If you see unfamiliar placeholders or need to check which tools are connected, see CONNECTORS.md.

1. Report type — determine which type of report the user needs:
   • Campaign report — performance of a specific campaign
   • Channel report — performance across a specific channel (email, social, paid, SEO, etc.)
   • Content performance — how content pieces are performing
   • Overall marketing report — cross-channel summary (weekly, monthly, quarterly)
   • Custom — user-defined scope
2. Time period — the reporting window (last week, last month, last quarter, custom date range)
3. Data source:
   • If ~~product analytics is connected: pull performance data automatically
   • If not connected: ask the user to provide metrics. Prompt with: "Please paste or share your performance data. I can work with spreadsheets, CSV data, dashboard screenshots described in text, or just the key numbers."
4. Comparison period (optional) — prior period or year-over-year for trend context
5. Stakeholder audience (optional) — who will read this report (executive summary style vs. detailed analyst view)

• 2-3 sentence overview of performance in the period
• Headline metric with trend direction (up/down/flat vs. prior period)
• One key win and one area of concern

• On track (meeting or exceeding target)
• At risk (below target but within acceptable range)
• Off track (significantly below target)

• Impressions and reach
• Click-through rate (CTR)
• Conversion rate
• Cost per acquisition (CPA)
• Return on ad spend (ROAS) or ROI
• Total conversions/signups/leads

• Emails sent, delivered, bounced
• Open rate
• Click-through rate
• Unsubscribe rate
• Conversion rate

• Impressions and reach
• Engagement rate (likes, comments, shares)
• Follower growth
• Click-through rate
• Top-performing posts

• Spend
• Impressions and clicks
• CTR
• CPC and CPM
• Conversions and CPA
• ROAS

• Organic sessions
• Keyword rankings (movement)
• Pages indexed
• Backlinks acquired
• Top-performing pages

• Pageviews and unique visitors
• Time on page
• Bounce rate
• Social shares
• Conversions attributed to content
• Top and bottom performers

• Total leads generated
• Marketing qualified leads (MQLs)
• Pipeline contribution
• Customer acquisition cost (CAC)
• Channel-by-channel summary

• Performance trend over the period (week-over-week or month-over-month)
• Notable inflection points and what caused them
• Seasonal or cyclical patterns observed
• Comparison to benchmarks or targets

• Top 3-5 wins with specific data
• Why these performed well (hypothesis)
• How to replicate or scale

• Bottom 3-5 performers with specific data
• Hypotheses for underperformance
• Recommended fixes

• Patterns in the data that are not obvious from the metrics alone
• Audience behavior insights
• Content or creative themes that resonated
• External factors that may have influenced performance (seasonality, news, competitive moves)

• What to do
• Why (linked to a specific insight from the data)
• Expected impact (high, medium, low)
• Effort to implement (high, medium, low)
• Priority (immediate, next sprint, next quarter)

• Top 3 priorities for the upcoming period
• Tests or experiments to run
• Targets for key metrics

• Top 3 metrics with week-over-week change
• What worked this week (1-2 bullet points with data)
• What needs attention (1-2 bullet points with data)
• This week's priorities (3-5 action items)

1. Executive summary (3-5 sentences)
2. Key metrics dashboard (table with MoM and target comparison)
3. Channel-by-channel performance summary
4. Campaign highlights and results
5. What worked and what did not (with hypotheses)
6. Recommendations and next month priorities
7. Budget spend vs. plan

1. Quarter performance vs. goals
2. Year-to-date trajectory
3. Channel ROI analysis
4. Campaign performance summary
5. Competitive and market observations
6. Strategic recommendations for next quarter
7. Budget request and allocation plan
8. Key experiments and learnings

• Lead with the metrics that map to business objectives (not vanity metrics)
• Show trends over time, not just point-in-time snapshots
• Include comparison context: prior period, target, benchmark
• Use consistent color coding: green (on track), yellow (at risk), red (off track)
• Group metrics by funnel stage or business question
• Keep dashboards to one page/screen — detail goes in appendix
• Update cadence should match decision cadence (real-time for paid, weekly for content)

1. Directional trends: is the metric consistently going up, down, or flat over 4+ periods?
2. Inflection points: where did performance change direction and what happened then?
3. Seasonality: are there predictable patterns by day of week, month, or quarter?
4. Anomalies: one-time spikes or drops — what caused them and are they repeatable?
5. Leading indicators: which metrics change first and predict future outcomes?

1. Chart the metric over time (at least 8-12 data points for meaningful trends)
2. Identify the overall direction (upward, downward, flat, cyclical)
3. Calculate the rate of change (is it accelerating or decelerating?)
4. Overlay key events (campaigns launched, product changes, market events)
5. Compare to benchmarks or targets
6. Identify correlations with other metrics
7. Form hypotheses about causation (and plan tests to validate)

• Linear projection: extend the current trend line forward (useful for stable metrics)
• Moving average: smooth out noise by averaging the last 3-6 periods
• Year-over-year comparison: use last year's pattern as a baseline, adjusted for growth rate
• Funnel math: forecast outputs from inputs (e.g., if we generate X leads at Y conversion rate, we will get Z customers)
• Scenario modeling: create best case, expected case, and worst case projections

• Short-term forecasts (1-3 months) are more reliable than long-term
• Forecasts based on fewer than 12 data points should be flagged as low confidence
• External factors (market shifts, competitive moves, economic changes) can invalidate trend-based forecasts
• Always present forecasts as ranges, not exact numbers

• Start with last-touch attribution if you have no model in place — it is the simplest and most actionable
• Compare first-touch and last-touch to understand which channels drive awareness vs. conversion
• Use position-based (U-shaped) as a reasonable middle ground for most B2B companies
• Data-driven attribution requires high conversion volume to be statistically meaningful
• No model is perfect — use attribution directionally, not as absolute truth
• Multi-touch attribution is better than single-touch, but any model is better than none

• Do not optimize one channel in isolation based on single-touch attribution
• Awareness channels (display, social, PR) will always look bad in last-touch models
• Conversion channels (search, retargeting) will always look bad in first-touch models
• Self-reported attribution ("how did you hear about us?") provides useful qualitative color but is unreliable as quantitative data
• Cross-device and cross-channel tracking gaps mean attribution data is always incomplete

1. Identify: which metrics are underperforming vs. target or benchmark?
2. Diagnose: where in the funnel is the problem? (impressions, clicks, conversions, retention)
3. Hypothesize: what is causing the underperformance? (audience, message, creative, offer, timing, technical)
4. Prioritize: which fixes will have the biggest impact with the least effort?
5. Test: design an experiment to validate the hypothesis
6. Measure: did the change improve the metric?
7. Scale or iterate: roll out wins broadly; iterate on inconclusive or failed tests

• Test one variable at a time for clean results
• Define the success metric before launching the test
• Calculate required sample size before starting (do not end tests early)
• Run tests for a minimum of one full business cycle (typically one week for B2B)
• Document all tests and results, regardless of outcome
• Share learnings across the team — failed tests are valuable information
• A test that confirms the status quo is not a failure — it builds confidence in your current approach

• Daily: monitor paid campaigns for budget pacing, anomalies, and disapproved ads
• Weekly: review channel performance, pause underperformers, scale winners
• Bi-weekly: refresh ad creative and test new variants
• Monthly: full performance review, identify new optimization opportunities, update forecasts
• Quarterly: strategic review of channel mix, budget allocation, and targeting strategy

• Use tables for data presentation
• Bold key numbers and trends
• Keep the executive summary concise (suitable for forwarding to leadership)
• Include a "detailed appendix" section for granular data if the user provided a lot of metrics

• Create a slide-ready summary of these results?
• Draft a stakeholder email with the key takeaways?
• Dive deeper into any specific metric or channel?
• Set up a reporting template you can reuse next period?"