Analytics

Detailed stats: real-time KPIs, 30-day sparkline, rule distribution, platform and source comparison.

Help Center / Analytics
What can I find on the Analytics page?

The Analytics page is GeniReply's advanced control panel. At the top you'll find 4 KPI cards giving you the situation at a glance:

  • Total comments — the overall number of comments intercepted by GeniReply across all platforms
  • Today — how many comments arrived on the current day
  • This week — volume over the last 7 days
  • Most active rule — the rule with the most matches overall, with a count

Below the KPI cards you'll find the 30-day sparkline chart: an SVG curve showing the daily comment volume over the last month. Hover over points to see exact values.

Further down you'll find three detail panels:

  • Rule distribution — bar chart with the top 10 most active rules and their match counts
  • By platform — Facebook vs Instagram comparison, to understand where your volume comes from
  • By source — Live (real-time polling) vs Historical import (backfill), to distinguish new comments from imported ones

Finally, the Reply statistics panel shows how many comments received an automated reply and how many remain unanswered.

What do the 4 KPI cards at the top of the page mean?

The KPI cards are real-time counters updated each time you open the Analytics page:

  • Total comments — all comments in the database, including those imported via historical backfill. This is the baseline for understanding your overall managed volume
  • Today — comments received via live polling on the current day (from 00:00 UTC). If it's zero, no new comments arrived today
  • This week — comments from the last 7 calendar days. Useful for weekly briefings
  • Most active rule — the name of the rule with the highest total match count, with the number beside it. If one rule dominates with very high percentages, it might be too generic
💡 If "Today" and "This week" are both zero but "Total comments" is high, your comments likely come from the historical import and live polling isn't receiving new comments yet. Check Connected Pages status.
How do I read the 30-day sparkline chart?

The sparkline chart shows a curve of daily comment volume over the last 30 days. The horizontal axis is time (one point per day) and the vertical axis is the comment count.

What to look for:

  • Sudden spikes — often coincide with new ad campaigns, viral posts or influencer mentions
  • Drops to zero — may indicate a problem with the Meta connection (webhook disabled, token expired) — check Connected Pages
  • Upward trend — good sign! More managed comments means more engagement under control

If the chart shows a high initial spike followed by low values, it's likely the effect of the historical import (backfill) that loaded past comments in bulk.

Can I export Analytics data to Excel/CSV?

From the Analytics page, click the Export CSV button in the top right. The downloaded CSV file includes all interactions in the selected period with these columns:

  • Date, Author, Comment text
  • Platform (Facebook/Instagram), Source (Live/Import)
  • Matched rule, Confidence, Actions performed

The CSV opens directly in Excel, Google Sheets or Numbers. Numbers and dates follow the international standard for maximum compatibility.

💡 For a complete report with charts, also use the Export PDF Report button which generates a printable document with KPIs, daily trends and rule distribution.
How do I filter statistics by time period?

At the top of the Analytics page you'll find a selector with 5 time ranges:

  • Today — current day only
  • 7 days — last week
  • 30 days — last month (default)
  • 90 days — last quarter
  • All time — from the very beginning, all interactions in the database

The filter applies to all sections on the page: KPI cards, sparkline chart, rule distribution, platforms and sources. Changing the range updates the data immediately.

💡 Use "7 days" for your weekly briefing and "90 days" for your quarterly client report.
What does the AI Learning section in Analytics show?

The AI Learning section at the bottom of the Analytics page gives you visibility into how GeniReply is learning from your corrections. It includes:

  • Total corrections — how many times you've reclassified a comment from the wrong rule to the correct one
  • Confusion patterns — pairs of rules the AI confuses most often (e.g. "Rule 3 → Rule 7" with N occurrences). Helps you identify rules with overlapping keywords
  • 30-day timeline — the trend of corrections over time. A downward trend means the AI is improving

The AI Learning section data is refreshed each time you open the page and respects the selected time filter.

💡 If you spot a pair of rules with many confusions, try making one rule's keywords much more specific, or use exceptions to distinguish them.
What is 'confidence' and how do I use it?

Confidence is a 0–100% value that indicates how sure the AI is about the matched rule. Guidelines:

  • ≥ 95% — reliable classification, standard action is safe
  • 70–94% — likely match, worth sampling in the Inbox
  • < 70% — the system uses the 🙈 Hide action (soft-hide) and asks you to confirm instead of acting on its own

In Analytics you'll find the confidence distribution: if you see too many comments below 70% it means the rules have keywords that are too generic.