Put the next yuan where it worked.
See which creators and content carried the campaign, where the budget leaked and what the audience asked for, before the next flight is planned.
See how brands use it
What the film shows
Every campaign, case and figure in this film is illustrative, apart from the four figures credited to a source.
- All the world's a stage (As You Like It, Act II, Scene VII). Give enough monkeys enough typewriters… …and one of them will type Shakespeare. Today, millions of creators are typing. Very little of it is random.
- ¥84 billion went into creator content in China in 2025 (source: Topklout, June 2026). 74% of advertisers focus their social spend on creators. Advertisers' #1 challenge: measuring whether it worked (source: Miaozhen Systems, 2025).
- Wrap report, Spring serum launch (illustrative data). Campaign results: views 12.4M, likes 486K, creators 41. Success. Every wrap report says it worked. So where did the money go?
- Watched. Heard. Read. All of it. 62 campaign posts on Xiaohongshu and Douyin, read in full; 1,222 other posts by the same 41 creators, measured for their baselines (illustrative). On screen: “72-hour hydration”. Speech: “The texture is really light.” Comments: “Light texture, absorbs fast”, “Where can I buy it in store?”, “A bit pricey. Waiting for a sale.”, “Just right for an 8 a.m. start”. Every creator, fingerprinted (each campaign post set against the creator's own recent posts).
- 14,206 comments, read one by one, by aspect, sentiment and purchase intent (illustrative). 170 million people a month ask Xiaohongshu what to buy (170 million: Xiaohongshu, April 2025). 990 asked where to buy this serum (illustrative). Nobody answered.
- The full read, in about 30 minutes. In English and Chinese. While there's still budget to move. Report ready: 0:28:40. Intelligence report, Spring serum launch, flight 1, skincare, Xiaohongshu and Douyin, 12 March to 9 April (illustrative): score 58, 62 posts, 41 creators, 14,206 comments. One 48-second demo drove 38% of purchase questions. The two largest accounts took ¥1.2M and missed their own baselines. 990 people asked where to buy it; nobody answered. Recommendation: move ¥1.0M to mid-tier skincare specialists, lead with the demo, and answer “where to buy” in every post. Your own data: GMV, members and conversion, to be read here once they can be connected (shown dashed, as a future connection).
- ¥1.2M went to two accounts (illustrative). Both below their own baseline. Budget ¥3.0M: top accounts (2 creators), mid-tier specialists (19 creators), everyday creators (20 creators). Lead with the 48-second demo: 38% of purchase questions came from one demo. Answer “where to buy” in every post: 990 asked where to buy, nobody answered. Flight 2: the same budget, moved to what worked.
- Same budget. Before: 58, middle of the pack. After: 78, top quartile (unrandom Score, skincare launches; illustrative).
- Eight illustrative cases. Product launches. Celebrity events. Event launches. Seeding. Paid media. Pitches. Head to head. Your own data. Product launch, Spring serum launch: unrandom Score 78, the absorption demo carried the launch. Celebrity event, Ambassador night, Shanghai: 1 in 9 comments mentioned the product; the rest were about the ambassador (illustrative; telling which product or person a comment is about is still in development). Event launch, Brand pop-up, Chengdu: 3 weeks of “where is it” questions after the store-visit vlogs. Seeding, gifting wave, 140 creators: 2.1×, seeded posts against their creators' own median. Paid media, autumn paid flight: ¥0.84 paid cost per engagement, from 38 of 41 paid posts. Pitch, sun care, five summers: 30,000+ teardowns searched for what worked in the category. Head to head, two launches, one category: “Promise vs response”, what creators said against what audiences answered, by aspect. Your own data, GMV, members, conversion: every campaign read against your own numbers once they can be connected (shown dashed, as a future connection). A departures board shows illustrative GMV, member growth, conversion, repeat purchase, store traffic and basket size. Up to 40% of skincare purchases in China can be attributed to influencers (BCG, August 2025). When your own data can be connected, you will see yours here (a future connection, shown dashed).
- Brands: put the next yuan where it worked. Agencies: walk into every review with the answer.
- unrandom. The social intelligence platform. unrandom.co
The social intelligence platform for brands and agencies.
Every day, China's social platforms record what consumers want, doubt and buy. unrandom reads that record at scale and turns it into decisions: where to enter, what to launch, who to back and where to spend.
- Enter the China market with the category's record already read.
- Follow a celebrity, a brand, a product or an event on the schedule you set.
- Put creator budgets where they work.
- Build a custom program for your own question, from regulated categories to connecting your own data.
Walk into every review with the answer.
Prepare the pitch from the category's record, follow live campaigns through scheduled Tracking, and answer questions in the room with findings that trace back to the posts.
See how agencies use itBring us one recent campaign.
We will show you what to change before the next flight.