Questionnaire and survey design

You are a survey design expert. Help me build the questionnaire for my research: objective [WHAT I NEED TO DISCOVER], respondent audience [WHO], application [ONLINE/IN-PERSON/TELEPHONE], and data use [INTENDED ANALYSIS]. Deliver: the instrument architecture built backwards (each question stems from an analysis need — the table question → variable → analysis that eliminates 'it would be cool to know' questions that only lengthen it), question wording without bias with your examples revised (the plagues: double question 'is the product good and cheap?', leading 'do you agree that...', ambiguity, jargon, impossible memory 'how many times in the last year...' — before/after each one), the right scales for each case (Likert how many points and why, with neutral point or not, NPS when appropriate, order and labels of options — and the balance that avoids pulling the answer), the strategic order of the questionnaire (easy warm-up, thematic blocks, sensitive and demographic at the end, filters and logical jumps), honest length (abandonment rate rises after [X] minutes — cutting what is not essential), mandatory pre-testing (5-10 people from your audience thinking aloud — what always reveals itself), opening texts (consent, anonymity, estimated time) and anti-bias collection design (avoid only engaged volunteers). Goal: an instrument that measures what you need — and that people answer through to the end.