budget-allocator
/budget-allocator
When the user allocates a campaign budget across channels, models unit economics, or stress-tests spend scenarios. Also use when the user mentions "budget split," "channel allocation," "CAC," "LTV," "payback period," "unit economics," "how much to spend," or "budget model." Produces channel-level allocations with expected CPA/ROAS, CAC-LTV unit economics, and graceful degradation scenarios.
- Author
- doddlehq ↗
- Repository
- doddleos ↗
- Licence
- MIT
- Last repository change
- 26.10.2026
Assessment and security
- Files in this package
- 2
- Scripts
- 0
- HTML comments
- 0
- URLs in SKILL.md
- 0
No signals detected. Repository changed on 26.10.2026, without references to retired models or APIs.
MIT, open: use and modify while retaining author attribution.
What it does and when to use it
When the user allocates a campaign budget across channels, models unit economics, or stress-tests spend scenarios. Also use when the user mentions "budget split," "channel allocation," "CAC," "LTV," "payback period," "unit economics," "how much to spend," or "budget model." Produces channel-level allocations with expected CPA/ROAS, CAC-LTV unit economics, and graceful degradation scenarios.
Triggers from the description
Original author description
When the user allocates a campaign budget across channels, models unit economics, or stress-tests spend scenarios. Also use when the user mentions "budget split," "channel allocation," "CAC," "LTV," "payback period," "unit economics," "how much to spend," or "budget model." Produces channel-level allocations with expected CPA/ROAS, CAC-LTV unit economics, and graceful degradation scenarios.
How it works
- Language & Quality Standards
- When to Use This Skill
- Initial Assessment
- Inputs Schema
- Outputs Schema
- Core Framework
- Budget Split Guidance
- Common Mistakes
- MCP Tool Integration
- Related Skills
SKILL.md preview
---
name: budget-allocator
id: doddle.marketing.budget-allocator
version: 1.0.0
blueprint: ./blueprint.yaml
description: When the user allocates a campaign budget across channels, models unit economics, or stress-tests spend scenarios. Also use when the user mentions "budget split," "channel allocation," "CAC," "LTV," "payback period," "unit economics," "how much to spend," or "budget model." Produces channel-level allocations with expected CPA/ROAS, CAC-LTV unit economics, and graceful degradation scenarios.
---
# Budget Allocator
The money model for a campaign. Takes the brief and the total budget, and returns per-channel allocations with rationale, expected CPA/ROAS, and the CAC/LTV/payback math that tells you whether the plan survives contact with reality — before a dollar is spent.
Executable via `blueprint.yaml` (DoddleOS graph). See Inputs/Outputs below.
## Language & Quality Standards
**CRITICAL**: Respond in the same language the user is using. If Vietnamese, respond in Vietnamese. If Spanish, respond in Spanish.
**Standards**: Token efficiency, sacrifice grammar for concision, list unresolved questions at end.
---
## When to Use This Skill
Apply this skill when:
- Stepping a total budget into channel allocations
- Modeling CAC, LTV, payback, or margin for a campaign
- Comparing lean / base / stretch spend scenarios
- Deciding whether a channel earns budget at all
- Pre-launch sanity-checks from a finance perspective
## Initial Assessment
Before modeling, understand:
1. **Money bounds**
- Total budget + timing (one-shot vs monthly)?
- Price or ACV of the offer? (proposes an assumption if missing, flagged)
2. **Channel shortlist**
- Any must-run or must-avoid channels (compliance, exec mandate)?
- Any minimum viable test budgets?
3. **Goals**
- From intake primary KPI: volume-target vs efficiency-target?
---
## Inputs Schema
| Input | Type | Required | Description |
|-------|------|----------|-------------|
| brief | string | yes | Campaign brief: goal, audience, channels |
| total_budget | number | yes | Total budget in account currency |
| price_or_acv | string | no | Deal size or price band; propose assumption if missing |
---
## Outputs Schema
| Output | Type | Description |
|--------|------|-------------|
| budget_split | json | Channel allocations with rationale and expected CPA/ROAS |
| unit_economics | markdown | CAC, LTV, payback, margin math + lean/base/stretch scenarios |
---
## Core Framework
### Step 1: Fix the Unit Economics
Compute with user numbers or explicitly-flagged assumptions:
- `LTV = ARPU × gross margin × avg lifetime` (state every input)
- `CAC ceiling = LTV ÷ target payback period (months)` (e.g. payback < 6 mo → max CAC ≈ 12 months ARPU × margin ÷ 6)
- Blended CAC ceiling for the campaign = the money constraint the plan must respect
### Step 2: Floor Every Channel
For each candidate channel: minimum viable budget (enough impressions/clicks to exit learning), expected CPA band, and ramp time. A channel below its floor gets 0, not a token amount that never exits learning.
### Step 3: Allocate with Rationale
Split by expected contribution to the primary KPI, not by habit:
- Proven channels (existing baseline data) get scale weight
- Test channels get capped test budgets (10–20% of total)
- Reserve 5–10% contingency; it is not extra channel budget
- Every line carries expected CPA/ROAS and its assumption
### Step 4: Stress-Test Scenarios
Emit three scenarios: lean (-30%), base, stretch (+30%). For each: expected signups volume, blended CAC, payback period, and the decision rule that moves money if actuals diverge.
### Step 5: Approval-Ready Summary
One-page allocation summary with Math, Assumptions, Limitations. Assumptions are the honest list — never fabricate benchmarks; label industry comparators as assumptions to validate.
---
## Budget Split Guidance
```json
{
"total_budget": 20000,
"currency": "USD",
"cac_ceiling": 150,
"contingency_pct": 8,
"channels": [
{
"channel": "meta-ads",
"amount": 8000,
"expected_cpa": [60, 110],
"rationale": "retargeting + lookalikes off existing pixel",
"test_share": 0.15
},
{
"channel": "seo/content",
"amount": 6000,
"expected_cpa": [35, 55],
"ramp_weeks": 6,
"note": "lower CPA, slow ramp; not cut in week 2"
}
],
"assumptions": [
"assumed ACV $1,200 (unconfirmed — needs finance check)",
"no top-of-funnel PR credited to conversion"
]
}
```
---
## Common Mistakes
### Strategy Mistakes
| Mistake | Why It's Wrong | Do This Instead |
|---------|----------------|-----------------|
| Equal splits | Spends according to taste, not math | CPA-based allocation with floor checks |
| No contingency | No budget to reallocate when a channel wins | 5–10% reserve, rule-governed |
| Channel budget without floor | $300 exits?? never exits learning | Minimum viable test budgets |
| Hidden CAC assumption | Model looks profitable via invented LTV | Every input labeled as fact or assumption |
### Execution Mistakes
| Mistake | Why It's Wrong | Do This Instead |
|---------|----------------|-----------------|
| Lifting benchmark numbers from memory | Fabricated data, wrong currency | ⚠️ NOT AVAILABLE + range assumptions, flagged |
| One-position allocation | Week-2 data has nowhere to go | Reallocation triggers pre-agreed (see measurement-plan) |
| Budget > channel saturation | Diminishing returns ignored | Cap per channel when CPA bends upward |
---
## MCP Tool Integration
| Tool ID | When to Use | Data | Fallback |
|---------|-------------|------|----------|
| doddle.tool.v1.ga4.getReport | Channel baselines, past CPA | Sessions, conversions | "⚠️ NOT AVAILABLE" |
| doddle.tool.v1.hubspot.deals | ACV, deal-size distribution | Deals | "⚠️ NOT AVAILABLE" |
Never fabricate benchmarks. If actuals are unavailable, use user-stated assumptions, label them, and list them in the summary.
---
## Related Skills
- `doddle.marketing.measurement-plan` — the KPI targets CAC/payback get checked against
- `doddle.marketing.paid-advertising` — lights implementation of the paid split
- `doddle.marketing.media-reallocator` (planned) — in-flight reallocation against these scenarios
Files in this package (2)
- SKILL.md 6 KB
- blueprint.yaml 1 KB
Versions
| Repository | Score | Licence |
|---|---|---|
| This page · doddlehq/doddleos | 100 | MIT |
| varunk130/ai-gtm-skill-library | 100 | MIT |
Database snapshot: 13.11.2026 · package version edc5b0fcecb8 (SHA-256 digest of the entire package)
Downloads in the last 30 days: 2 · ranking
Report a problem with this skill: harmful content, incorrect descriptions or an author’s removal request.
Similar skills
- 95ltv-cac-ratio-optimizer
Calculates LTV, CAC, the LTV to CAC ratio and payback period from your customer and spend data, by channel, cohort and first product, then sets max CPA and ROAS targets and budget shifts. Use when asking about customer lifetime value, unit economics, payback period or profit beyond ROAS. - 100gtm-channel
When the user wants to compare acquisition channels on unit economics — CAC, LTV:CAC and payback. Use for 'which channel is best', 'compare channels', 'channel economics', 'paid vs organic vs partner', 'where should we spend'. Inputs are assumptions the user supplies. - 100unit-economics
Use when the user wants to calculate or sanity-check unit economics — LTV, CAC, the LTV:CAC ratio, and CAC payback — for a subscription, ecommerce, or services business. Also use when the user mentions lifetime value, customer acquisition cost, payback period, churn,… - 90unit-economics
Compute the SaaS unit-economics quartet — CAC, LTV, LTV:CAC ratio, and CAC payback period — and assess whether the business acquires customers profitably. Use when evaluating growth health or comparing acquisition channels. - 90unit-economics-cac-ltv-payback
Activate when: deciding whether to spend more on growth; 'is this business actually profitable per customer', 'can we afford ads', raising or budgeting; CAC, LTV, payback, contribution margin. Do NOT activate when: pre-revenue with no cost data (estimate ranges instead) or the… - 100unit-economics-check
Checks CAC, LTV and payback for a subscription or app business, finds which input makes the numbers unreliable, and sets a safe CAC ceiling. Use when the user asks "is my CAC okay", "calculate LTV", "payback period", "can we afford to scale", "unit economics check", "ROAS…