Good data, good science: simple advice for better CCD photometry