Explain Changes (Change Insights)

Change Insights answers the critical question: "Why did my metrics go up or down?" with AI-powered causal analysis.

What It Is

When your traffic, rankings, or visibility change significantly, understanding the cause is essential. Change Insights:

• Compares two time periods you select
• Detects meaningful changes across multiple data sources
• Generates explanations for what caused the changes
• Provides actionable recommendations to capitalize on wins or recover from losses

How It Works

Data Sources Analyzed

Change Insights examines changes across:

Analysis Process

1. Select two periods (e.g., this month vs. last month)
2. System calculates deltas for all metrics
3. AI identifies correlations between changes
4. Narrative is generated explaining the findings
5. Recommendations provided based on analysis

Using Change Insights

Accessing the Feature

1. Go to Insights in your project navigation
2. Or click Explain Changes from the Reports page

Selecting Time Periods

Choose from:

• Preset ranges: Last 7 days, Last 30 days, Last quarter
• Custom dates: Pick specific start and end dates
• Compare options: Previous period, Same period last year

Reading the Analysis

#### Summary Section

High-level overview of what changed and why, written in plain language.

#### Metric Breakdowns

Detailed view of each metric showing:

• Absolute change (e.g., +500 clicks)
• Percentage change (e.g., +15%)
• Contributing factors

#### Drill-Down Details

Click any metric to see:

• Page-level breakdown
• Keyword-level analysis
• Correlation with other changes

Example Insights

"Organic clicks increased 23% primarily due to improved rankings for 'product comparison' queries. The rankings improvement correlates with the FAQ schema added to 5 product pages on March 15th."

"Answer Ownership dropped 12% because competitor XYZ started appearing in ChatGPT responses for your key brand queries. Consider creating more definitive content targeting these prompts."

Fact-Checking

Change Insights uses strict fact-checking:

1. Only reports verified data from connected sources
2. Distinguishes correlation from causation clearly
3. Provides confidence levels for each insight
4. Links to source data for verification

Best Practices

1. Check after major changes: Algorithm updates, site launches, content updates
2. Compare meaningful periods: Avoid comparing holidays to normal weeks
3. Act on insights: Use recommendations to drive your strategy
4. Track follow-up: Re-check after implementing changes

FAQ

Q: How far back can I compare?

A: As far as your connected data sources retain data (typically 16 months for GSC).

Q: Why are some insights marked "low confidence"?

A: When there are multiple possible causes or insufficient data, confidence is lowered.

Q: Can I export the analysis?

A: Yes. Click "Export" to download as PDF or include in a report.