Engagement Insights
Understand how comments, mentions, and DMs contribute to your engagement metrics — and how SonetHub measures response rate, sentiment, and top-performing posts.
Engagement Insights
The Engagement page surfaces a layer of insight on top of raw mention and comment data: aggregated trends, sentiment over time, response performance, and the posts generating the most conversation. It is distinct from the post engagement rate shown on the Analytics page (which measures likes + shares relative to reach).
Available on: Growth, Pro, and Business plans (requires the Engagement monitoring feature)
What Counts as Engagement
SonetHub tracks three types of inbound interaction as "engagement":
| Type | Description | Example |
|---|---|---|
| Comment | Someone leaves a comment on one of your posts | "Love this product!" on an Instagram post |
| Mention | Someone @-mentions your account in their own post or story reply — not on your content | A tweet that says "@YourBrand great job!" |
| DM (Direct Message) | A direct message sent to your account inbox | A message asking about pricing |
Note: DMs are tracked by the Inbox feature and included in response-rate calculations only when they have been read or replied to inside SonetHub. DM content never appears in the Engagement dashboard — only counts.
All three types share the same triage and sentiment pipeline. Which types appear in the Engagement dashboard depends on the platform:
- Comments on your posts — available on all platforms that support comment retrieval (Instagram, Facebook, Threads, YouTube, LinkedIn, TikTok Business, Bluesky).
- External @-mentions — available primarily on Twitter/X (via the mentions timeline). Instagram and Facebook deliver mentions via webhook in real time; other platforms do not expose a general mention search.
- DMs — response rate only; content is private.
Metric Definitions
Volume
The total count of comments + mentions (and optionally DMs) received in the selected time period. Volume is the raw signal; it tells you how busy the period was but not whether the activity was positive or required action.
Sentiment Split
Every inbound item is classified by AI as Positive, Neutral, or Negative immediately after arrival. The split is displayed as a ratio (e.g. 42 / 31 / 12) and as a time-series chart. Items where the AI cannot determine sentiment (very short text, emoji-only, unsupported language) are excluded from the three counts but included in the volume total.
- Positive — appreciation, praise, excitement, support
- Neutral — questions, informational comments, uncategorised
- Negative — complaints, criticism, frustration, negative comparisons
Response Rate
Response Rate = replied items ÷ total items
An item is counted as "replied" when SonetHub records a replied_at timestamp — meaning a reply was sent through SonetHub (via the Engagement inbox, the AI agent, or an automation rule). Replies made directly on the platform without going through SonetHub are not reflected here.
Response rate covers comments and mentions. DMs are excluded unless a reply was sent from the SonetHub Inbox.
Median Response Time
The median time between when an item arrived (mentionedAt) and when the first reply was sent (repliedAt), measured in minutes. Only items that have been replied to are included in the median. The median is used rather than the mean because a handful of very slow replies (e.g. items discovered days later) would otherwise inflate the average.
Displayed as minutes (e.g. "14m") or hours (e.g. "2h") depending on the value.
Sentiment Over Time
The chart plots positive, neutral, and negative counts per calendar day across the selected date range. Days with no activity appear as zero. This lets you spot spikes in negative sentiment correlated with a product launch, a viral post, or an external event.
Category Mix
Every item is also classified by category:
| Category | Typical signals |
|---|---|
| Praise | "Amazing", "love it", stars, hearts |
| Complaint | "Broken", "disappointing", "refund", "never again" |
| Question | "How do I…?", "Where can I…?", "Is this available…?" |
| Spam | Unrelated links, repetitive text, known spam patterns |
| General | Everything else |
The category-mix bar chart helps prioritise: a large Complaint or Question bar indicates workload that needs a reply; a large Spam bar suggests an automation rule would reduce noise.
Top Posts Driving Engagement
The Top posts list shows the posts that generated the most inbound engagement during the period. Only posts attributed to one of your published posts (via a post_id link) appear — true external @-mentions that are not associated with a specific post are excluded.
Engagement per post is the sum of likeCount + replyCount across all comments tied to that post. A post with 50 comments each liked once scores higher than a post with 10 comments and no likes.
Clicking a post title navigates to the post detail page.
Most Active People
Shows the authors who engaged most frequently with your accounts in the period. This is volume-based (count of items from each author), not weighted by sentiment or reach.
Sparse-Data Caveat
Engagement data density varies significantly by platform and by the age of your connected accounts:
- Instagram and Facebook deliver mentions via webhook with no historical backfill. When you connect a new account, only mentions that arrive after the connection date appear. Comments on older posts arrive via the 30-minute polling window (last 7 days of posts only).
- Twitter/X is the primary source of true external @-mentions (people talking about you in their own posts). Mention history is limited by the Twitter API tier in use.
- LinkedIn, YouTube, TikTok Business, Bluesky support comments on your own posts but do not expose a general mention search — so the "mention" type for these platforms means only comments, not external references.
- TikTok (regular) and Pinterest do not provide comment or mention access via the Content Posting API.
As a result, the Engagement dashboard is most useful once your accounts have been connected for at least 30 days and have been active during that period.
Plan Requirement
Engagement monitoring and the Insights dashboard require the Growth plan or higher. On Free and Starter plans, the Engagement page is not available.