Adweek 复盘 2026 年美国政治广告，120 亿美元投放未能扭转选情

2026 political advertising was a disgrace. 

Campaigns torched billions on placements that had little-to-no impact on voter behaviors. Many ads backfired, and in a few cases — such as Texas and Ohio — did so spectacularly. 

AdImpact says 2026 was the most expensive election year in history (including both presidential and midterms cycles), with $12 billion spent on ads. Most of that money was backed Republicans, but the GOP wipeout makes you wonder if it had any effect. Could things have gone worse for the party if campaigns and PACs had run no ads at all? Doubtful. 

The most egregious examples are truly embarrassing for the losing campaigns. In Texas, Elon Musk’s over-the-top spending to back Trump loyalist Ken Paxton created an opening for James Talarico to surge on a “Texas is not for sale” message. 

In Ohio, Vivek Ramaswamy’s desperate and gross ad featuring Amy Acton’s mother may go down as the worst miscalculation in political advertising history. 

It wasn’t just Republicans who hurt their chances with terrible ad strategies. In Maine, Janet Mills had to bow out of the Dem primary after negative ads targeting Graham Platner backfired and her support dried up. 

Beyond that, a whole bunch of deceptive and undisclosed AI-generated ads triggered negative media coverage and may have run afoul of the Federal Elections Committee. 

What lessons can we draw? 

It’s not that political ads never work, but they can’t accomplish what many campaigns seem to want them to do. 

Specifically, ad spending will never overcome a broad-based shift in voter sentiment of the kind we just saw. Republicans outspent Democrats by $200 million in Texas, by $100 million in Ohio, by $84 million in Michigan. If such princely sums were spent to promote snacks or mobile games rather than political brands, the results would be real and highly measurable. Where’s the return on all this campaign spending? Nowhere to be found. Turns out, no amount of money could sway angry Texas voters to back Ken Paxton. When people have turned their back on a whole product category — whether MAGA-style politics or sugary drinks — the individual products and brands in that category stand no chance. 

Also, 2026 campaigns vastly overrelied on a last minute onslaught of ads to change minds. Relentless October spending just dials up the saturation, inviting voter annoyance or outright hostility. The very best use of political advertising may be to build awareness for a candidate early in the cycle. Once that opportunity has been squandered, all that remains is to browbeat voters who already know (or think they know) your candidate. This is probably more true in a midterm race, when voters are on the whole more informed than in general elections. 

Finally, please stay away from AI-generated creative — especially if it’s meant to deceive or confuse. The use of AI in political ads dramatically ramped up over the last two years, from $50,000 to at least $137 million. According to The Wesleyan Media Project, such ads are heavily skewed toward Republican campaigns. Fully 81% of AI-created ads supported GOP candidates and issues, compared with just 12% of Democrats. This makes some sense, since voters on the left generally have been louder in their disapproval of AI. But the public ire toward AI is increasingly bipartisan. The most telling thing about AI-generated ads is how many of them exclude disclosures — suggesting campaigns either know they’re unpopular, are deliberately misleading voters, or both.  

No lie: The 2026 political ad outlay may have been the dumbest money spent in the history of advertising. Even so, media and tech platforms will be glad to take political money from the same undiscriminating sources again in two years’ time. It’s kind of a travesty. 

The good news is, our relentless politics means campaigns will have the chance to mend their ways all too soon. I’d urge you to think like a smart commercial marketer: Find brands that are worth promoting, build the story early and learn from the mistakes of the past.