Pepsi's Fifth Attempt at a Product Nobody Asked For
How I work these cases: Before my MBA starts, I'm practicing the case method on my own. I give an AI a business case and ask it to release the details in stages, the way a class discussion unfolds. I give my read at each stage without looking anything up. Then the real outcome is revealed and I compare it with my reasoning. I call this working a case cold. These posts are my write-ups.
This case uses the facts available at the launch of Pepsi Next. I worked from them with no knowledge of the outcome.
Pepsi Next launched in February 2012, and the case ends at launch without saying how it went. It was a mid-calorie cola with 60 calories per can, about 60% fewer than regular Pepsi. It was aimed at people who had left cola for water or energy drinks over health concerns but never liked the taste of diet soda. The case mentions almost in passing that this was Pepsi's fifth attempt at a mid-calorie cola since the 1970s, and that the first four all failed within a couple of years.
Central question
Can a mid-calorie cola find real demand between people who don't care about calories and people who have already accepted diet soda's taste?
Key ideas
My first move was to split the market into two groups: people who don't care about calories and drink whatever tastes good, and people who do care and have already accepted diet cola's taste. A mid-calorie product sits in the gap between them, and neither group wants it. That was my whole thesis, built before I saw a single number.
The case pushed back with a group I hadn't accounted for: people who care about sugar but dislike diet cola's taste, and who left the category instead of switching to Diet Pepsi. They are different from both of my groups. They are a real gap that a better-tasting reduced-sugar product might fill.
Then the numbers arrived, and they looked strong. Blind taste tests came back enthusiastic. Market tests in Iowa and Wisconsin beat internal targets on trial, repeat, and incremental sales. At national launch, Pepsi Next reached almost 1% market share within weeks, which Pepsi's own CEO called unusually fast for a new product. I said I didn't see a reason it might fail.
My synthesis
That was a mistake I could have caught. Two exchanges earlier I had separated trial from repeat purchase, and I didn't apply that distinction to the data. Free sampling drives trial, and early numbers like these mostly measure it. They didn't tell me whether someone who tried the product once came back and paid for it again.
Once I applied my own distinction to the data, my read flipped. A product built on curiosity produces an early spike whether or not it is good. The real test was always what happened after the free samples ran out.
What happened next
Market share fell from about 1.0% to 0.6% within a few months. Convenience stores reported excellent distribution and weak repeat sales. Pepsi launched two new flavors shortly after, probably to revive a brand that was already losing momentum. Pepsi Next was discontinued in the US in 2015. That was the fifth failure in the same category, for nearly the same reason as the first four.
Questions I still have
How do I distinguish trial-driven early metrics from genuine repeat-purchase demand before the data has had time to mature?
This was Pepsi's fifth mid-calorie cola attempt since the 1970s. What does it take for a company to treat a repeated category failure as a pattern and not a fixable execution problem?
Final perspective
What I want to keep from this case is the mid-case wobble: numbers that looked convincing talked me out of a sound framework, and I caught it before it became my answer. Staying with a sound framework when the data seems to disagree is the harder test.