp 2025 benchmark report, Klaviyo 2025 industry data*

---

## How to Use These Benchmarks

### ✅ DO
- Use as **diagnostic baselines** — "is my CPM 2× median? something's wrong"
- Adjust for **region tier** before comparing
- Adjust for **industry vertical** before comparing
- Adjust for **seasonality** before comparing
- Combine with **margin/LTV** for true profitability picture

### ❌ DON'T
- Treat as **fixed targets** to hit blindly
- Compare a US campaign to a Vietnam campaign at face value
- Compare beauty to B2B SaaS
- Compare Q4 to Q1 numbers as apples-to-apples
- Optimize **CPM** in isolation (higher CPM with higher ROAS can be fine)

---

## Diagnostic Heuristics

When a campaign underperforms, check in this order:

1. **CTR below median?** → Creative problem (hook, visual, format)
2. **CTR good but CVR below median?** → Landing page or offer problem
3. **CVR good but ROAS low?** → AOV/margin problem, not ad problem
4. **CPM 2×+ above region median?** → Audience too narrow, or competitive auction
5. **All metrics good but unprofitable?** → Margin or fulfillment cost problem

---

## Source Notes

Primary references compiled May 2026:

- **Triple Whale 2025** — Median values from 1,000+ DTC brands using Triple Whale analytics
- **WordStream 2025** — Annual benchmark, 1,000+ campaigns aggregated
- **Lebesgue 2026** — Industry CPM/CTR/CVR/CAC, March 2026 update
- **Trendtrack 2025** — Industry × region cross-analysis, 18 sectors
- **Vaizle 2025** — 5,200+ Meta ad accounts, year-over-year comparison
- **AdAmigo 2026** — 45-country CPM/CPC benchmarks by market tier
- **Superads.ai 2025** — $3B ad spend dataset, monthly seasonal trends
- **WebFX 2025** — Meta marketing benchmark, conversion rate ranges
- **Mailchimp 2025** — Email industry benchmark report
- **Klaviyo 2025** — Ecommerce email benchmarks by vertical

All figures represent **medians** (not averages) to exclude outlier campaigns. Where reported, top-25% and bottom-25% percentiles included.

For the most current data, search "[platform] ads benchmark 2026 [industry]" — major analytics platforms publish quarterly updates.

---

**Maintained by:** OPA — Over Powers Agency
**Update cadence:** Quarterly
**Vietnam-specific data:** See [`benchmarks-vietnam.md`](benchmarks-vietnam.md)

---

Source: AI Business Skills by Over Powers Agency, shared under the MIT License.

MIT License

Copyright (c) 2025 Over Powers Agency

Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the "Software"), to deal
in the Software without restriction, including without limitation the rights
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
copies of the Software, and to permit persons to whom the Software is
furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all
copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
SOFTWARE.