Trends & release impact

What it shows

Two stacked time series across the window of your choice:

• Conversation volume — daily count, with a 7-day moving average overlay to smooth weekday cycles.
• Average quality score — daily mean of scored replies, same smoothing.

Optional overlays: negative sentiment rate, FRT p90, containment rate. Toggle them on per chart.

Release markers

Every merge to main, master, prod or production in your connected GitHub repos appears as a vertical line on the chart. Hover to see the PR title, author and commit hash. Click to open the GitHub commit.

Before / after analysis

Click any release marker to open the impact panel. ResponZ computes the deltas over the 7 days after the release vs the 7 days before, on the same weekday alignment:

• Volume delta (absolute and %).
• Score delta (absolute, on a 0–100 scale).
• Negative sentiment share delta.
• Top 3 clusters that moved up or down.

This is what makes the "did it help?" question answerable. A release that drops support volume on its target cluster by 15% within 7 days is a hit. A release that raises negative sentiment by 4 points without dropping volume is a regression — even if no one filed a bug.

Manual annotations

Some milestones aren't releases: a marketing campaign launch, an incident, a price change, a competitor outage. Add them manually from the chart context menu. They render as dashed vertical lines so you can distinguish them from product releases.

Workflows

Post-launch checklist (D+7)

Open the release marker for the launch you care about. Read the delta panel. If volume on the targeted cluster dropped, ship the screenshot to the launching team. If it rose, escalate before the next sprint plan.

Quarterly business review

Pick the full quarter, screenshot the volume + score chart with all release markers visible. It's the single best one-slide narrative of what shipped, what it cost in support, and where quality landed.