BUILDER · charts & dashboards built live · synthetic data
Ask a question about the campaign data.
The assistant picks a chart and draws it here. Export it or save it.
Compare opens vs clicks by year
Which audience has the most unsubscribes?
Show click rate trend over time
Top 5 campaigns by opens
One question → one chart. Several lines → a dashboard (one panel each). ⌘/Ctrl+Enter or Build.
Dashboard Links
Define navigation between dashboards. A link adds an "Open →" action to the source dashboard's charts — click a data point to jump. Managed here, no DevOps needed.
→
Defined links
Real campaign data. Engagement rates from 20,000 campaign sends (2021–2024). We lead with rates — open rate, click-through, unsubscribe rate — because they're the reliable signal; absolute volumes in the source are synthetic-scale and left out on purpose. Filter by campaign or audience to slice the trend.
Open rate by year
Email content characteristics. Extracted from 1,244 QC emails (subject + body feature pipeline): sentiment, tone, readability, structure. Descriptive for now — tying these to engagement is pending a data join, but the content patterns themselves are real.
⚠ Demonstration — synthetic signal. The QC campaign data has no real relationship between content and engagement, so to show how driver analysis works, a realistic signal was injected into a copy of the data. The correlations below are honestly computed — but on constructed data, not live campaign performance. This illustrates the analysis for review; it is not a finding about real campaigns.
Compose an email using these drivers
Generate an email that deliberately applies the drivers above. Start from a real sample campaign, or describe your own goal.
Ask about the campaign data. Question the real aggregated rates in plain language. Answers are grounded strictly in the data — Claude only cites numbers we computed and will say when something can't be answered or when engagement is steady across the board.
Delivery & bounce analysis. Real campaign data — delivery and failure rates across 1,244 sends. This is the "reduce rejection" view: overall delivery health, where bounces concentrate, and outlier sends worth investigating.
⚠ Demonstration. Estimates a likely open-rate range for a draft by applying the demonstration driver model (synthetic signal). Illustrates how predictive scoring would work; not a validated forecast on live performance.