A phrase starts appearing
Captions are mined for product names whose arrival rate jumps. Searching for brands you already know only finds brands you already know.
Chris Camillo's idea is that consumer behaviour is visible to ordinary people months before it reaches a filing. The hard part was never noticing that something is popular. It is knowing whether popular means anything. Here is the method SociArb uses to decide, and the checks that stop most candidates.
SociArb keeps watching until the trend, sources, company connection, stock, and timing are strong enough. Weak ideas are held back.
Captions are mined for product names whose arrival rate jumps. Searching for brands you already know only finds brands you already know.
Attention is scored against that brand's trailing 90 days, not against an absolute view count. A big brand's ordinary week beats any fixed threshold.
If every beauty product spikes the same day, that is seasonality or an algorithm change. The category's move and the platform's move both come off.
Purchase language, people saying it is sold out, complaints. A clip can take a million views while nobody underneath it wants the product.
The product is matched to an SEC-registered issuer, and the issuer's own filings size how much revenue could plausibly be exposed.
Earnings inside three days, a stock that already ran, seeded campaigns, and thin creator breadth all stop an alert.
Can you open the posts, and were they measured more than once?
Is attention accelerating, and are the comments about buying?
How much of this company's revenue could the product touch?
Is this the right issuer, and is that reproducible from the product name?
Are earnings close, or has the stock already moved?
The scores help organize the research. A serious risk can still stop the alert.
The same tabs and checks must be rebuilt whenever the evidence changes.
The sources, company, stock price, earnings date, and updates stay together.