Heineken has strangers with opposing political views build furniture together before either learns what the other believes

the problem

Ads about tolerance preach to people who already agree

background

what everyone would do

The standard format for a tolerance message is to put opposing views in the same room and let them debate, or simply present the message to viewers directly. Both fail for the same reason: leading with the point of disagreement activates identity defense before any human connection has a chance to form, so the people most resistant to the message are exactly the ones who tune out or dig in the moment the disagreement surfaces.

what they saw

Publicis London saw that the sequence, not the content, was the actual obstacle — the same two strangers who would defend their position instinctively if told the disagreement first might behave completely differently if they'd already formed a working bond before the disagreement was revealed. The insight was that a cooperative task completed together changes what a subsequent disagreement means to the people involved: it's no longer a stranger's alien opinion, it's a disagreement with someone they've already started to trust.

the move

Publicis London recruited three pairs of strangers under the pretext of a social-experiment shoot, gave each pair a physical task — building a bar and other furniture together — before either person knew the other's stance, then showed each participant a pre-recorded clip revealing their partner's opposing view on feminism, climate change, or transgender identity, and let each pair choose, buzzer-style, to walk away or stay for a beer and talk it through.

why it works

Having pairs build furniture together before either party knows the other's political stance lets a real cooperative bond form under conditions with nothing to defend — there's no belief on the table yet, just a shared task. Once that bond exists, revealing the disagreement changes what the information means: it now has to compete against an already-formed positive impression of a specific, likeable person, rather than landing on a blank slate primed for identity defense. This is exactly the mechanism Stanford's Strengthening Democracy Challenge megastudy (32,000 participants, published in Science) found when it tested the Worlds Apart video against 24 other interventions and found it the single most effective at reducing partisan animosity — the sequencing itself, cooperation before disclosure, outperformed interventions built on facts, shared identity messaging, or direct persuasion.

the payoff

All three filmed pairs chose to stay and talk. The film reached over 40 million views across social channels, and Publicis London/Heineken won Bronze Lions in both the Cyber and PR categories at Cannes Lions 2017. Five years later, Stanford's Strengthening Democracy Challenge megastudy (roughly 32,000 participants) independently tested dozens of interventions for reducing US partisan animosity and found the Worlds Apart video was the single top-scoring intervention — real academic evidence the sequencing mechanism, not just the ad's warmth, did work. Contemporaneous coverage flagged the obvious caveat: a produced ad with three curated pairs is not proof the effect generalizes uncurated, and whether any lasting social change followed for Heineken itself was never demonstrated.

where it breaks

The mechanism depends on the shared task being genuine enough to produce a real bond before disclosure — a task too trivial, too short, or too obviously staged to build any real rapport leaves nothing for the subsequent disagreement to compete against. It also depends on the disagreement being disclosed by the person themselves rather than imposed by a third party or discovered accidentally, since the reveal has to preserve the human relationship already built rather than reframe it as a manipulation the participant didn't consent to — contemporaneous criticism of the ad itself flagged exactly this risk, noting some pairings (like leaving a trans woman alone with someone expressing transphobic views) put unequal burden on one participant. And a three-pair produced film is not proof the effect generalizes uncurated at scale; the academic megastudy is what actually validated the mechanism causally, and even that showed the effect works for partisan animosity specifically — a different, smaller set of interventions moved anti-democratic attitudes themselves, so the same technique should not be assumed to transfer automatically to every kind of intergroup conflict.

what came after

The 'shared task before disclosure' structure became a reference point cited well beyond advertising, including in academic research on reducing partisan hostility, as a working example of sequencing cooperation ahead of conflict rather than leading with it.

references

1. [1]How to get Democrats, Republicans to strengthen democracyCornell Chronicle, reporting on the Strengthening Democracy Challenge megastudy published in Science, 2024news.cornell.edu
2. [2]Is This 2017 Heineken Ad the Key to Saving Democracy?InsideHook, 2022insidehook.com

keep it

same kind of clever

• 1911 · Aalsmeer Flower Auction (Royal FloraHolland)Aalsmeer sells its dying flowers by racing a price downward instead of bidding one upward
• 1911 · Roald Amundsen's South Pole expeditionAmundsen marked his supply depots sideways to the route, not along it, so a lost party could still walk into one
• 1841 · P.T. Barnum (American Museum)Barnum cleared his overcrowded museum by making the exit look like an exhibit