Public methodology — how every forecast is built

Forecasts that
show their work

Every Color Intelligence prediction is built inductively from cited evidence — runway looks, social signals, and consumer emotion drivers. Nothing is asserted that cannot be traced.

We never read industry forecasts

No WGSN, no trend-agency reports, no Pantone seasonal announcements go into our pipeline — as input, calibration, or sanity check. If we consumed industry forecasts, our predictions would converge on the same consensus everyone already has. Instead, every call is derived from primary evidence of what consumers are feeling and wearing. That independence is the point: when we agree with the consensus, it is corroboration; when we disagree, it is information you cannot get anywhere else.

From scope to scored forecast

Six steps, each leaving a trail you can inspect.

01

Scope

A forecast starts from a market scope — category, demographic, region, season — not from a color. We never start from "what color is trending" because that question already assumes someone else's answer.

02

Evidence gathering

For that scope we collect primary signals: runway looks from recent seasons, social conversation and visual signals, and cultural shifts visible in consumer behavior. Each signal is stored with its source so it can be cited later.

03

Emotion mapping (P2VP)

Signals are mapped to consumer emotion drivers using our P2VP framework — 25 emotions organized in 9 clusters across 3 groups. Emotion is the lens that explains why a signal matters: color adoption follows what consumers are feeling, not what the industry publishes.

04

Color derivation

Colors are derived inductively from the cited evidence — the palette is what the evidence supports, not a curation. If the evidence for a color is thin, the color does not make the forecast.

05

Confidence scoring

Each prediction carries a confidence score computed from its evidence: how much evidence supports it, how diverse the sources are, how recent the signals are, and whether independent signal types corroborate each other. The score is never an opinion — it is a property of the evidence shown next to it.

06

Public scoring

Seasonal calls are logged publicly before the season starts, then scored after the season lands — hits and misses alike. The track record is the methodology's accountability loop.

What the confidence score means

A confidence number with no visible basis is marketing. Ours is a function of the evidence displayed beside it — remove the evidence and the score falls.

Evidence volume

How many independent signals support the prediction.

Source diversity

Runway, social, and cultural signals weigh more together than any one alone.

Recency

Fresh signals count more; stale signals decay out of the score.

Corroboration

When unrelated signal types point at the same emotion driver, confidence rises.

The citation promise

Every prediction in a Color Intelligence report lists the runway looks, social signals, and P2VP emotion drivers it rests on. If you cannot see why we called a color, treat the call as unmade — and tell us.

Hold us to it

We log our seasonal calls publicly before each season and score them after it lands — hits and misses.

See the track recordTry Color Intelligence