Are polls useful anymore?

In 2018 pundits were predicting a Democrat wave in the midterm elections. Didn't happen. In 2016 they predicted a Hillary Clinton landslide over Donald Trump for the presidency. Didn't happen. In the U.K. that same year they predicted that Britons would vote to stay in the European Union during the Brexit vote. Didn't happen.  Later in 2017, Theresa May, leader of the conservative party in the U.K. had called a snap election to capitalize on the Brexit vote and was expected to win an expanded majority in parliament. Didn't happen. And most recently, this week the Trump supporting Australian Prime Minister won a re-election no one thought he had a chance of winning. He was supposed to be trounced by the Labor Party.  Didn't happen.

Are polls useful anymore?  True there is an argument that polls can be used to direct a narrative  - voters can be asked leading questions, they can be oversampled from Democrat-leaning populations  and they can just be deliberately misinterpreted.  But there are other issues less sinister.  With the increase in mobile phone usage and a decline in landline phones, there are potential issues with how samples are drawn. Polling methodologies seem to be not working.

While pollsters need to rethink their methodologies, the broader lesson for voters, is to ignore the polls and just vote.  Voting after all, is what really matters.