In the spring of 1985, the world’s largest soft-drink company voluntarily destroyed its flagship product — relying on data that looked flawless.
In the late 1970s, little tables with paper cups began appearing in shopping malls across America. Pepsi promoters offered passersby two unlabeled drinks and asked them to choose which tasted better. The campaign was called the Pepsi Challenge, and its results were unpleasant for Coca-Cola: in blind tests, people reached for Pepsi more often. The formula worked elegantly — Pepsi was slightly sweeter, the first sip more pleasant, and in a short comparison that was enough to win. Coca-Cola was losing market share: by the early 1980s, the gap between the two giants had narrowed to a few percentage points, and the trend was not reversing. For a company that had spent nearly a century thinking of itself as synonymous with soda itself, this was not competitive pressure. It was existential anxiety.
Coca-Cola’s leadership could have said: a blind test is a trick. It does not reflect real purchases. And they would have been right. But instead, the company decided to beat Pepsi on its own turf. If consumers preferred a sweeter taste in short trials — then give them a sweeter taste. The logic seemed ironclad: measure what people like, make exactly that, sell it. Thus was born the secret Project Kansas — a program to develop a new formula that would beat both Pepsi and Coca-Cola itself.
The scale of the research is striking even now. Around 200,000 consumers in 25 cities tried versions of the new formula. The results were unequivocal: in blind tests, the new taste beat old Coca-Cola by 55% to 45%. When participants were told that one cup contained Coca-Cola and the other a potential replacement, preference for the new taste rose to 61%. The new formula beat Pepsi, too. The data looked like a verdict in favor of change: what sane manager would turn down a product that wins in every measurement?
Chairman and CEO Roberto Goizueta and president Donald Keough saw the numbers as a map to victory. Goizueta was a Cuban émigré, a chemist by training, a man who had risen through the company on the strength of rational thinking. Keough was a charismatic salesman raised in an Irish family in Iowa. Both were smart enough to understand the risk, and confident enough in the data to take it. The problem was not that the executives were stupid — the problem was the instrument. A blind tasting measured the sensation of a sip. It did not ask the question that actually decided everything: “What will you feel when the old Coca-Cola disappears from the shelves forever?”
One thing still puzzles me: some of the tests did include the wording about replacement. Participants were told that the new taste might take the old one’s place. And even then, most chose the new formula. But between “might replace it” in the context of a survey and the actual disappearance of a product lies an abyss. A person answering a hypothetical question and a person discovering that their usual drink no longer exists are two different people.
On April 23, 1985, Coca-Cola held a press conference and announced that the original formula, which had existed for 99 years, was being replaced. New Coke — sweeter, smoother, more modern. Goizueta spoke of courage and evolution. Pepsi rejoiced that day — and made no attempt to hide it. The company had called more than 200 reporters in advance, supplying them with prompts for pointed questions at Coca-Cola’s press conference. Pepsi declared a day off for its employees: as if to say, our main competitor has just admitted that our taste is better.
For the first few days, New Coke sales went well enough — curiosity did its work. But then the storm began, and it gathered force each week. The company’s hotline in Atlanta, which normally handled around 400 calls a day, began to choke. At the peak — up to 8,000 calls daily. People did not ask for explanations. They demanded. They shouted. They cried. One caller told the operator that Coca-Cola was the only thing that had remained unchanged since his father died, and now they had taken that away too.
Letters arrived by the sackful. Some were addressed to “Chief Dodo” — an insult built around the image of the extinct flightless bird, a synonym for stupidity. People wrote as if a piece of their home had been taken from them. It was not a metaphor: for many Americans in the South, Coca-Cola really was woven into everyday life — Sunday dinner, a barbecue, a trip to the baseball game. A habit built into a ritual becomes invisible until someone touches it. Touch it — and it turns out to be a pressure point.
Consumers organized themselves faster than the company could respond. Gay Mullins of Seattle founded Old Cola Drinkers of America, putting up 30,000 dollars of his own money and claiming 100,000 supporters. At the same time, the Society for the Preservation of the Real Thing and the Cola Clan appeared — the latter had around 3,000 members. People bought up the remaining old Coca-Cola from store shelves and kept bottles in their basements — some enthusiasts stockpiled as many as 900 bottles, spending up to 1,000 dollars to do it. Soda was turning into a currency of nostalgia, an artifact of a vanished world.
The convenient version is that New Coke tasted bad. That is not true. In blind tests, it won — and won convincingly. Another convenient version is that Coca-Cola staged the scandal on purpose, so it could bring back the classic formula in triumph. That is not true either — one look at Goizueta and Keough’s faces at the press conferences that summer is enough to understand: these men were not acting. They had made a mistake, and the mistake was sincere.
So what happened? A blind test is a microscope with one eyepiece. It sees taste perfectly. It does not see everything else. When someone takes a sip from a paper cup, they are judging a liquid. When someone takes a bottle of Coca-Cola from a store shelf, they are not buying a liquid — they are buying continuity. A habit stretching across years. A red label that sat on the table when they were ten years old. The test measured the tongue’s reaction. The revolt was memory’s reaction.
There is another nuance that often gets missed. A short tasting — one or two sips — is systematically biased in favor of the sweeter drink. The first sip of something sweet is more pleasant. But when someone drinks an entire can, then another, then buys a case for the week, excessive sweetness starts to wear them down. A sip test and a bottle test are different experiments, and Coca-Cola conducted only the first. In essence, the company optimized its product for conditions in which nobody consumes it in real life.
By early July 1985, people inside the company already understood: the data no longer mattered. The market is not a laboratory, and two hundred thousand tastings weighed less than thousands of letters from people who felt betrayed. On July 10–11, 79 days after New Coke launched, Coca-Cola announced the return of the original formula. It was given the name Coca-Cola Classic.
The news of the return was significant enough that ABC interrupted the daytime broadcast of General Hospital for it — the announcement was delivered by anchor Peter Jennings. For the soap opera’s audience, it was an almost surreal moment: the drama on screen was halted for a drama about soda. Coca-Cola’s hotline received 31,600 calls in two days after the announcement — but now people were calling to give thanks. At a press conference, Donald Keough uttered a line that became perhaps the most honest corporate admission of the decade: he said the company had not been either dumb enough or smart enough to plan all this in advance.
New Coke did not vanish at once — it remained on shelves alongside Classic for several more years, gradually losing sales until it became a ghost, and then was discontinued altogether. Coca-Cola Classic, by contrast, survived a sales surge: its return drew attention that no advertising campaign could buy.
It seems important to me that this story is not about corporate stupidity. The people making the decision were competent. The data had been collected properly. The methodology was standard for the industry. The mistake lay deeper: in the very idea that consumer choice can be reduced to a sensory response isolated from context. Coca-Cola tried to solve a cultural problem with an engineering method — and discovered that culture does not fit inside a paper cup.
The difference between a sip and a purchase is the difference between dating and marriage. On a date, the brighter one wins. In a marriage, the winner is the one who is more familiar, more dependable, who is built into the daily rhythm so thoroughly that their absence feels like a physical loss. Coca-Cola was a marriage for millions of people, and the company offered them a divorce on the grounds that another candidate was a better kisser.
After 1985, the consumer-goods industry began treating blind tests differently. It did not abandon them — but it stopped treating them as the only compass. Companies began asking the questions Coca-Cola had failed to ask: what happens if the product disappears? What role does it play in the buyer’s life beyond taste? Coca-Cola brought back its formula, restored trust, and by the end of 1985 Classic sales were surpassing both New Coke and Pepsi. Goizueta remained CEO until his death in 1997, turning Coca-Cola into one of the most valuable companies in the world. Keough retired in 1993. Neither ever called New Coke a success — but both acknowledged that the return of Classic gave the company an emotional bond with consumers that no advertising campaign could have created. Gay Mullins disbanded his organization after the classic formula returned and went back to ordinary life — no one ever reimbursed his thirty thousand dollars.