How many interviews do I need for an opinion poll?

How many interviews are generally accepted as a credible sample size for a poll? The answer is not quite as straight forward as the question seems.
For an opinion poll or PR survey to be seen as credible by journalists and readers, you need to have a sufficient number of interviews. Most polls published in newspapers will be 1,000 or 2,000 if they are looking to provide a sufficiently accurate view of the national population, split by region, age and gender. Which should you use - 1,000 or 2,000? Technically, a well structured sample of 1,000 Nationally representative interviews should be sufficient. But from a practical point of view, the fact that some publications and journalists favour 2,000 interviews is inevitably right at the top of the list of considerations. However good the survey is, however creative the idea, if it is not seen to have a robust sample, the poll will not gain traction among target readerships. There are, however, general exceptions to this. It is possible to run smaller sample sizes with certain target groups e.g. parents with children of a particular age - perhaps 500 interviews or fewer. Similarly several hundred interviews with business respondents may be seen as sufficient. But even here, ultimately we need to be sensitive to the needs of the publications and channels we are seeking coverage.