Does a joke in a cold email actually help, or just feel like it does?
A fair look at what the actual research on humor and persuasion says, where a joke helps a cold email and where the same joke quietly costs you the reply.
By Norbelys Chirinos, Co-founder
Founder-reviewed ·How we research and correct articles
Ask ten people whether a joke belongs in a cold email and you’ll get ten confident, contradictory answers. “It shows personality, it stands out.” “It’s unprofessional, it undermines you.” Both camps are working from vibes, not evidence. The actual research on humor and persuasion is more interesting than either position — and it explains exactly why the same joke can work beautifully for one sender and quietly sink another.
The honest finding: humor helps and hurts you at the same time
The clearest evidence comes from advertising research, not email specifically. Martin Eisend’s 2009 meta-analysis combined 369 correlations from the published experimental literature on humor in ads and found a genuinely two-sided effect: humor significantly increases attention and improves attitude toward the message itself — but it also significantly reduces perceived credibility of the source. Those two effects aren’t a contradiction in the data; they’re the same mechanism seen from two angles. A joke earns attention because it breaks the expected pattern of a sales message. That same pattern-break is what makes a skeptical reader trust the sender slightly less — a funny message reads as less serious, and “serious” is exactly what a B2B buyer is unconsciously scanning for when deciding whether to take an unsolicited email at face value.
That trade-off is the honest answer to “does humor work”: it depends entirely on which side of the trade-off matters more for a specific email. An email trying to get opened and remembered in a crowded inbox benefits from the attention half of the effect. An email trying to establish that the sender is a credible vendor for a five-figure contract is fighting against the credibility half.
Humor is a bet, and the research says it can go either way
A separate line of research gets at why humor feels risky even when it works. Bitterly, Brooks, and Schweitzer’s 2016 study in the Journal of Personality and Social Psychology ran eight experiments manipulating whether a person’s attempt at humor landed or fell flat, then measured how it changed perceptions of that person’s status, confidence, and competence. Successful humor raised all three. Unsuccessful humor — the same attempt, just perceived as not funny or inappropriate for the context — lowered them. The finding isn’t “humor is good” or “humor is bad.” It’s that a humor attempt is a genuine bet with two outcomes, and the downside of losing the bet is measurably worse than the downside of never having tried.
That framing maps directly onto a cold email, where the sender gets exactly one attempt, no read on the room, and no ability to recover if the joke misses. A comedian bombing on stage gets to read the silence and adjust the next line. A cold email that misses the joke just sits in an inbox, unread past the first sentence, with the sender’s credibility already discounted and no chance to walk it back.
| Humor that lands | Humor that misses | |
|---|---|---|
| Perceived status | Higher — reads as confidence | Lower — reads as poor judgment |
| Attention earned | ✓ | ✓ |
| Source credibility | Slightly reduced regardless | Reduced further |
| Recoverable in the same email? | — | — |
Where the bet actually pays off
The research points to a few conditions where humor’s upside is more likely to beat its downside in a cold email specifically:
Self-deprecating, not other-directed. A joke at your own expense — about the outreach itself, about how many cold emails the recipient probably gets — carries almost none of the credibility risk, because it doesn’t require the reader to agree the joke is objectively funny; it just has to read as self-aware. “I know this is the eleventh cold email in your inbox this week, I’m not going to pretend otherwise” isn’t really a joke so much as an honest observation with a light touch, and it costs you nothing if it doesn’t land, because there’s no punchline to fail.
Situational, not scripted. Humor tied to something real and specific about the moment — a genuinely odd coincidence, an honest aside about the outreach process — reads as spontaneous. A pre-written joke that could be sent to anyone reads as a bit, and a bit that misses is worse than no bit at all, because it signals the email is templated rather than personal.
Light, not structural. A single wry line inside an otherwise straightforward, respectful email risks little. Building the entire pitch around a comedic premise raises the stakes on the joke landing to “does the whole email work,” which is a much worse bet.
Where it reliably backfires
The Eisend meta-analysis and the Bitterly/Brooks/Schweitzer status research point at the same failure modes from different directions: humor about the recipient or their company (rather than about the sender or the situation) risks reading as mockery rather than levity — the exact “unsuccessful attempt” condition that cost status in the status study. Humor stacked on top of an already weak pitch compounds the credibility hit the meta-analysis found, because a reader already unconvinced by the offer has one more reason — “this feels unserious” — to stop reading. And humor in a first-touch email to a senior or risk-averse buyer (finance, legal, compliance roles) is playing the bet against an audience that’s disproportionately scanning for credibility signals in the first place — precisely the dimension the research says a joke puts at risk.
The audience variable neither study measured directly
Neither the meta-analysis nor the status study set out to measure this, but it’s worth naming as a real risk factor rather than leaving it implicit: humor is one of the least portable things in written language. What reads as a light, confident aside in one region or industry can read as flippant, confusing, or even rude in another — sarcasm in particular translates badly, because the reader has to correctly infer that the literal words don’t mean what they say, and a stranger reading a cold email from an unfamiliar sender has the least possible context to make that inference correctly. A B2B seller running outbound across multiple countries or industries is effectively running the Bitterly, Brooks, and Schweitzer bet with worse odds on every send outside their own cultural and professional context, because “did this land” depends on shared reference points the sender can’t verify are actually shared. That’s not a reason to avoid humor in international outreach entirely — it’s a reason to weight it toward the safest category described above (self-deprecating, about the outreach itself) rather than anything that depends on the reader picking up on tone or irony correctly on a first, cold read.
The test that actually settles it for your list
Both studies agree on one thing that matters more than either finding individually: the effect of humor depends heavily on the audience and the specific joke, which means no blog post — this one included — can tell you definitively whether a joke will help your next campaign. What the research gives you is a map of the trade-off, not a verdict. The only way to actually resolve it for a given segment is to test a version with a light, self-deprecating aside against a straightforwardly serious version of the same email, and judge by reply rate — not opens, which a joke can inflate simply by being more forwardable or screenshot-worthy without actually earning more genuine replies.
Humor in cold email, quickly
Is there a safer type of humor for a first cold email to a stranger?
Self-deprecating asides about the outreach itself tend to carry the least risk, because the research on humor and status suggests the downside of a joke mostly comes from the reader deciding the sender misjudged the moment — and a joke that pokes fun at your own position rather than the reader's is much less likely to be read that way, even if it doesn't land as intended.
Does humor work better later in a sequence than in the first email?
Generally yes, and it follows directly from the status research: once some rapport exists from an earlier reply, a joke has more context to be read charitably, and the credibility cost of an unfamiliar sender being 'not serious enough' matters less than it does on a completely cold first touch.
Should B2B sellers avoid humor entirely to be safe?
The research doesn't support that either — Eisend's meta-analysis found real upside in attention and message attitude, not just risk. The honest read is that humor is a genuine trade-off, not a free win or a hard no, and the sizing of that trade-off depends on the audience, the seniority of the buyer, and how central the joke is to the email rather than a single incidental line.
Does the format matter — is a funny GIF or meme different from a written joke?
The status research was about attempted humor generally, not written text specifically, but the same logic applies with an added risk: a visual joke is even less portable across audiences and even less recoverable if it misreads the room, and it adds a rendering and deliverability dependency a plain-text line doesn't have. Written, situational humor is the lower-risk version of the same bet.
Test the trade-off instead of guessing at it
If you’re going to bet on a joke, bet on being able to measure whether it actually paid off. Norbelys lets you A/B test a lighter, self-deprecating version of a step against a straightforwardly serious one on the same segment, and pins the winner by real reply rate using only human-verified opens and replies — not raw pixel fires that a forwardable, “look at this email” joke can inflate without actually earning more genuine responses. Norbe, the built-in AI operator, can also draft both a serious and a lighter variant of a step from the same brief, so testing the trade-off doesn’t cost you an extra hour of writing. See how the plans compare or start testing your own tone on Norbelys.