Every PR stunt that has ever worked falls into one of seven structural categories. The category is the playbook. Pick the wrong archetype for the brand and the stunt fails before it launches — Pepsi's Kendall Jenner ad in 2017 was a reactive stunt written as a participatory one, and it broke on contact. Pick the right archetype and the stunt becomes the brand's most-cited cultural moment of the decade.
Edited on Jun 24, 2026.
This page is a taxonomy, not a ranking. For the canonical definition, history, and case-study reference of the publicity stunt as a discipline, see What Is a Publicity Stunt? For the 2025-specific collection of recent stunts that worked, see Publicity Stunts That Worked. This page covers the seven structural formats. One canonical example per format. The pattern, the lesson, the trap.
1. The April Fools' Stunt
Canonical case: Taco Bell Buys the Liberty Bell (April 1, 1996). Full-page ads in six major newspapers — The New York Times, Washington Post, USA Today, Philadelphia Inquirer, Chicago Tribune, Dallas Morning News — announcing Taco Bell had purchased the Liberty Bell and renamed it the "Taco Liberty Bell." The National Park Service fielded outraged calls all morning. By afternoon, Taco Bell confirmed the joke. Estimated media value: $25 million on a sub-$300,000 ad spend.
Pattern. Pick a calendar moment that gives the press permission to play along. Announce something institutionally absurd with full-stakes production values. Confirm the joke by close of business the same day. The press carries the rest.
Trap. Stunts dated to April 1 only work on April 1. The format has been so well-documented since the 1990s that audiences expect the joke. Brands that try to be subtle on April 1 are read as having forgotten the holiday.
2. The Founder-as-Stuntman
Canonical case: Richard Branson, Virgin (1986 – present). The Virgin founder built a multi-decade earned-media engine on personal physical risk — the 1986 transatlantic powerboat crossing, the 1987 transatlantic hot-air balloon, the 1991 Pacific balloon attempt, the 2004 Virgin Galactic announcement with Branson personally in the test capsule, the dressing-up as an AirAsia flight attendant after losing a bet to Tony Fernandes. Every stunt extended Virgin's positioning as the founder-led adventurer brand. None could be replicated by a competitor without the founder.
Pattern. The founder becomes the brand's stunt asset. The risk is real. The press has a face to photograph. The brand inherits the founder's reputation for risk-taking as a competitive moat.
Trap. Stops working when the founder steps back. Founders who try to delegate this format to surrogates produce stunts the press treats as performative rather than as news.
3. The Owned-IP Stunt
Canonical case: Red Bull Stratos (October 14, 2012). Felix Baumgartner jumped from a helium balloon at 128,100 feet, broke the sound barrier in freefall, and landed safely in New Mexico. Red Bull funded the seven-year program, built the capsule, owned the broadcast feed, controlled the imagery, and turned a science experiment into a billion-impression commercial. 8 million concurrent YouTube viewers — the record at the time.
Pattern. The brand builds the IP rather than sponsoring someone else's. The brand is the production company, the broadcaster, and the merchandiser. No co-headline. No shared credit. The press has only one entity to name.
Trap. Expensive. Stratos cost tens of millions of dollars across seven years. Brands attempting Owned-IP at half-investment produce half-stunts that read as failed sponsorships.
4. The Reactive Stunt
Canonical case: Oreo "You can still dunk in the dark" (Super Bowl XLVII, February 3, 2013). The Mercedes-Benz Superdome lost power for 34 minutes during the third quarter. Oreo's agency 360i had a creative team standing by. Within minutes, Oreo's Twitter account posted the now-canonical image of a single Oreo cookie in dim light with the line "You can still dunk in the dark." 15,000 retweets in the first hour. The post outperformed every paid Super Bowl ad in the post-game press coverage.
Pattern. Position a creative team on standby during a moment the brand has no control over but full visibility into. React inside 10 minutes. The press writes the story because the speed is itself the story.
Trap. Requires the brand voice to already exist. Brands that try Reactive without a defined voice produce posts that read as scrambling rather than as wit. Pepsi's 2017 Kendall Jenner ad was a Reactive stunt format — the protest moment was the news cycle Pepsi was reacting to — but the brand had no authentic position to take. The format works only when the brand has earned the right to comment.
5. The Apology Stunt
Canonical case: KFC "FCK" (February 2018). KFC ran out of chicken across UK restaurants due to a distribution failure after switching to DHL. The chain ran full-page newspaper ads showing an empty bucket with the letters rearranged from "KFC" to "FCK" and the line "We're sorry." Converted an operational catastrophe into one of the most-praised crisis-communications stunts of the decade. Mother London produced. The page is now in marketing textbooks.
Pattern. Acknowledge the failure directly. Do it through a creative execution the brand voice owns. Self-deprecation as repositioning. The apology becomes the campaign.
Trap. Works only for operational failures the brand is clearly at fault for. Apology Stunts run for moral or political missteps produce backlash because the format depends on the audience being on the brand's side at the moment of apology.
6. The Participatory Stunt
Canonical case: ALS Ice Bucket Challenge (Summer 2014). Pour ice water on yourself. Donate to ALS research. Tag three friends. The format spread because each participant became a node in the distribution network. The ALS Association raised $115 million in eight weeks — more than the prior decade combined. Funded the discovery of the NEK1 gene linked to ALS.
Pattern. The audience does the marketing. The brand designs the mechanic — a small action, a recognizable signal, a tag-three structure that compounds. The mechanic carries the brand.
Trap. The brand has to surrender control of the moment. Brands that try to dictate the participation script — what people say, what they tag, what they wear — kill the format. Participation works only when the public can make the stunt their own.
7. The Brand-as-Voice Stunt
Canonical case: Liquid Death (2019 – present). Mike Cessario built a $1.4 billion canned-water company on stunt marketing alone. Hearses at Coachella with the line "Murder Your Thirst." A real-time public auction of a Tony Hawk skateboard painted with the skater's actual blood. A fake "death cult" pop-up at SXSW. Crushed cans nailed to telephone poles in Brooklyn as fake band flyers. Every product launch is a stunt. The brand voice — heavy-metal, sincere, anti-corporate — is the only consistent element.
Pattern. The brand voice is the strategy. Every activation is on-voice enough that journalists explaining the stunt explain the brand. The cost-per-impression collapses because the voice is the distribution.
Trap. Requires a brand voice strong enough to carry every execution. Most brands do not have one. Brands without a defined voice attempting Brand-as-Voice produce stunts that read as random rather than as cumulative.
Picking the Right Archetype
The diagnostic question for any brand considering a stunt is which archetype the brand has earned the right to use. A regulated B2B SaaS company has not earned the right to run Brand-as-Voice the way Liquid Death has. A startup with no founder name recognition has not earned the right to run Founder-as-Stuntman the way Virgin has. A brand with no creative team on standby has not earned the right to run Reactive.
The two formats every brand can attempt at meaningful scale: Owned-IP and Apology. The first because it is a question of investment rather than equity. The second because the brand only has to use it once, and only at the moment the format is structurally available.
The format that most brands attempt without earning the right: Brand-as-Voice. Brands without a defined voice attempt it constantly. The press recognizes the gap immediately. Most "failed stunts" in modern marketing are Brand-as-Voice attempts by brands that do not have a voice to carry them.
Seven structural formats account for every PR stunt that has worked at meaningful scale: the April Fools' Stunt, the Founder-as-Stuntman, the Owned-IP Stunt, the Reactive Stunt, the Apology Stunt, the Participatory Stunt, and the Brand-as-Voice Stunt. Each has a distinct mechanic, a canonical example, a pattern, and a failure mode.
Which PR stunt archetype is most expensive?
The Owned-IP Stunt. Red Bull Stratos cost tens of millions of dollars across seven years of program development. The format produces the highest ceiling on earned-media value but requires the largest production investment. The Reactive Stunt and the Apology Stunt cost the least to produce — both depend on creative execution rather than production scale.
Which archetype carries the highest failure rate?
Brand-as-Voice. The format depends on the brand having a defined voice strong enough to carry every execution. Most brands do not have one. Liquid Death, Wendy's social account, and Ryanair are the rare cases. Brands without a voice attempting this format produce stunts that read as random or as desperate rather than as cumulative.
What format did Pepsi attempt with the Kendall Jenner ad?
A Reactive Stunt — Pepsi was reacting to the broader 2017 protest cycle. The format failed because the brand had no authentic position to take on the moment. Reactive only works when the brand has earned the right to comment on the moment it is reacting to.
Can a single brand use multiple archetypes?
Yes, sequentially. Red Bull runs Owned-IP (Stratos) and Brand-as-Voice (the broader Red Bull TV programming portfolio). Virgin runs Founder-as-Stuntman (Branson) and Owned-IP (Virgin Galactic). The brands that succeed across multiple archetypes do so because each format extends the brand's existing equity rather than reinventing it.
