Explain a Marketing Performance Change
Design an analytics investigation flow that helps a marketer move from an unusual metric change to a defensible explanation.
The brief
Understand the problem
Background
Marketing dashboards can show that conversions changed without explaining whether the cause was campaign performance, tracking quality, attribution settings, or normal variation.
User context
Alyssa sees reported conversions drop by 24 percent after a website release. Spend and traffic look stable, so she needs to determine whether performance or measurement changed.
Product problem
A polished chart can imply certainty even when source data is incomplete. Investigators need comparison context, annotations, data quality, and a record of hypotheses.
Objective
Create a guided analytics investigation from anomaly detection through evidence-backed explanation.
What to design
Define the experience
Required experience
- Marketing overview with anomaly and data-quality signals
- Metric breakdown by channel, segment, and time
- Timeline annotations for campaign and product changes
- Investigation notes and shareable conclusion
Screens and states
- Performance overview
- Metric explorer
- Annotated timeline
- Data quality detail
- Investigation summary
Core user flow
Follow the critical path
- 01
Alyssa opens the conversion anomaly from the overview
- 02
She compares channels and finds the decline concentrated on mobile web
- 03
She overlays the website release and tracking health
- 04
She records a measurement hypothesis and supporting evidence
- 05
She shares a conclusion that distinguishes known facts from open questions
Product rules
Requirements and constraints
Requirements
- Show metric definition, attribution model, and data freshness
- Support comparison with an appropriate prior period
- Allow drill-down without losing the original anomaly context
- Place operational events and data-quality issues on the same timeline
- Separate observation, hypothesis, evidence, and conclusion
- Make exports preserve filters and caveats
Constraints
- Source systems update at different times
- Attribution models can produce different answers
- Some segments are too small for reliable conclusions
- Historical tracking definitions may have changed
Reality check
States worth considering
Finish line
What to deliver
- Design five desktop screens that turn a metric anomaly into a transparent investigation narrative.
Optional direction
Visual resources
Use these as a starting constraint if you want one. They are not part of the required solution.
Montserrat & Crimson Text
Clear interface writing gives people the confidence to understand what changed and decide what to do next.
Montserrat & Crimson Text
Clear interface writing gives people the confidence to understand what changed and decide what to do next.