A campaign report becomes hard to trust when the search terms change halfway through it. Fix the name, slogan, hashtag, and date range first. Then every returned post can be traced back to the query that found it.
SocialListeningAPI searches LinkedIn, X, Reddit, TikTok, YouTube, and Google through the MCP. It returns normalized records with source URLs. Claude or ChatGPT can prepare the brief, while scheduling, history, and calculations stay in external tools.
Step 1: Freeze the campaign monitoring query list
Create one query for each official and common variation:
- Campaign name
- Campaign hashtag with and without
# - Slogan
- Product and campaign name together
- Event, partner, or spokesperson name
- Common misspelling
Record the campaign start date and end date. Keep brand-only searches separate so normal brand mentions do not inflate campaign results.
Step 2: Search through the MCP
Create a SocialListeningAPI account, claim the 100 free credits, and connect the MCP to Claude or ChatGPT through OAuth.
Use a prompt that requires both sources and limits:
Search LinkedIn, X, Reddit, TikTok, YouTube, and Google for each campaign phrase below. Return the
platform, matched phrase, published date, author, available engagement, short context, and source
URL. Do not estimate total reach. State when a source returns no useful matches.
Campaign phrases:
- campaign name
- #campaignname
- campaign sloganOne phrase across these six sources costs up to 7 credits. Run the full query set once and remove phrases that mostly find unrelated posts.
Step 3: Keep campaign history outside the MCP
An external scheduler can start each approved run. Store the source URL, query, platform, published time, first-seen time, and available engagement in an external database. Deduplicate by source URL.
Calculate counts and changes outside SocialListeningAPI. Returned results are a directional sample, not a complete measure of every campaign impression or conversation.
Step 4: Write the sourced campaign report
Group results into questions, creator posts, partner posts, praise, complaints, and unrelated matches. Include representative URLs and the exact query set. Have a person verify any claim about performance, reputation, or customer response.
Any Reddit reply needs human approval after reading the community rules. Posting and reply drafting are not SocialListeningAPI features.
Conclusion
Use the same query list and storage rule for the whole campaign. That gives you a comparable report with source examples, even when source coverage and available engagement differ by platform.
Continue with Build Brand Mention Alerts With an API.
