Stella Artois' 'The Artois Probability' computes odds on beer in old paintings — Cannes GP

Stella Artois dates back to 1366, when the Den Hoorn brewery opened in Leuven, and the brand wanted that heritage to feel tangible rather than abstract. GUT Buenos Aires partnered with the Museo Nacional de Bellas Artes to make the idea physical: if the beer in old paintings could plausibly be Stella, the brand's history was painted directly into European art history.

Turn 600 years of beer brand history into data: an algorithm analyses the beer in classic paintings and prints the probability it is Stella Artois.

The agency built an algorithm that analysed paintings by Old Masters — including works by Van Gogh and Monet — scoring each one on the year it was painted, the geographical location, the type of glass, and the colour of the liquid, then cross-referencing those variables against the brand's historical records to produce a percentage probability. The result was a series of print ads and outdoor billboards that labelled famous paintings with their 'Artois Probability', turning art history into playful, data-driven speculation.

The campaign won the Creative Data Grand Prix at Cannes Lions 2023 and a Gold Clio in 2024, and entered the D&AD awards archive. Its lasting effect was reframing how the brand's 600-year history is told: instead of claiming heritage, the work made viewers check every painted beer glass for a Stella Artois.

Stella Artois has one of the oldest brand stories in the world: the Den Hoorn brewery opened in Leuven in 1366, Sebastian Artois bought it in 1708, and 'Stella' joined the name in 1926. The problem was that a 600-year history is hard for a modern drinker to feel. GUT Buenos Aires found a way to make it visible — by looking at the beer painted in European art.

The agency developed an algorithm that analysed classic paintings and scored each one on variables such as the year it was painted, its geographical location, the type of glass depicted and the colour of the liquid, cross-referencing them against Stella Artois' historical records. Each painting received a probability score: the chance that the beer in it was actually a Stella Artois. The resulting print ads and outdoor billboards labelled masterpieces with their 'Artois Probability', inviting art and beer enthusiasts to see the brand hiding in art history.

The idea worked because it was honest speculation rather than a claim: the campaign never said the masters drank Stella, only that there was a probability — a frame that made the heritage argument feel playful instead of pompous. The Museo Nacional de Bellas Artes partnership gave the data cultural weight, and the copy treated the whole exercise as a fun what-if.

Juries agreed: The Artois Probability took the Creative Data Grand Prix at Cannes Lions 2023, won a Gold Clio in 2024, and now sits in the D&AD awards archive — proof that a heritage campaign can be data-driven, humorous and old-master-classic at the same time.

It converted an abstract claim (600 years of heritage) into something concrete and verifiable — a probability printed next to each painting.

The museum partnership gave the data cultural authority, so the campaign felt like art history, not advertising.

The playful what-if frame avoided the arrogance of claiming Van Gogh drank Stella, making the idea more credible and shareable.

The algorithm itself was the creative asset: the technology was the message, not just a tool behind it.

It gave viewers a new habit — checking the beer in any painted scene — which embedded the brand into cultural observation.

Heritage claims land better as play than as history lessons: a probability game made 600 years of brand history concrete, shareable and something audiences could verify themselves.

After the Creative Data Grand Prix at Cannes 2023, The Artois Probability collected further honours including a Gold Clio in 2024 and a place in the D&AD awards archive. Ads of the World and other creative archives continue to host the case study, and the campaign is widely cited as an example of using data creatively to dramatise brand heritage rather than to personalise media.