Yeah, the polling industry has an existential crisis on their hands. If we don't trust polls at all, we'll have to go about predictions in a completely different way.
Polling wasn't wrong(it was more off in 2012) it's that the data was interpreted wrong, especially by pundits and the media, because they couldn't fathom a Trump win. State poll did mess up a bit though.
From what I could tell, national polling was pretty on par (
Clinton +2 or so), but individual state polls were wrong, which led to the wrong predicted EV map, such as:
https://projects.fivethirtyeight.com/2016-election-forecast/. As you can see, polling in Florida (
Clinton +1), North Carolina (
Clinton +1), Pennsylvania (
Clinton +3), and more states Trump won were off. And FiveThirtyEight was one of the more favorable aggregators toward Trump.