I used to develop and sell novel earnings forecasts which would outperform analyst consensus. Typically my research would have a <1% margin of error on forecasting a specific KPI directly relevant to corporate revenue and profit for an equity. The best hedge funds were able to successfully trade on this information because they combined it with a significant amount of complementary data.
But I would not want to develop a trading strategy based on that data alone. These guys did very well considering the information they had to work with and how primitive trading earnings is. The real money in insider trading is in more illiquid OTC markets, or in mergers/acquisitions.