Weekly share-of-voice report

Set it up

https://app.octolens.com/api/mcp/v2

Set up a weekly recurring task (Monday mornings) using the playbook below. Brand: "{BRAND_NAME}". Competitors: {COMPETITORS}.

What it does

• ✓Computes mentions and share of voice for your brand and each competitor, week over week
• ✓Tracks sentiment splits and per-platform volume
• ✓Flags share shifts of 2+ points and volume moves over ±30%
• ✓Pulls example mentions to explain every spike — no number without a cause
• ✓Writes a sub-400-word report with a one-line takeaway on top
• ✓Says 'nothing meaningful changed' when nothing meaningful changed

Mentions come from every platform Octolens monitors:RedditX (Twitter)LinkedInHacker NewsGitHubYouTubeBlueskyStack OverflowTikTokDEV.toNewsNewslettersPodcastsWeb

Delivers: A weekly share-of-voice report: table, deltas, sentiment shifts, and the quoted, linked mentions behind every change.

Inputs

How your agent runs it

1. 1

   Load the schema — The agent pulls the analytics schema and maps keyword IDs for your brand and competitors.

   analytics_contextlist_keywords
2. 2

   Run the numbers — SQL over the mentions data: counts, share of voice, sentiment split, and platform breakdown — this week vs last.

   analytics
3. 3

   Spot the changes — Share shifts of 2+ points and ±30% volume moves get flagged for explanation.

4. 4

   Find the cause — For every flagged change, pulls the actual mentions that drove it — quoted and linked.

   list_mentions
5. 5

   Write the report — Takeaway first, then the table, sentiment, evidence, and a watchlist. Under 400 words.

Tools used

• analytics_contextOctolens MCPGet the analytics schema and valid filter values
• analyticsOctolens MCPRun read-only SQL queries for counts, shares, and trends
• list_keywordsOctolens MCPMap keyword IDs to brand and competitor names
• list_mentionsOctolens MCPPull the example mentions behind each notable change

All 23 Octolens MCP tools are documented on the MCP page and in the docs.

The full playbook

Replace the placeholders, then paste this into your agent as a recurring task.

Every week, do the following:
1. Call analytics_context to load the analytics schema, valid filter values, and SQL tips. Call list_keywords to map keyword IDs to names — you need "{BRAND_NAME}" plus the competitor keywords: {COMPETITORS}.
2. Using the analytics tool, compute for the last 7 full days and the 7 days before that:
   - mention count per keyword (brand and each competitor)
   - share of voice: each keyword's mentions as a share of the tracked total
   - sentiment split (positive / neutral / negative) per keyword
   - mention count per platform for "{BRAND_NAME}"
3. Identify what changed:
   - share-of-voice gains or losses of 2+ percentage points week over week
   - any keyword whose mentions moved more than ±30%
   - any platform where brand mentions spiked
4. For each notable change, pull 1–3 example mentions via list_mentions
   (sorted by relevance) to explain what actually drove it. Quote them,
   with links. Never report a spike without showing what caused it.
5. Write the report in this structure:
   **Share of voice — week of {date}**
   1. The one-line takeaway (who gained, who lost, why)
   2. Share-of-voice table: keyword | mentions | share | Δ vs last week
   3. Sentiment: anything that shifted meaningfully, with an example quote
   4. What drove the changes: the evidence from step 4, quoted and linked
   5. Watchlist: anything to keep an eye on next week
6. Post the report in chat.
Keep it under 400 words excluding the table. Numbers first, adjectives
never — if nothing meaningful changed, say so in one line and keep the
report short.