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

# Budget Planning (Global)

> A budget is not a fixed number — it is a decision tool: where you place bets and when you pull back. The leader must know total spend, ROI per channel, what to increase, and what to cut. Run `10-reverse-kpi-global` first to derive the minimum budget from the revenue target before allocating anything.

## Information gathering

Read `.agents/product-marketing-context-global.md`, the marketing plan, and the previous period's report if available. If missing, ask up to 4 questions:

1. **Total budget and planning period?** How much, for a month, a quarter, or a year (USD)?
2. **Revenue target for the period?** Needed to compute marketing as a percent of revenue and the minimum viable budget.
3. **Active channels and last period's ROI?** Meta / TikTok / Google / email + SMS / affiliates / agency — which has the best ROAS or CPA?
4. **Any large campaign needing a reserve?** Launch, Black Friday / Cyber Monday, or another seasonal peak.

## Principles

1. **Budget follows performance.** Do not stay rigid to the opening plan. Channels that work get more; channels that do not get cut.
2. **Always hold 10-15% in reserve.** Never allocate everything on day one. The reserve covers unexpected opportunities and incidents.
3. **Every dollar has a measurable KPI.** "Branding you cannot measure" is not an acceptable line item.
4. **Do not spread evenly.** Concentrate on the best-performing channel. New channels get a minimum test allocation (10-15%) first.
5. **Plan backwards from revenue, not from channels.** The budget is the output of a revenue equation, not an arbitrary number.
6. **Do not divide the annual budget evenly across quarters.** Median Meta CPM in Q4 runs roughly 26% above Q1 (`references/benchmarks-global.md`), so identical monthly spend buys materially less reach in Q4. Either front-load acquisition into cheap Q1 months or explicitly reserve a larger Q4 allocation for the same volume.

## Workflow

### 1. Choose the budget-setting method

| Method | How it works | Strength | Weakness |
|--------|--------------|----------|----------|
| Percent of revenue | Budget = X% of expected revenue | Simple | Ignores growth stage |
| Backwards from targets (recommended) | Set revenue > derive CAC/CPL > derive spend | Logical, evidence-based | Needs historical CAC/CPL |
| Competitive benchmark | Match category spend norms | Avoids under-investment | Ignores your own conditions |

Seven-step backwards calculation (full detail in `10-reverse-kpi-global`):

```
1. Target revenue
2. / AOV                        = orders needed
3. / close rate (30-50%)        = hot leads needed
4. / nurture CVR (50-60%)       = warm leads needed
5. / lead capture CVR (20-40%)  = total reach needed
6. total leads x CPL            = minimum ad budget
7. + content + tools + agency   = total marketing budget
```

Build 3 scenarios: Conservative (CVR below category median, CPL above), Base case (at median), Optimistic (post-optimization). Plan against Base; prepare cashflow against Conservative.

### 2. Allocate by bucket

| Bucket | Suggested share | Purpose |
|--------|-----------------|---------|
| Paid ads (performance) | 50-65% | Meta, TikTok, Google, retargeting |
| Content production | 10-15% | Video, design, copywriting |
| Creators / UGC / affiliates | 5-10% | Social proof, third-party reach |
| Agency / freelancers | Per scope | If used |
| Tools / software | 2-5% | Ad tools, design, email/CRM, analytics |
| Campaign reserve | 10-15% | Opportunities and incidents — NOT pre-allocated |

### 3. Allocate by channel

Starting split (adjust to your category and your own data):

| Channel | Share of ad budget | Objective | Expected CPA / ROAS |
|---------|--------------------|-----------|---------------------|
| Meta Ads (Facebook + Instagram) | 45-60% | Lead / purchase | |
| TikTok Ads | 15-20% | Awareness / lead | |
| Google Search + Shopping | 15-25% | High-intent capture | |
| Email + SMS (Klaviyo, Mailchimp, Attentive) | 5-10% | Retention, nurture, retargeting | |
| New channel (test) | 10-15% | Validate | |

For B2B, shift toward LinkedIn and Search — LinkedIn CPM runs $30-100+, so plan a higher CPL and a longer payback window. Allocation logic: channels with strong ROAS or CPA get a larger share; new channels get the minimum test allocation first; never divide evenly.

### 4. Split by funnel and by test vs scale

Standard performance split inside each channel:

| Purpose | Share | Note |
|---------|-------|------|
| Testing (new creative + audiences) | 30% | Always keep testing — creative fatigue arrives fast |
| Scale (confirmed winners) | 50% | Only scale campaigns with winning data |
| Retargeting | 15% | Warm audience — cheapest CPA in the funnel |
| Lookalike / similar audiences | 5% | Expansion from converter seeds |

By funnel: TOFU (reach/awareness), MOFU (lead/nurture), BOFU (conversion/retargeting). The ratio depends on stage: a launch skews TOFU, a mature program skews BOFU and retention.

### 5. Allocate by campaign and month

| Campaign | Month | Budget (USD) | Objective | KPI |
|----------|-------|--------------|-----------|-----|
| Always-on (retargeting + nurture) | Full period | | | |
| Campaign A (launch or seasonal) | | | | |
| Testing budget | Full period | | New channels or angles | |
| Reserve | — | 10-15% | | |

Seasonality overlay before locking monthly numbers:

| Period | Media cost vs Q1 | Planning implication |
|--------|------------------|----------------------|
| January | ~-21% below Q1 average | Cheapest acquisition window — stock budget here if cashflow allows |
| Q2-Q3 | Rising | Normal operating cadence, build audiences for Q4 |
| October | ~+10% | Start of the expensive window |
| November | ~+27% (peak) | Reserve extra budget or accept lower volume |
| Q4 average | ~+26% above Q1 | Do not plan Q4 at Q1 unit costs |

In US election years, political spend displaces commercial inventory and adds roughly 20-40% to CPM in the last 30 days before the vote — build that into Q4 plans for US-targeted campaigns.

### 6. Scale-up and stop-loss rules

**Increase budget when (BOTH conditions hold):**

- ROAS above 1.5x target for 7+ consecutive days, OR CPA at or below 70% of target with sufficient volume
- Audience not yet saturated (frequency below 2.5)
- Increment: +20-30% per step, never a sudden doubling

**Cut or pause when (stop-loss):**

| Condition | Action |
|-----------|--------|
| ROAS below break-even for 14 days after enough data (break-even on contribution margin before ads, not gross margin) | Cut the channel or campaign, move budget to what works |
| CPA above 1.5x target for 7 days | Change creative or audience |
| CPA above 3.0x target once spend >= 3x target CPA (below that, not enough data; house heuristic ladder in `03-performance-eval-global`) | Propose a pause (rule 7 draft block, named approver); replace creative or audience before restarting |
| Frequency above 4 (urgent tier; heuristic) | Audience saturation — stop scaling, refresh creative or expand audience; frequency alone never pauses an ad set |
| Spend above 110% of the monthly plan | Cap daily budgets, review the allocation |

Weekly CPA review rule (expressed relative to your own target, since absolute CPA varies 3-5x between regions and industries):

| CPA vs target | Frequency | Verdict |
|---------------|-----------|---------|
| <= 1.0x target | <= 2.5 | WIN — scale +20-30% |
| 1.0-1.5x target | <= 2.5 | Watch — monitor |
| > 1.5x target, under 7 days | any | Watch — keep monitoring, no verdict yet |
| > 1.5x target for 7 days | any | Optimize — change creative or audience |
| > 3.0x target (spend >= 3x target CPA) | any | Propose a pause through rule 7 (draft block, named approver) |

Step 0: spend below 3x target CPA = not enough data, no verdict. Source: house heuristic ladder in `03-performance-eval-global` (heuristic).

Set the target itself using `references/benchmarks-global.md` for your region and industry, not a global average. **Never increase budget while CPA is deteriorating** — fix creative or tracking first.

### 7. Tracker and review cadence

| Month | Planned budget | Actual spend | Variance | ROAS / ROI | Note |
|-------|----------------|--------------|----------|------------|------|
| Month 1 | | | | | |
| Month 2 | | | | | |
| Month 3 | | | | | |
| Quarter total | | | | | |

| Cadence | Work | Output |
|---------|------|--------|
| Weekly | Review CPA and ROAS per channel, apply the step-6 decision rules | Immediate adjustment when a threshold trips |
| Monthly | Review the full allocation, reconcile plan vs actual | Rebalance for next month |
| Quarterly | Review overall allocation strategy and LTV:CAC | Strategy adjustment |

Unit economics health — LTV:CAC:

| Ratio | Meaning | Action |
|-------|---------|--------|
| < 1:1 | Losing money on every customer | Stop scaling, fix product or price |
| 1:1 to 3:1 | Break-even to viable | Optimize conversion, raise AOV |
| > 3:1 | Healthy | Begin scaling |
| > 5:1 | Strong | Increase ad budget aggressively |

## Output structure

File name: `budget-plan-[brand-name]-[YYYYMMDD].md` — contains: I. Budget overview (total, revenue target, marketing as % of revenue, period) · II. Bucket allocation · III. Channel allocation · IV. Test vs scale split · V. Campaign and monthly allocation with seasonality overlay · VI. Decision rules and stop-loss thresholds · VII. Plan-vs-actual tracker · VIII. Review cadence.

## Related skills

- `10-reverse-kpi-global`: derive the minimum budget from the revenue target — run before allocating.
- `00-marketing-plan-global`: the budget plan is the detailed version of the plan's budget section.
- `54-media-plan-global`: turn the ad budget into a concrete campaign and ad-set level media plan.
- `21-ads-audit-global`: audit a channel that repeatedly trips the stop-loss thresholds.
- `07-marketing-report-global`: budget reconciliation inside the monthly report.

## Quality checklist

- [ ] Budget derived backwards from the revenue target, with 3 scenarios
- [ ] All 3 allocation layers present: bucket > channel > campaign/month
- [ ] Test vs scale split explicit (testing 30 / scale 50 / retargeting 15 / lookalike 5)
- [ ] 10-15% reserve held back, not pre-allocated
- [ ] Scale-up and stop-loss thresholds use concrete numbers (ROAS, CPA vs target, frequency, days)
- [ ] New channels receive only a 10-15% test allocation
- [ ] Every budget line has a measurable KPI attached
- [ ] Monthly plan accounts for Q4 media inflation rather than dividing the year evenly
- [ ] Plan-vs-actual tracker exists by month
- [ ] Weekly, monthly, and quarterly review cadence with defined outputs
- [ ] LTV:CAC checked before any scale decision

---

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.