In addition to the current hype about Gen Z, digital agencies love to sell another seductive idea: that elections can be won by slicing the electorate into ever-smaller segments and whispering perfectly tailored messages into each voter’s ear. It sounds modern and scientific yet is highly overrated.

Let me explain to you the three reasons why I think so.

The first problem is technological reality. The data behind microtargeting is far messier than its evangelists admit. About half of the clicks on the internet are accidental. Traffic is polluted by bots and click farms. Ad supply chains are stacked with middlemen. Yet campaigns routinely treat digital metrics as if they were precise instruments. I won’t rule out that AI will fundamentally improve the situation, but at the moment there are still plenty of reasons not to blindly trust social media statistics.

Second is the personalization myth. We assume that the more customized a political message is, the more persuasive it becomes. Recent evidence suggests, however, that that might not necessarily be true. Research in the Harvard Journal of Press and Politics shows that there is a tipping point where tailored messages stop feeling relevant and instead start feeling creepy. Voters do not feel understood. They feel watched.

But the biggest damage done by microtargeting is strategic. Time and attention are a zero-sum game. When the architecture of a campaign is built around algorithms and models, the core message loses its punch. The result is a foggy campaign that cannot clearly explain what it stands for.

Big, bold steps usually win elections. Micro-optimization, on the other hand, often masks strategic macro-level weakness.

I remember that during the last U.S. presidential campaign I read reports about how Democrats were sophistically targeting Polish American voters in the state of Pennsylvania. At the time, I took this as an early warning sign that Democrats were getting lost in the data and Trump did indeed ultimately carry the state with a margin of 1.7%.

As a matter of fact, most parties and politicians are a mass product. They therefore need to win something close to a majority of votes. Yes, there are exceptions. For example, a small party operating in a proportional system is more like a niche product. They need to appeal to a small part of the electorate. In such a situation (or in any extremely close race), microtargeting may indeed offer marginal advantages at the edges. But such occasions are rather rare.

This year alone, I have won more than 10 competitive election campaigns. In most cases, the candidate or the party was a mass product. They don’t win by sounding different to everyone. They win by sounding clear to everyone.

Think about it. How many mass products do you know that microtargeted their way to becoming a mass product? Just look at the world’s strongest brands. Apple. McDonald’s. Coca-Cola. They all became global giants through the power of mass communication and message discipline. They repeated simple ideas until they became cultural facts.

That being said, campaigns do of course need a targeting strategy. You must know who you are trying to reach. Equally important is to decide who you don’t want to reach. But broad categories are enough to do that. Beyond a certain point, obsessing over hyper-segmentation is a distraction. Data should inform strategy, not replace it.

Future electoral success won’t belong to those who analyze each and every individual tree, but to those who can see the forest.

PS: Check out my channel on YouTube where the latest video is also on this topic.