Big Launches are Hard to Predict – The Biggest Marketing Blunder of all Time^[a]
Ronny Kohavi
29 July 2024

I recently saw yet another presentation that claimed that the Pepsi challenge back in the 1980 was flawed because they gave people just a sip and not the whole can, and therefore people prefer a sweeter first sip.  The implication is that the market research that Coca-Cola did before the introduction of New Coke, in response to the Pepsi Challenge, had an obvious flaw—how dumb could they be?  As with many such tales, the simple explanation is inaccurate, and the story is more complex.  Without knowing more, just think how unlikely it is that professionals doing market research for a famous brand like Coca-Cola would be so naïve; indeed, they were not.

Here is a quick summary from a great book on the topic: The Real Coke, The Real Story by Thomas Oliver.  Here are some excerpts with minor rewordings for readability:

• In the 50s, coke outsold Pepsi two to one, but by 1984, they were leading by a mere 4.9%, and in grocery-stores, they were trailing by 1.7%.  Coca-Cola spent far more than Pepsi on advertising, it was competitively priced, and was more widely distributed.   If marketing was not the problem, it must be the product itself.  
• The Pepsi Challenge, a campaign by the competition, presented results of comparative taste tests across America, showing a clear preference for Pepsi.  The advantage was slim: 52 to 48, but it gave Pepsi the legal right to claim their product was superior.

Coca-Cola countered with the message: “One sip is not enough,” that is, a brief taste might come out in favor of the sweeter drink, but with a whole glassful, you would appreciate the body of “zip” of Coke.

• Stout, who had a Ph.D. in economics and headed the marketing-research department, claimed that given all the data about the lack of influence of other factors, you “have to begin asking about the taste.”
• A test was done with 200 households of heavy soft-drink users.  The new trial product was put into the regular coke bottles and two cases were shipped to each family, saying it was “processed differently.”  The response was not rebellion but overwhelming approval.  
• The technical division persisted in its efforts to brew up a Coke that would be more pleasing than Pepsi and that would arouse the public with a taste that could not be ignored. In September 1984, they said they had it….in blind taste-tests the new formula beat Pepsi by a margin of six to eight points, whereas Pepsi had hammered old Coke by anywhere from ten to fifteen points in earlier tests. “That’s an eighteen-point swing,” Stout declared. Still more mind-boggling were results showing that Pepsi drinkers who previously chose Pepsi over Coke something like seventy to thirty in blind taste-tests now picked new Coke 50 percent of the time.
• On April 23, 1985, executives of the Coca-Cola Company held a press conference in New York City, offering a new drink with a new taste, under the name Coca-Cola.  The change, they claimed, was based on nearly 200,000 consumer taste tests, which had revealed a resounding preference for the new flavor.  
• President Donald R. Keough said the numbers represent a “staggering superiority” and [CEO] Goizueta told the press that it was “the surest move the company ever made.”
• Within forty-eight hours of the press conference, according to several studies, more people knew about the new taste of Coke than knew who was president of the United States.
• When the company was questioned about new Coke’s potential, it responded: “It can’t fail.” And when asked if there was any possibility of bringing back the old Coke, spokesmen unequivocally said, “Never.”
• Yet just 79 days later, on July 11, 1985, the executives publicly apologized to the American people, saying they would reissue the original Coca-Cola formula under the name Coke Classic.
• The company rebranded the new formula “Coke II” in 1990 before it was eventually abandoned in 2002

Britannica’s article on New Coke claims the introduction of New Coke is the biggest blunder of all time.

My thoughts:

1. In the Hierarchy of Evidence, A/B tests (randomized controlled experiments) are near the top, with the top spot held for meta-analysis of such experiments.  As you go down the hierarchy, the level of trust should go down.  Market research is hard to get right, even by experts.  The backlash, which was actually seen in the studies, was stronger than estimated.  President Keough said “all the time and money and skill poured into consumer research on the new Coca-Cola could not measure or reveal the deep and abiding emotional attachment to original Coca-Cola felt by so many people … the passion for original Coca-Cola—and that is the word for it: passion—was something that caught us by surprise.”
2. Big launches are risky.   Coca-Cola could have introduced this as a second line, like they did with diet coke, but two reasons were given against this at the time:

1. Concern about being in a position of having the flagship brand not be number one in taste.
2. Two Cokes would split Coke’s market share and Pepsi would become the number-one soft drink, an intolerable situation for Coca-Cola.

Hindsight is 20/20, and the big launch was a mistake.  Coca-Cola did have both cokes available for several years afterwards.

3. Coca-Cola could have slowly tweaked the taste over time.  There’s a famous story where eBay’s background color was yellow. When they tried to change it to white, users revolted, so they changed the color slowly, one level at a time, over a year (see shaping user behavior).
   Here is an interesting data point from the book: “Stout [head of market research] was conducting taste tests, and in blind tests new Coke was still beating Pepsi, but Pepsi was winning overwhelmingly in tests that identified the brands. The very idea of a new Coke was inspiring an extremely negative reaction.”  Coca-Cola has tweaked the taste over time, especially with the switch from sugar to high fructose corn syrup, but it was not a big launch.  Could they have done the same and tweaked slowly towards the new reformulation?   Maybe.

Other resources

• https://en.wikipedia.org/wiki/New_Coke
• All about New Coke video.
  Goizueta claims it’s about the consumer connection to the brand.
• “New Coke” – is it?
  The research failed to consider the sentimental value of the consumers in relation to the product. 
• The 'New Coke'​ Epic Fail
• Don Keough Speech Classic Coke video
  Fun humility lessons.
  Some critics will say: Coca-Cola has made a marketing mistake; some cynics say that we planned the whole thing.  The truth is : we’re not that dumb, and we’re not that smart.
• The ‘New’ Coke: Market Research Failure. And …. how to avoid it!
  Another claim that it’s the sip-test problem, which appears to be inaccurate.