We use rich social media data and large language models to classify retail investors’ trading strategies, showing that strategy adoption is highly dynamic, and responds to news, past performance, and social feedback. Sentiment from fundamental analysis posts positively predicts future returns, whereas sentiment from technical or other strategy posts negatively predicts returns. Technical sentiment is strongly correlated with net retail buying, especially among Robinhood investors. Retail order flows are substantially more informative when linked to fundamental analysis sentiment. These results demonstrate how retail investors form strategies and the conditions under which strategy adoption enhances or diminishes their order flow informativeness.