Every day we look at millions of real design choices and stitch three streams of data together to spot what’s gaining momentum::
Frequency on its own isn’t enough. We employ GPT models to interpret context and intent, filtering out one‑off experiments and brand-specific outliers. This way, the colours we highlight reflect genuine cross‑project trends rather than noise.:
Raw counts don’t tell the whole story. Our language model looks at how and why each colour is used - whether it’s a background, a button, body text or an accent. That helps us spot patterns:
In short, the model helps us interpret the data; it doesn’t make the trend decisions itself.
A hue earns a spot on the trending list when:
Because we track usage over time, the trends evolve naturally, reflecting what’s happening in the design world right now.
Our trending list is refreshed every week. We continuously collect new data from colour picks, website scans and palette images, then recompute the trends on a regular cadence. This way, the colours you see reflect what designers are actually using right now, without overreacting to short‑lived spikes.