Ask the Audience Lifeline

Well, in my opinion, the stack of the results from the ATA lifeline depends on the difficulty of the questions, which is usually different in the show's production and the people who are in the audience (referring to your guess that many audience members might not know the answer). I believe that some audience members usually either know it or guess the answer, but at the end, the contestant has to decide whether to go for the audience or reject their most voted opinion. So, I think three elements are important to make the results of Ask the Audience helpful:

1. The contestant's decision

2. The decision of the audience members

3. The difficulty of the specific questions 

It may be possible, but I think I'd rather consider my statement remaining a theory. 

The way I see it, the harder the question, the less likely the audience will actually have a definitive answer. It's a bit rare to see an audience that actually knows the answer to harder questions because, well, everybody's different. Don't expect much to come out of the audience because there is the possibility that more people that could be just guessing the answer.

I think some of it has to do with how ridiculously difficult the questions tend to get for the higher rounds in the syndicated broadcast. Maybe it has to do with the fact that they're a lot more US-centric than questions in the primetime broadcasts, but whenever I watch an episode of primetime Millionaire, I tend to be able to make it to the second safe haven without too many problems, while syndicated episodes have me stunned as early as question 7 or 8, and it does not get easier from there.

I'm not sure if they're doing it to preserve costs. What it has led to, though, is no grand prize winner in almost 10 years - and the last one was pretty forced, too.

As a former contestant, I can say that Ask the Audience easily is the least helpful lifeline. In my case, the audience yielded TWO ties (that's right; 41% picked two different answers and 9% picked the other two answers). I honestly didn't expect the audience to know what Liberty measles are (I certainly didn't), but Ask the Audience almost has become a throwaway lifeline. In fact, the producers told contestants at a pre-show taping to avoid using Ask the Audience except for two instances: questions about pop culture and questions about the geographic region in which the show is taped. This leads me to believe that Ask the Audience actually is more for viewers than contestants.

I was on the US version of the show in 2014. This was when the difficulty of the questions was randomized, so I got some tricky questions early on in the game. There also was a Jump the Question lifeline instead of the 50-50. I used that lifeline pretty earlier in the game and got lucky because the dollar amount of the question I skipped turned out to be just $100. My strategy was to use that lifeline early in the game if necessary while low-dollar values were still on the board. The random dollar amounts definitely added to the excitement and the anxiety of the game for me!
I thought of my producer's advice to use the Ask the Audience lifeline only for questions about pop culture and questions about the geographic region in which the show is taped recently. On one episode a contestant incorrectly guessed on a question related to a U.S. senator who had received a death threat as head of a gaming commission. After the contestant lost, Chris Harrison told him that the Las Vegas audience probably could have helped him come up with the right answer: Harry Reid of Nevada.

For a few seasons, the dollar value of each of the first 10 questions was not revealed until contestants answered each question. That's what made it so exciting and nerve-wracking for me as a contestant! I had a look of utter fear on my face as I skipped a question; then I was absolutely relieved to learn it was worth only $100. I later was elated when I discovered that a really easy question that I answered correctly was worth $25,000. The tricky part of this version of the show was that contestants only got to keep half of their winnings if they left before correctly answering 10 questions. Of course, this encouraged contestants to keep playing until they usually lost all of their winnings, except for their $1,000 parting gift. That meant that I left with $20,250 instead of $40,500. As much as I wanted to keep playing, I had watched too many contestants lose all of their winnings on a guess, so I decided to take the money and run! Absolutely no regrets here.

When I was a contestant, there was a running total of winnings for the first 10 questions. However, contestants only kept half of their winnings if they left before answering 10 questions correctly. The following isn't the episode that I appeared on, but it will give you an idea of the randomized version of the game:
www.youtube.com/watch?v=zSbIClBF1iI