How to Ask Viewers What They Want Without Bad Data

Ask viewers about the last time they faced the problem, not whether they would watch a future video. Past behavior is a record. Future intent is a guess, and people guess generously when asked nicely.

That one change fixes most of the bad data creators collect. A question about a past event forces memory, and memory carries the detail you need to make something real.

Ask about the last time

Instead of 'would you watch a video about X', ask 'tell me about the last time you dealt with X'. The person reconstructs an event, and the event hands you the specific friction they hit. That friction is your hook.

Follow up with 'and then what happened'. People rush past the useful part because they assume it is boring. Your job is to slow them down enough to describe it.

• What did you try before you found something that worked?
• What made the last guide you used fall short?
• Where did you get stuck when no one could help?
• What would have saved you time on that specific day?

Do not ask what people would watch. Ask what they did last Tuesday when the problem showed up.

Avoid polls that measure nothing

A poll that asks 'which video should I make next' mostly measures which option has the funniest wording. Everyone who answers is self selected, and the loudest options win regardless of demand. Use polls to pick between ideas you already know matter, not to discover the ideas themselves.

When you do run a poll, keep answer choices in the audience's own words and leave one open field. That field is where the surprising intent shows up, the thing you did not think to offer.

Key takeaways

• Ask about past behavior instead of future intent.
• Use one follow-up to reach the useful detail.
• Treat polls as a tiebreaker, not a discovery tool.
• Keep one open answer field for what you missed.