We are doing a study on shampoo buyers in Walmart stores across the US and we send them online surveys. Based on past studies, we know that of the people who start the survey, 80% qualify for the study, and only 40% of these people who do qualify actually complete the survey. Of the people who complete the survey, 10% of respondents are removed due to invalid or low quality answers. If we want 1000 valid responses for this particular study, how many people do we need to have start the survey?