A present invention for estimating a confidence interval includes: receiving a dataset comprising information on software product defects found by a first customer and a second customer, the information comprising numbers of cases of software product defects found by the first customer and the second customer; determining a probability for each case of found software product defect using the frequencies and a total number of software product defects found; determining a plurality of bootstrapping draws from the cases of found software product defects in the dataset; determining an array of conditional probabilities that the first customer may find a given software product defect given that the second customer found the given software product defect, using the numbers of the cases and the probabilities corresponding to the cases in each bootstrapping draw; and determining a confidence interval for the determined plurality of arrays of conditional probabilities.