Humor works best on products people buy for fun (snacks, fashion, beer, perfume) and barely moves the needle for routine commodities (paper towels, batteries).
For expensive functional products (insurance, electronics, SaaS), humor only works if the joke is about what the product does — not just adjacent to it.
A meta-analysis of 43 humor-in-advertising studies sorted the rules out cleanly.
THE MECHANISM
Hedonic products run on feeling. The job of the ad is to make the buyer feel something, and that feeling attaches to the brand by simple association. A funny ad makes the buyer feel good near the logo, and the good feeling is the value proposition. Snack, fashion, beer — the joke doesn't have to be about the product, because the product is in the feeling business.
Functional products run on argument. The buyer is asking "does this work?" and a joke doesn't answer that. The good feeling evaporates by the time they're weighing features. The only humor that survives the trip from ad to brand attitude is humor that is the argument — a punchline that doubles as a benefit claim.
For low-involvement commodities, neither tactic moves the needle. The buyer isn't engaging with the ad in the first place.
TEST IDEAS
Match the joke to the product. Hedonic brand (snacks, fashion, beer)? Any joke that lands. The feeling lifts the brand on its own.
Functional brand (mattress, SaaS, insurance)? Make the punchline be the benefit claim — not a gag wrapped around one.
Commodity (paper towels, batteries)? Skip the humor. Spend the effort on key features.
THE STUDY
Martin Eisend (2009). A meta-analysis of 369 correlations from 43 humor-in-advertising studies. Humor lifts ad attitude strongly (r = .37) and brand attitude about half as much (r = .19) — and the brand-attitude lift varies by product. Hedonic products: humor works, related or unrelated. High-involvement functional: only related humor (β = .12). Low-involvement commodities: nothing.
(caveats: most underlying studies were US-based and used student samples — the author notes this biases effects upward, but the directional pattern replicates across the dataset.)
Source: Eisend, M. (2009). "A meta-analysis of humor in advertising." Journal of the Academy of Marketing Science, 37(2), 191–203.