| name | report-automation |
| description | Designs automated reporting workflows with data sources, schedules, distribution lists, and template standardization. Use when eliminating manual reporting work. |
| allowed-tools | Read Write Glob |
| metadata | {"author":"matthewhitcham","version":"1.0"} |
Report Automation
When to Use This Skill
Use this skill when you need to:
- Automate recurring business reports that are currently manual
- Design a reporting workflow with scheduled delivery
- Standardize report templates across the business
- Reduce hours spent compiling data into presentations or spreadsheets
DO NOT use this skill for one-time analyses, dashboard design, or building automation scripts. This is for planning the automation workflow and template design.
Core Principle
AUTOMATE THE REPORT, NOT THE THINKING — AUTOMATED DATA DELIVERY FREES TIME FOR ANALYSIS AND DECISIONS INSTEAD OF COPY-PASTING NUMBERS.
Phase 1: Brief
Required Inputs
| Input | What to Ask | Default |
|---|
| Reports to automate | "Which recurring reports take the most time?" | Must be provided |
| Current process | "How are these reports created now? (manual spreadsheet, copy-paste, tool export)" | Manual spreadsheet |
| Frequency | "How often is each report needed? (daily, weekly, monthly)" | Weekly |
| Audience | "Who receives each report? (founder, team, clients, investors)" | Founder and team leads |
| Data sources | "Where does the report data come from? (GA4, Stripe, CRM, spreadsheets)" | Multiple sources |
| Tools available | "What tools do you use or would consider? (Google Sheets, Zapier, Looker, Databox)" | Google Sheets + Zapier |
GATE: Confirm brief before proceeding.
Phase 2: Design
Automation Assessment
For each report, evaluate:
- Data collection: Can data be pulled automatically? (API, integration, export)
- Transformation: Does data need calculations or reformatting?
- Presentation: Can the template be standardized?
- Distribution: Can delivery be automated? (email, Slack, dashboard link)
Automation Levels
| Level | Description | Tools |
|---|
| Manual with template | Standardized template, manual data entry | Google Sheets |
| Semi-automated | Data pulls automated, human reviews before sending | Zapier + Sheets |
| Fully automated | Data to delivery with no human touch | Looker, Databox, custom |
GATE: Present the automation plan with recommended level per report.
Phase 3: Build
Deliverables
1. Report Automation Specification
- Each report mapped: data source → transformation → template → delivery
- Automation level and tools per report
- Schedule and trigger definitions
- Distribution list per report
2. Standardized Report Templates
- Template for each report type with placeholder sections
- Consistent formatting: header, date range, key metrics, charts, commentary section
- Brand-consistent styling guidelines
3. Implementation Guide
- Step-by-step setup for each automation
- Integration configuration instructions
- Testing procedures (verify data accuracy before automating delivery)
- Error handling: what happens when a data source fails?
4. Maintenance Checklist
- Monthly: verify all automations are running and data is accurate
- Quarterly: review whether reports are still needed and useful
- Annually: audit the full reporting stack for redundancies
Phase 4: Polish
Time Savings Calculation
Document for each report:
- Time spent before automation (hours/week)
- Time spent after automation (hours/week)
- Annual time saved
- Dollar value of recovered time
Iteration Recommendations
After 1 month of automated reports, gather feedback: Is anything missing? Is anything ignored? Adjust templates and frequency based on actual usage.
Example 1: Weekly Business Metrics Report
Before: 2 hours every Monday pulling numbers from GA4, Stripe, and Mailchimp into a Google Doc.
After: Google Sheets with Supermetrics auto-pulling data, summary auto-generated, Slack notification sent Monday 8am. Human reviews and adds commentary (20 minutes).
Example 2: Monthly Client Report (Agency)
Before: 4 hours per client assembling analytics, ad performance, and deliverable summaries.
After: Looker Studio dashboard with auto-refreshing data, PDF auto-generated and emailed to clients on the 1st. Human adds strategic commentary (30 minutes per client).
Anti-Patterns
- Automating bad reports — if nobody reads the report, automating it just sends garbage faster. Validate usefulness first.
- Over-engineering the first version — start semi-automated. Prove the template works, then add full automation.
- No human review layer — fully automated reports without periodic accuracy checks will eventually send wrong numbers. Build in spot-checks.
- Too many reports — automation makes it easy to create reports. Fight the urge. Fewer, better reports beat more reports.
- Ignoring data freshness — a "daily" report pulling from a source that updates weekly is misleading. Match report frequency to data availability.
Recovery
- Data sources do not have APIs: Use scheduled CSV exports + Google Sheets IMPORTDATA or Zapier file triggers as a bridge.
- Reports serve different audiences: Create one data source with multiple views — executive summary for leadership, detailed breakdown for operators.
- Automation breaks frequently: Simplify. Complex chains break more often. Reduce integration points or use more robust tools.
- User does not trust automated numbers: Run automated and manual reports in parallel for 2 weeks. Compare and resolve discrepancies before retiring the manual process.
