Crystal Pepsi

Crystal Pepsi failed because people tried it and didn't come back, says David Novak, the Pepsi executive behind it. He blames a taste that wasn't enough like Pepsi and his own rush to market over the bottlers' warnings.

What it was

Crystal Pepsi was a clear, caffeine-free cola that PepsiCo launched in the US in 1992, in a regular and a diet version.

PepsiCo's 1992 annual report called it "a unique-tasting clear cola with 100% natural flavors, no caffeine and no preservatives." It added that Time magazine had named it one of the best new products of 1992.

CNN reports that David Novak, a former CEO of Yum! Brands, is credited with the idea from his time as a Pepsi executive.

Novak told the Masters of Scale podcast in 2024 that the soft drink business was slowing while water and Clearly Canadian were taking off. His answer was a Pepsi with no color.

He also said it was the only Pepsi soft drink up to then launched at a premium price, which the bottlers set.

What happened

Pepsi put Crystal Pepsi into US test markets in 1992 and took it national at the end of the year.

Novak said it went into a test market in Boulder, Colorado, and that Dan Rather, then the CBS News anchor, reported the launch. By late September it was on shelves in Salt Lake City, where the Deseret News put it in front of a taste panel.

In December 1992 Pepsi announced plans for national distribution in regular and diet versions, "promising a major advertising campaign during the Super Bowl," the Associated Press reported. It had test-marketed the drink for six months first.

The following week, Coca-Cola USA president M. Douglas Ivester unveiled Tab Clear, which the AP described as "a clear, sugar-free and calorie-free version of its 29-year-old diet cola Tab."

Ivester said Coca-Cola kept the Coke name off it because a clear Coke would be an oxymoron that "does not compute." Paine-Webber analyst Emanuel Goldman said the company wanted to avoid a repeat of its mid-1980s fiasco with New Coke, when, in the AP's words, it "introduced a new Coke formula that bombed."

Crystal Pepsi sold well at first. PepsiCo's 1993 annual report said its case sales offset a decline in Pepsi and Diet Pepsi that year, and that it "particularly aided first quarter results, but did not significantly impact full year profits."

The same report said the drink was being refined and would come back in 1994 "in a nondiet version only." PepsiCo's 1994 annual report then recorded "significant declines in the Crystal Pepsi brands."

CNN reported that it was pulled off the market in early 1994.

PepsiCo has brought it back since. It was available for two days in December 2015 through a sweepstakes on Pepsi's app, and in June 2016 PepsiCo said it would sell it for a limited time from July 11 in Canada and August 8 in the US, CNBC reported. Fortune reported it back on US shelves on August 8. In January 2022 Pepsi announced a 30th-anniversary contest that would send six bottles each to 300 winners, CNN reported.

Why it failed

Crystal Pepsi failed, by its creator's account, because it didn't taste enough like Pepsi and he rushed it to market.

What PepsiCo said

PepsiCo's 1993 annual report gave "softening trends in the Crystal Pepsi brands" as the reason it was refining the drink for a 1994 relaunch. Neither that report nor the 1994 one says why sales softened.

What reporting showed

When the Deseret News testing panel tried it in September 1992, one family thought it tasted "a little like ginger ale or Sprite." A couple who called themselves "true-blue Pepsi drinkers" weren't too impressed.

Another panelist said that blindfolded, they'd have taken it for regular Pepsi. The paper's verdict was that "some Pepsi drinkers and a lot of teens" liked it.

Fortune wrote in 2016 that Crystal Pepsi had lasted about a year "after failing to catch on with consumers."

What later analysis found

In 2007 Novak told Fast Company that Crystal Pepsi was the "best idea" he ever had and the "worst executed," CNN reported. "People were saying we should stop and address some issues along the way, and they were right," he said. "It would have been nice if I'd made sure the product tasted good."

On Masters of Scale in 2024, he said Pepsi's bottlers told him it didn't "taste enough like Pepsi." He was after new drinkers and, he said, "I didn't care what they said."

The bottlers also told him that everybody would try it but "they're not going to come back." "Guess what, they didn't come back," Novak said.

He said Pepsi "rushed it into the marketplace," that in a few markets the drink was "sort of cloudy and hazy," and that "we really didn't put that into enough testing." In the same interview he said consumers who tried it before the test market "told us it tasted great."

Harvard Business School professor Amy Edmondson cites Crystal Pepsi in a passage on skipping basic homework, according to The Globe and Mail's summary of her book Right Kind of Wrong. The summary says the launch "failed to take into account clear drinks in clear bottles deteriorate easily and subsequently taste lousy."

The stage it reached

In the product life cycle, Crystal Pepsi never left the introduction stage, the launch period when a product manager should measure early retention rather than sign-ups.

Trial wasn't its problem. Plenty of people tried it, by Novak's account, and PepsiCo said it particularly helped first-quarter results in 1993. I think that strong start hid the retention problem until the full-year numbers came in.

It never reached growth, so I count it as a failure, not a decline. Its returns in 2015, 2016 and 2022 were limited-time promotions.

The lesson

I think the lesson of Crystal Pepsi is the bottlers' warning, which Novak says he ignored.

By his account, consumers told Pepsi the drink tasted great. The bottlers, who set its price, said it didn't taste enough like Pepsi and that people would try it and not come back.

Before a national launch, write down the repeat rate your test markets have to hit, and measure it. Then give the people closest to your customers, like Pepsi's bottlers, a real way to stop the launch through your product feedback loop.