2026 campaign-assessment score
How Mediaforce (百分点舆情) scored
56 / 100
Mediaforce pairs public-opinion monitoring, including video recognition and named Chinese and overseas networks, with communication-effect models; no report output is public.
| Criterion and evidence | Points |
|---|---|
Campaign explanationVideo analysis reads covers, embedded text and narration and finds logos, products and spokespeople, text analysis extracts selling points, and brand communication effect evaluation is a named service. No output shows these connected across one campaign. | 18 / 30 |
Reaction analysisPosts and user comments from Douyin, Kuaishou and Xiaohongshu feed sentiment judgment, clustering and entity tagging. The page shows no aspect-level results and does not separate creator reaction from product reaction. | 12 / 20 |
Finished assessmentSystem analysis reports, analyst briefings and event review reports are listed. The page includes no sample, and the balance between generated and human-written work is unstated. | 12 / 20 |
Creator baselineFour named models cover reach, expected effect, media friendliness and self-media influence, and competitor comparison is listed. The formulas are unpublished, and no model compares a post with that creator's usual results. | 5 / 10 |
China and international coverageWeChat, Weibo, Douyin, Kuaishou and Xiaohongshu are named alongside Facebook and Twitter. The page gives no detail on depth or access on either side. | 6 / 10 |
Source trailThe page describes source monitoring and retrieval. It shows no path from a report finding to the post behind it. | 3 / 10 |
Evidence differs by product. No product was tested head to head. Scoring methodology · Publisher disclaimer
Mediaforce is the public-opinion system from Percent Technology (百分点科技). It treats campaign reach and reputation risk as one job. Percent Technology
Its monitoring covers WeChat, Weibo and short video within 14 categories of social platform, posts and comments from Douyin, Kuaishou and Xiaohongshu, more than 6,000 TV programs, domestic and foreign news sites, and overseas networks such as Facebook and Twitter. The video layer reads cover images, embedded text and voice narration, and can look for a particular logo, product or spokesperson. Text analysis identifies entities, clusters articles, extracts communication selling points and estimates paid-poster activity. Percent states sentiment accuracy above 98%, a vendor figure with no published test. Mediaforce
For the measurement question, Percent lists four named models: reach (ROC), expected communication effect (EOC), media friendliness (DMF) and a self-media influence index (WMI). They are offered for evaluating marketing communication, media relations and media buying. Brand communication effect evaluation is sold as quantitative and qualitative research for brand, marketing and PR teams, and the deliverables listed are system reports, analyst briefings and event review reports. Mediaforce
The video recognition stands out. A team can find a logo inside a video when the caption says nothing about it. Percent publishes the model names without their formulas and offers no sample report, so a campaign lead has little to show how a score was built or which posts stand behind it.
Mediaforce monitors a brand’s public exposure across Chinese and overseas channels. unrandom starts from the posts in one creator campaign and writes the assessment of that campaign. unrandom product