Wow I find it extremely expensive. The OP's survey had 14 questions. To get statistically relevant results we would have to spend 2100$.
As I understand their advertisement, they ask the same user the full survey but over visiting several sites? As one question costs 10ct, I wonder if you can conclude the survey after "do you know bitcoin?" as knowing that 99% don't know bitcoin will make the rest 1% answers kind of statistically irrelevant. (I assume you can create subquestions. Everything else would be stupid.)
Edit:
Custom audience
Target a specific population like dog owners, moms, or golfers using a threshold question (e.g. "Do you own a dog?”).
$0.50 per response for follow-up questions to that audience or $750.00 for 1500 responses (recommended for statistical significance, minimum incidence rate of 5%)
Assuming less than 1/5 of the general population qualifies to answer these bitcoin questions, this would make the survey 5 times more expensive.