2026 campaign-assessment score
How Archive scored
74 / 100
Archive turns a finished campaign into written reports with renew-or-drop recommendations and comment analysis, on Instagram, TikTok and YouTube only.
| Criterion and evidence | Points |
|---|---|
Campaign explanationCampaign-level Deep Research ranks creators and reports what resonated, drawing on AI reading of transcripts, captions and video. YouTube filters see only a preview image and metadata, on-screen text is not separately documented, and no complete sample report is public. | 24 / 30 |
Reaction analysisComment analysis reports themes, sentiment, brand mentions, intent and signature comments, and Campaign sentiment covers what resonated or backfired. Post-level sentiment does not separate reaction to the brand from reaction to an influencer. | 16 / 20 |
Finished assessmentCampaign performance reports name creators to renew or drop and the top changes for next time, and are shared as a no-login link, PDF or Markdown. No complete public example could be opened. | 16 / 20 |
Creator baselineCampaign reports give engagement by followers and by views and impressions as a multiple of follower base, Competitor Insights benchmarks EMV and impressions, and creator profiles show averages. Its documentation does not compare a post with that creator's usual results. | 6 / 10 |
China and international coverageArchive captures Instagram, TikTok and YouTube and documents capture limits by content type. No Chinese network is named. | 4 / 10 |
Source trailA report's downloadable context lists source posts with transcripts and retrieved comments, and shared links need no login. Cited posts are selected examples, and Archive warns of occasional dead links. | 8 / 10 |
Evidence differs by product. No product was tested head to head. Scoring methodology · Publisher disclaimer
Archive collects a brand’s creator posts and writes reports on them. A campaign runs as Creator Monitoring, with a known roster and an end date, or Brand Monitoring, which draws on the workspace’s sources and can stay open. Campaigns
AI Filters read each post’s visuals, transcribed speech and caption, then tag it against rules the workspace defines. Tone and music are out of reach, because audio is read only through its transcript. On YouTube the filters see a preview image and metadata. How AI Filters work Archive’s July 2026 announcement says Deep Research recognizes the product, listens to the voiceover and reads the captions. Deep Research announcement
Two of Deep Research’s nine reports are written for a finished campaign. Campaign performance ranks creators best to worst, says who to renew or drop and why, and lists the changes to make next time. Campaign sentiment covers what resonated or backfired. A separate Comment analysis report reads audience themes, sentiment, intent and signature comments. Comments are an optional add-on, capped per post. Reports handle up to 10,000 posts and go out as a no-login link, PDF or Markdown. Deep Research reports
A downloadable context file holds each source post’s caption, transcript, engagement and retrieved comments, down to the commenter’s handle. Archive also warns that reports can occasionally carry dead links.
Post-level sentiment does not separate feeling about the brand from feeling about the influencer. Sentiment Radar, the untagged-content detector, looks only at video cover images and only at creators above 50,000 followers. Radar
Archive captures three networks: Instagram, TikTok and YouTube. Content types
Archive also runs creator programs, from search to usage rights. unrandom focuses on assessing a defined campaign, with Xiaohongshu and Douyin alongside TikTok, Instagram, YouTube and X. unrandom product