# Marketing ROI
Use this skill to prove whether marketing work created measurable business value.
This skill is deliberately more rigorous than a normal campaign report. Its job is to separate evidence from assumptions, connect marketing activity to revenue or qualified pipeline, quantify uncertainty, and produce a client-believable proof package.
Before intake, read `./references/dependencies/conversint-current/references/skill-operating-standards.md` and `./references/dependencies/conversint-current/references/output-paths.md`. This skill owns ROI proof and readiness-gap artifacts; tracking implementation, campaign operations, ad creative, CRM writes through `/crm`, Drive publishing through `/gdrive`, and project-task updates remain with their owning skills.
What This Skill Owns
- Marketing ROI model design
- Baseline and counterfactual planning
- Attribution evidence review
- Incrementality test design
- Campaign, website, SEO, content, and ad impact measurement
- Funnel economics: spend, leads, qualified leads, opportunities, revenue, margin, CAC, ROAS, payback
- Client-facing ROI proof packs
- Brochure proof pages and case-study signal standards when an offering brochure needs credible prior-work evidence
- Internal measurement gap audits
- Spreadsheet model specifications for ROI analysis
- Escalation point for economically meaningful learnings from `/marketing-learning-loop`
What This Skill Does Not Own
- Pre-sale proposal pricing: use `/proposal-value-analysis`.
- Tag/pixel implementation details: use `/analytics-tracking`.
- Paid campaign launch structure: use `/paid-ads`.
- Ad copy production: use `/ad-creative`.
- Organic post creation: use `/social-content`.
- Cross-channel learning memory and test decisions: use `/marketing-learning-loop`.
- Full brochure creation, design QA, website resource updates, and Drive publishing: use `/offering-brochure`.
Required Context Load
Before analysis:
- If this is for an active project, read `./references/dependencies/conversint-current/references/active-projects.md`, then load the relevant project ledger with:
```bash scripts/python lib/project_manager.py report --project <project-id> ```
2. If the customer has local files, read relevant artifacts under `customers/<Customer Name>/`, especially: - meeting notes - project ledger - proposal value analysis - scope or engagement draft - website or campaign notes - prior reports
3. For Conversint positioning and banned language, read: - `./references/dependencies/conversint-current/references/org/identity.md` - `./references/dependencies/conversint-current/references/org/anti-patterns.md` - `./references/dependencies/conversint-current/references/org/pricing.md` only if discussing pricing, value-based fees, or the public ROI frame
When measuring or describing Conversint's own marketing, also read: - `./references/dependencies/conversint-current/references/marketing-message-governance.md` - `research-thought-leadership/conversint-messaging-optimization/active-message-catalog.md`
Do not treat a message as active, publishable, or proven from a prior report alone. Use the catalog's channel status and the current dated learning ledger; route a status change to `/marketing-learning-loop`.
4. If producing a client-facing Markdown or DOCX deliverable, read: - `./references/dependencies/conversint-current/references/client-deliverable-standards.md` - `./references/dependencies/conversint-current/references/conversint-brand.md` - `./references/dependencies/conversint-current/references/output-paths.md`
5. Read the relevant supporting references: - Always read `./references/dependencies/conversint-current/references/skill-operating-standards.md` - Always read `./references/data-access-checklist.md` - Always read `./references/roi-models.md` - Always read `./references/attribution-and-incrementality.md` - Read `./references/dependencies/conversint-current/references/paid-ads-audit-engine.md` when the value question involves ad spend, ad platform metrics, paid landing pages, paid lead generation, or campaign optimization - Read `./references/dependencies/conversint-current/references/digital-marketing-funnel-best-practices.md` when assessing a paid landing-page, CTA, form, offer, or lead-quality path - Read `./references/client-proof-pack.md` when producing client-facing output - Read `./references/spreadsheet-model-spec.md` when producing a workbook
Intake
Resolve these before doing math:
- Customer or business name. - Campaign, channel, project, or website change being measured. - Measurement purpose: - forecast before launch - tracking plan before launch - mid-campaign readout - post-campaign ROI proof - dashboard design - attribution audit - board/client-facing report - Date range and comparison period. - Primary business outcome: booked calls, qualified leads, opportunities, closed revenue, gross profit, retained revenue, saved labor, or avoided spend. - Sales cycle length. - Average order value, contract value, gross margin, close rate, and lead-to-opportunity rate if available. - Ad spend, agency/vendor spend, internal labor cost, and implementation cost. - Data sources available and access status.
If key values are missing, do not block. Build low/base/high ranges and clearly label assumptions.
Required Access Checklist
Use `./references/data-access-checklist.md` to identify what access is needed. Minimum viable access depends on the claim:
- To prove **traffic or engagement**: analytics access and campaign UTMs.
- To prove **lead generation**: form submissions, call tracking, CRM/source fields, and lead quality labels.
- To prove **pipeline**: CRM opportunities tied to source/medium/campaign or manual source review.
- To prove **revenue ROI**: closed-won revenue, gross margin, spend, and a defendable attribution or incrementality method.
- To prove **profit ROI**: revenue, gross margin, discounts, refunds, fulfillment cost, campaign cost, and labor/vendor cost.
If access is unavailable, produce an **ROI readiness gap report** rather than pretending ROI is proven.
ROI Proof Levels
Use this ladder to classify the strength of evidence:
| Level | Name | Evidence standard | Claim allowed | |---|---|---|---| | 0 | Not instrumented | No reliable tracking | "We cannot prove impact yet." | | 1 | Activity tracked | Spend, impressions, clicks, traffic | "The campaign generated activity." | | 2 | Conversion tracked | Leads/forms/calls tied to UTMs or source | "The campaign generated tracked leads." | | 3 | Quality tracked | Leads scored or qualified in CRM | "The campaign generated qualified demand." | | 4 | Pipeline tracked | Opportunities tied to campaign/source | "The campaign influenced pipeline." | | 5 | Revenue attributed | Closed revenue tied to campaign/source with caveats | "The campaign produced attributed revenue." | | 6 | Incrementality tested | Holdout, geo split, time-series, or controlled experiment | "The campaign likely caused incremental value." |
Most small businesses start at Level 1-3. The skill should move them up the ladder over time.
Analysis Workflow
1. **Define the value question.** - Example: "Did the brochure + landing page + paid traffic create incremental qualified pipeline for FACCorp?"
2. **Map the funnel.** - Spend or effort - Impressions/reach - Clicks/traffic - Engaged sessions - Conversions - Qualified leads - Opportunities - Closed revenue - Gross profit - Retention or repeat value
3. **Build the measurement architecture.** - UTM schema - conversion events - source capture - CRM fields - call tracking - lead quality rules - offline conversion import, if applicable
4. **Establish baseline and comparison.** - Prior period - same period last year, if seasonality matters - untreated geography/audience/page, if possible - expected run-rate without campaign
5. **Calculate economics.** - Use `./references/roi-models.md`. - Separate revenue, gross profit, and cash payback. - Include all spend and material labor/vendor costs. - Use low/base/high assumptions.
6. **Assess attribution strength.** - Use `./references/attribution-and-incrementality.md`. - Use `./references/dependencies/conversint-current/references/paid-ads-audit-engine.md` to check whether tracking, source capture, and platform-reported conversions are credible enough to support the ROI claim. - Separate direct attribution from assisted influence and incrementality. - Do not present platform-reported conversions as finance-grade proof by default.
7. **Produce decision guidance.** - Continue, stop, scale, narrow, fix tracking, fix landing page, or retest. - State what would change the conclusion.
8. **Feed the learning loop.** - Any ROI conclusion that changes audience, message, channel, offer, website, or budget strategy should be written back to `/marketing-learning-loop`.
Core Formulas
Use the formulas from `./references/roi-models.md`. Default definitions:
```text Attributed revenue = closed revenue tied to campaign/source Gross profit = attributed revenue * gross margin Total campaign cost = media spend + production cost + platform/tool cost + external labor + material internal labor Net profit impact = gross profit - total campaign cost Marketing ROI = net profit impact / total campaign cost ROAS = attributed revenue / media spend CAC = total campaign cost / new customers Payback period = CAC / monthly gross profit per customer Pipeline ROI proxy = qualified pipeline value * expected close rate * gross margin / total campaign cost ```
When revenue is not available, do not call lead metrics ROI. Call them leading indicators or pipeline proxies.
Client-Facing Standards
When the output goes to FACCorp or another client:
- Start with the conclusion and confidence level.
- Show the evidence ladder level.
- Explain assumptions in plain English.
- Distinguish observed facts, modeled estimates, and recommendations.
- Include a tracking gaps section.
- Avoid false precision.
- Avoid claiming causality unless incrementality evidence supports it.
- Use "attributed" or "estimated" where appropriate.
- Include a "what we need next to improve proof" section.
- For brochure proof pages, keep proof concise, defensible, and buyer-facing. Avoid unnecessary meta language such as "anonymized examples" unless legal/compliance requires it. Use "selected examples from prior work" or similar language when client names are intentionally omitted.
Branded DOCX and Markdown deliverables use the joint-founder attribution default from `./references/dependencies/conversint-current/references/org/identity.md` / `./references/dependencies/conversint-current/references/client-deliverable-standards.md`. Ask about attribution only when the user requests a solo byline.
Output Modes
ROI Readiness Audit
Use when the business cannot yet prove ROI.
```markdown # <Customer> - Marketing ROI Readiness Audit
## Bottom Line ## Current Proof Level ## What Can Be Proven Today ## What Cannot Be Proven Yet ## Required Access ## Tracking Gaps ## Measurement Architecture To Implement ## 30-Day Proof Plan ## Risks And Caveats ```
ROI Model
Use when data and assumptions are enough for calculation.
```markdown # <Customer> - <Campaign> ROI Model
## Bottom Line ## Date Range ## Proof Level ## Data Sources ## Funnel Metrics ## Economics ## Low / Base / High Scenario ## Attribution Assessment ## Incrementality Assessment ## Decision Recommendation ## What Would Change The Conclusion ## Access And Tracking Gaps ## Sources / Data Pulls ```
Client Proof Pack
Use when presenting results externally.
```markdown # <Customer> - Marketing ROI Proof Pack
## Executive Summary ## What We Changed ## What Happened ## What We Can Prove ## ROI / Pipeline / Leading Indicator Model ## Confidence And Caveats ## Next 30 Days ## Appendix: Data Sources And Assumptions ```
Spreadsheet Handling
Create a spreadsheet only when explicitly requested.
When creating one:
- Read `./references/dependencies/conversint-current/references/spreadsheet-artifact-standards.md`.
- Read `./references/spreadsheet-model-spec.md`.
- Keep the Markdown explanation as the canonical narrative.
- Use editable assumptions and formula-driven outputs.
- Include low/base/high scenarios.
- Verify formulas before reporting completion.
Saving Artifacts
For client-specific Markdown, save under:
```text customers/<Customer Name>/<Campaign or Project> - Marketing ROI <Audit|Model|Proof Pack>.md ```
Commit Markdown artifacts to `main` with:
```bash scripts/commit-md-to-main.sh \ "customers/<Customer Name>/<File>.md" \ "Update Marketing ROI artifact: <Customer Name>" ```
Do not auto-upload to Drive unless the user explicitly asks.
Return Format
Return
- Proof level.
- Bottom-line ROI or readiness conclusion.
- Data sources used.
- Key assumptions.
- What can be proven now.
- What cannot be proven yet.
- Access still needed.
- Recommended next measurement step.
- Artifact path or GitHub URL if saved.