How Kitsunoko scores products
Every score on Kitsunoko is derived from five real data sources — no black box, no proprietary magic. Here's exactly what we measure, how each signal contributes, and where the model falls short.
The short version
We query five platforms every two hours — TikTok, Google Trends, Amazon, Reddit, and news APIs. Each platform returns a raw signal (search velocity, BSR rank change, mention count, etc.) that we normalise to a 0–100 scale and combine into a single Trend Score using a weighted average. A score above 65 triggers a GO DEEPER recommendation — meaning the organic demand signal is strong enough that we think the product warrants serious evaluation.
35%
TikTok velocity
28%
Google Trends slope
22%
Amazon BSR movement
10%
Reddit sentiment
5%
News coverage
* Weights are approximate and tuned periodically based on signal accuracy. TikTok carries the most weight for fast-moving consumer goods; Google Trends is the strongest predictor for evergreen products.
The five signal sources
TikTok
Highest weight
We track keyword search velocity and hashtag momentum on TikTok Shop and TikTok organic. TikTok is the single strongest leading indicator for fast-moving consumer goods — a product that starts trending here typically reaches Amazon and Google 4–8 weeks later. We measure how fast search volume is accelerating, not just its raw level.
What we measure
- ✓Search velocity (week-over-week acceleration)
- ✓Hashtag view count growth
- ✓Video upload rate for product-related terms
Google Trends
High weight
Google Trends provides the broadest consumer intent signal across all demographics. We pull 7-day, 30-day, and 90-day relative search interest for each product keyword and compute the momentum slope — a product rising from 20 → 60 interest in 30 days scores higher than one sitting at 80 with no movement.
What we measure
- ✓30-day relative interest slope
- ✓Geographic spread (narrow geo = early trend; wide geo = peaking)
- ✓Related query breakouts ('rising' terms near the product)
Amazon BSR
Medium-high weight
Amazon Best Seller Rank gives us purchase intent, not just curiosity. A product climbing BSR means real buyers are converting, not just searching. We weight upward BSR movement more than absolute rank — a product moving from BSR 80,000 to 20,000 in 30 days signals stronger opportunity than one parked at BSR 5,000 with no movement.
What we measure
- ✓BSR rank change over 30 days
- ✓Review velocity (new reviews per week)
- ✓Price spread vs competitor listings
Medium weight
Reddit is the early adopter signal. Authentic discussion of a product in relevant subreddits — before it hits mainstream channels — is a reliable leading indicator of organic demand. We look at post count, upvote ratio, and comment sentiment across dropshipping, product review, and niche-specific communities.
What we measure
- ✓Post count in relevant subreddits (last 30 days)
- ✓Upvote ratio as a proxy for genuine enthusiasm
- ✓Comment sentiment (positive vs negative mentions)
News & Media
Supporting signal
News coverage can accelerate a trend's mainstream adoption. We scan news APIs for product category mentions, press coverage of viral products, and brand mentions. This signal carries the lowest weight because it lags social signals — a product appears in news after it has already started trending, not before.
What we measure
- ✓News article count mentioning the product or category
- ✓Publication quality weighting (major outlets vs blogs)
- ✓Coverage freshness (last 7 days vs last 30 days)
Update frequency
Every 2 hours
Pipeline runs on a fixed schedule — scores refresh automatically.
Product images
RapidAPI / eBay CDN
Product photos sourced via Real-Time Product Search API. Attribution shown when an external image is displayed.
Supplier data
CJDropshipping API
Verified supplier pricing, MOQ, and ship times pulled directly from CJDropshipping's live catalog.
What the decisions mean
GO DEEPER
Score 65–100
Strong organic demand across multiple platforms. Investigate supplier cost, competition density, and ad angles. These products have a real window — but act before it closes.
MONITOR
Score 40–64
Rising signal but not yet at inflection. Add to your watchlist and check again in 7–14 days. Could either break out or fade — more data needed.
PASS
Score 0–39
Insufficient organic signal to recommend investigation right now. May still be a valid product — just not trending in a way our data sources capture.
Known limitations
Organic signals only
Scores reflect organic consumer interest — not ad spend. A heavily advertised product with low organic interest will score lower than a genuinely viral product. This is intentional: we measure what consumers want, not what brands are pushing.
English-language sources
Our current signal sources are English-first. Products trending primarily in non-English markets (e.g., a product viral in Japan but not yet translated to English content) may be underscored.
Niche products may score lower
A product with a small, passionate audience (e.g., a specific fly-fishing accessory) will score lower than a mass-market equivalent even if conversion rates in that niche are excellent. The 0–100 scale is calibrated to general consumer goods.
2-hour data latency
Scores update every 2 hours. A product can go viral in the 2-hour gap between runs. If you see a product exploding on TikTok right now, its Kitsunoko score may not reflect that for up to 2 hours.
Competition scores are estimate-based
Competition labels (Low / Medium / High) are estimated from the number of active sellers we detect on Amazon and our platform crawl — not from a real-time marketplace API. Treat them as directional, not precise.
See the scores live
Every product on our platform comes with its individual signal breakdown — so you can see exactly which sources are driving each score.