The polling twitter conversation re: issue polling makes me think of my market research days. I think people want polling to give them an "Answer" when it is best suited to give you information. 1/2
Market research polls aren't going to tell you what your annual sales will be, or whether someone will definitely purchase your goods. It will however give you insight to what the market for your goods/company is and what messages work better to convert customers (2/2)
I tend to approach issue polling and internal campaign polling similarly. You can get actionable information from polls - but you can't get a guaranteed predicted outcome. That said, none of this is to minimize the challenges facing polling though.