Data & scoring methodology
Here's exactly how scores are built, and what is sample data today.
Where the data comes from
Today, ads shown in AdSpindle come from a clearly-labelled sample catalogue. A live integration (TikTok Shop product and creator-video data) is in development; this page will be updated when it ships.
Product-level sales and revenue in the sample catalogue are illustrative estimates. Once live ingestion ships, revenue will be estimated from sold-count deltas multiplied by list price — still directional, not accounting data.
The catalogue is populated with a deterministic synthetic dataset so you can evaluate every feature. It is modelled on TikTok Shop-style dynamics but is not live market data.
Trend score (0–100)
Trend scores combine sales momentum, growth rate and lifecycle position (exact weights may change as the formula ships). When live ingestion runs, scores are recomputed on each ingest and stored to power trend charts and 7/30-day deltas. The public sample catalogue uses modelled scores and history for demonstration.
Lifecycle phases — emerging, rising, peaking, saturated, declining — are derived from where a product sits on its own trajectory, not from absolute score. A product can score 60 and be a better opportunity than one scoring 90 on the way down.
Conversion score (0–100)
Ad conversion scores estimate creative strength: hook quality, angle clarity, audience fit and production signals. When an ad is analysed with AI, the model reviews the actual creative (image) plus its copy and metadata, and the structured breakdown — hook, angle, audience, strengths, weaknesses, recommendations — is stored and shared across the platform.
Scores are comparative tools for prioritising research, not guarantees of performance.
Freshness
When live ingestion is connected, the catalogue is scheduled to ingest daily at 06:00 UTC and trend metrics append without rewriting past rows. Today the demo and public pages use a fixed sample catalogue. AI analyses are cached for 7 days and can be re-run on demand.
Limitations we're upfront about
Revenue estimates are modelled, not reported. Once live ingestion ships, data will lag the source by up to a day. Conversion scores are model judgements — pressure-test them against your own market knowledge before spending. This site currently shows sample data only; do not treat demo or public trend pages as live TikTok Shop performance.