Impact Analysis
Expand one deploy report into traffic, conversion, speed, revenue, and incident context.
Impact Analysis answers the larger question after a deploy, pricing change, campaign, or incident:
"What changed around this report, and what else should the team read before acting?"
What the page combines
Impact Analysis expands the Deploy Impact Report with:
- the latest production release summary
- traffic and engagement movement
- configured conversion signals
- Speed Insights context
- revenue attribution
- manual change markers and anomaly incidents
How to read it
Start with the directional explanation
The explanation is intentionally cautious.
It is there to help you investigate, not to claim perfect causality.
Read these parts first:
- verdict
- confidence
- supporting evidence
- recommended next actions
- cautions and current warnings
Then inspect the current release context
If a recent release exists, Callra reuses the release summary directly instead of inventing a separate release-health model.
That keeps the report and impact stories aligned.
Then inspect the top breakdowns and merged timeline
Use the breakdowns to see where movement is concentrated.
Use the merged timeline to line up:
- releases and deployments
- manual notes
- revenue events
- anomaly incidents
Important limits to keep in mind
Today the revenue side of Impact Analysis is intentionally conservative:
- revenue attribution is last-touch only
- mixed currencies should be treated directionally
- low-volume windows may not support strong conclusions
- timing overlap can suggest impact without proving causality
Recommended workflow
- Open Releases first when the latest deploy looks suspicious.
- Open Impact Analysis when you need the larger traffic, signup, checkout, speed, or revenue picture.
- Open Reports and incidents when the team needs follow-up, escalation, or delivery history.
- Use CLI or MCP when the same investigation needs to be repeated by a person or an agent.
Good uses for manual change markers
Manual markers are useful for business changes that are real but not deploys, such as:
- pricing changes
- newsletters
- campaigns
- incident notes
- launch-day annotations
Next step
- Need scheduled follow-up: Reports and incidents
- Need repeatable machine investigation: Agent, CLI, and MCP workflows