Social Analyst

The Social Analyst is one of the five Research analysts (the social listening analyst, with deep coverage of e-commerce platforms like Amazon). It reads how people actually talk and behave on social and commerce platforms, then synthesizes the cross-channel picture into a single output. Where the Category and Consumer Analysts draw from published reports, this analyst works from platform-native material: posts, threads, review themes, creator content, and long-form cultural commentary.

What it does

The analyst produces a unified social listening summary with an executive summary plus five main sections: four thematic sections (each anchored on a tension or consumer truth, supported by platform-native blockquotes from the channels where the signal is richest) and a final Other Social Insights section that captures cross-channel patterns the four themes don't cover.

Citations are platform-native. The thematic sections trace back to actual posts, Reddit threads, Amazon review patterns, or creator content rather than generic summaries. Each blockquote carries a platform tag, so you can see which channel contributed which insight and where the same signal surfaces across multiple platforms. Output length is roughly a one-pager equivalent of structured prose and tables.

Where to find it

Inside a brand or campaign, navigate to Research. After the Research run completes, the Social Analyst output appears as the fifth of five analyst tabs on the Research page.

How it works

When you start a Research run, all five analysts run in parallel. The Social Analyst reads the brand profile, the campaign brief if you're running campaign-scoped Research, and any attached documents. It then queries social-listening tools across the platforms below. Each platform gets multiple queries probing emotional triggers, routines, tensions, and identity-building behavior in your category.

The analyst pulls from at least three of the following five platforms based on where behavioral signal is richest for your category:

• Reddit — community discussion, critical discourse, price scrutiny, and category-specific subreddits.
• Instagram — lifestyle aesthetics, visual signals, creator content, and brand-positioning cues.
• TikTok — emotional storytelling, trends, creator culture, and how categories show up in short-form video.
• Amazon — review themes, purchase validation, complaint patterns, and gifting behaviors.
• Medium / Substack — reflective long-form analysis and cultural commentary on the category.

No single platform dominates the citations. The mix reflects where consumers in your category actually engage. Findings are synthesized into per-platform notes, then the cross-channel summary.

The full Research run takes about 6-8 minutes. The Social Analyst's runtime varies depending on how many platforms have rich signal for your category.

What you can adjust

Document attachment — before starting a Research run, select documents from the Documents page (PDFs, Word files, Markdown, CSVs, Excel sheets) to attach. The analyst treats them as high-priority input alongside the platform queries. Prior social listening reports, audience research with quoted posts, or platform-specific audits are particularly useful here.

Refine — after a run completes, submit feedback via the Refine button. The platform routes it to the analysts that need to re-run. Each Refine creates a new version. Prior versions stay in history and can be restored.

Limits and known issues

• Public posts only. The analyst sees what's publicly indexed on each platform. Private accounts, closed communities, and gated forum content are not readable.
• Platform indexing varies. Reddit and Amazon reviews are well-indexed for English-language content. TikTok's discoverability is shallower for niche categories. Instagram coverage depends heavily on hashtag and creator surfacing. The output flags which platforms returned thinner signal for your category.
• Snapshot, not real-time. The signals are read at run time and reflect what was visible then. Cultural moments that broke after the run may not appear. Re-run periodically for time-sensitive work.
• English-first source bias. The social listening leans English-language. For non-English markets, the platform mix typically shifts (less Medium/Substack, more region-specific platforms when available) but the bias remains.
• Beta software. Validate cited posts and the framing of any sensitive thread before briefing teams or committing budgets.