Retail sentiment data shows the percentage of retail trading accounts positioned long versus short on a given pair, aggregated across brokers. Unlike COT (which reflects institutional futures positioning), this captures everyday individual traders.
Retail positioning is one of the few datasets in trading that's commonly used as a contrarian indicator — meaning when it reaches an extreme, it's often read as a signal against the crowd rather than with it. The logic: retail traders as a group have historically tended to be positioned wrong at extremes, in part because the same patterns that draw a lot of retail attention to one side of a trade (a strong recent trend, a popular narrative) are often already mature by the time positioning becomes that lopsided. A common threshold is treating anything above roughly 75% one-sided as a strong contrarian flag, and 60-75% as a milder lean.