In my work as a global political consultant, I use market research (both quantitative and qualitative) every day as a basis to plan successful election campaigns. But as a tool to predict the outcome of the election, polls (at this point in time) are next to irrelevant. In fact, I find it almost amusing how pundits on television zealously discuss changes in the polls of a couple of percentage points.

First, the polling market in the U.S. is seriously skewed. Polls have been substantially off for several election cycles in a row. Ever since the surprise election of Donald Trump in 2016, U.S. pollsters have claimed that they have fixed the issue(s), but it remains to be seen. And there are indeed several things that would need to be addressed.

Keep in mind that a U.S. presidential election is not a nationwide, but de facto a state by state election. It’s not only a matter of how many votes a candidate gets, but where he gets them. Last time, Joe Biden lead substantially in the nationwide polls going into election day, but when it was all said and done, less than 50’00 people in three states decided the election (control of the House of Representative was decided by an even smaller number of voters in 2022). In other words, if Kamala Harris is ahead in nationwide polls by a few percentage points, she’s probably losing the election.

On a more technical side, it has also become common in the industry to look at the average of polls. But this is based on the assumption that the sample size would be the main factor influencing the quality of a poll, and that pollsters have all the other factors equally under control. This might not necessarily be the case.

Linked to this is the issue of herding among pundits and pollsters. When everybody on TV says that Hillary will win (2016), that there will be a huge repudiation of Trump (2018), that it’s not a close election (2020), that Democrats have no chance to hold on to the House (2022), it takes courage to go against the mainstream or even to raise an issue.

After the issues regarding polling methodology comes the interpretation of the results. Let’s assume for the sake of argument that the polls are accurate at the current moment. But with two months to go until the election, it really doesn’t matter if a candidate is two points ahead or behind. There is a margin of error to consider and many things will happen until the election. We don’t know how people will see the economy by then, how the wars in Ukraine and Gaza will develop and, most importantly, who will show up and vote.

If anything, the most noteworthy thing in the polls right now is not the vote question, but the favorability ratings of the candidates. Harris was able to significantly improve her favorability ratings since becoming a candidate and if she can maintain this until election day, that would put her in a good position.

Politicians often have a love/hate relationship with polls. When they like the results, they forget all the limitations of polling. When they dislike the results, they dismiss them altogether. The truth is often somewhere in between, and at this point in time, the polls basically indicate that the U.S. presidential election is a close race.

PS: An earlier version of this article was published in the U.S. magazine Newsweek. You can read it here.