Compare Periods in the MCP: Prove Your Impact to Clients
If you’re an agency, you know the challenge: AI search optimization is new enough that clients want proof it’s working. “We improved your rankings” doesn’t hit the same way as “Your visibility went from 37% to 48% — that’s a 30% increase.”
We built period-over-period comparisons into every analytics tool in the Knowatoa MCP. Now you can pull month-over-month or week-over-week comparisons directly into Claude, ChatGPT, or whatever AI assistant you’re using to build client reports.
How it works
Every MCP tool that returns metrics now accepts date ranges and automatically compares them to the previous period. Ask Claude:
“Show me my visibility for September 2026 compared to August”
And you get back exact numbers:
json
{
"mention_rate": {
"current": 48.0,
"previous": 37.0,
"change_points": 11.0,
"change_percent": 29.7
}
}
ChatGPT mention rate went up 11 percentage points — a 30% improvement. That’s the kind of specific, data-backed statement that wins client renewals.
What you can compare
Period-over-period works across every analytics tool:
Visibility — See if more AI services are mentioning your client’s brand
Sentiment — Track if mentions are becoming more positive over time
Citations — Watch which domains AI services cite more or less
Competitor analysis — Show how your client is gaining ground on competitors
Per-question performance — Drill into which questions are improving
Per-AI-service breakdowns — See which platforms (ChatGPT, Perplexity, Gemini, etc.) are moving
Every metric returns four values: - Current period — The timeframe you asked about - Previous period — An equal-length comparison window (or specify your own) - Absolute change — The raw difference - Percentage change — The relative improvement (or decline)
Example: Building a client report
Let’s say you’re preparing a monthly review. You can ask Claude:
“Compare my client’s visibility between September and August 2026”
Claude calls the MCP with:
json
{
"name": "get_visibility",
"arguments": {
"site_id": "tpc_abc123",
"start_date": "2026-09-01",
"end_date": "2026-09-30"
}
}
The response comes back with site-wide metrics and a per-AI-service breakdown:
json
{
"period": {
"start": "2026-09-01",
"end": "2026-09-30",
"refresh_count": 12
},
"compare_to": {
"start": "2026-08-02",
"end": "2026-08-31",
"refresh_count": 11
},
"metrics": {
"mention_rate": {
"current": 48.0,
"previous": 37.0,
"change_points": 11.0,
"change_percent": 29.7
},
"average_position": {
"current": 2.3,
"previous": 3.1,
"change_points": -0.8,
"change_percent": -25.8
}
},
"series": [
{
"name": "ChatGPT",
"metrics": {
"mention_rate": {
"current": 60.0,
"previous": 45.0,
"change_points": 15.0,
"change_percent": 33.3
}
}
},
{
"name": "Perplexity",
"metrics": {
"mention_rate": {
"current": 40.0,
"previous": 30.0,
"change_points": 10.0,
"change_percent": 33.3
}
}
}
]
}
Claude then formats this into a narrative: “Your visibility improved 30% in September, with ChatGPT mentions up 33% and your average position improving from 3.1 to 2.3.”
You can copy that directly into your client deck, or ask Claude to format it as a table, chart, or bullet points.
Beyond month-over-month
The comparison period defaults to the same length as your query period, but you can specify any window:
“Compare Q3 2026 to Q1 2026”
Or:
“Show me last week vs. the week before”
The maximum period is 366 days, so you can do year-over-year comparisons once you have enough history.
Competitor comparisons
One of the most powerful uses: showing clients how they stack up against competitors over time.
Ask Claude:
“Compare my client’s visibility to their top three competitors for September vs. August”
You get back each brand’s metrics side-by-side:
json
{
"brands": [
{
"name": "Your Client",
"is_target": true,
"metrics": {
"mention_rate": {
"current": 48.0,
"previous": 37.0,
"change_points": 11.0,
"change_percent": 29.7
}
}
},
{
"name": "Competitor A",
"is_target": false,
"metrics": {
"mention_rate": {
"current": 52.0,
"previous": 48.0,
"change_points": 4.0,
"change_percent": 8.3
}
}
}
]
}
Your client’s visibility jumped 30% while their competitor only grew 8%. That’s a story.
Sentiment tracking
Period comparisons work the same way for sentiment:
json
{
"sentiment_score": {
"current": 0.42,
"previous": 0.31,
"change_points": 0.11,
"change_percent": 35.5
},
"positive_count": {
"current": 15,
"previous": 10,
"change_points": 5,
"change_percent": 50.0
}
}
“Sentiment score improved 36%, with positive mentions up 50%.”
Getting started
If you’ve already connected Knowatoa to Claude, ChatGPT, or another MCP-compatible assistant, period comparisons work automatically. Just add date ranges to your questions:
- “Show visibility for September 2026”
- “Compare sentiment between this month and last month”
- “How did competitor mentions change in Q3?”
Your AI assistant handles the date math and calls the right MCP tool with the right parameters.
If you haven’t connected the MCP yet, setup takes about a minute. Check out our MCP docs for step-by-step guides for Claude, ChatGPT, Codex, and more.
Which tools support period comparisons
All of them:
-
get_visibility— Mention rates and ranking positions -
get_sentiment— Sentiment scores and positive/neutral/negative counts -
get_citations— Citation domains and categories -
get_competitors— Per-brand visibility and sentiment -
get_series_performance— Full breakdown per AI service (dates required) -
get_question_performance— Per-question × AI service (defaults to last 30 days) -
get_competitor_sentiment— Detailed sentiment per competitor -
get_negative_mentions— Negative sentiment diagnostics
More details: Period-over-Period Reporting in the MCP
This is the kind of reporting feature that helps agencies justify their work and keep clients. Instead of exporting CSVs and building pivot tables, you ask Claude a question and get the comparison instantly. Then you spend your time on what matters: explaining the trend and recommending what to do next.