Pepsi's Fifth Attempt at a Product Nobody Asked For

This case was worked cold, using the moment-of-launch facts for Pepsi Next, no outcome in sight until after the analysis was done.

The central question

Pepsi Next launched in February 2012, and the case doesn't tell you how it went. It ends at launch: a mid-calorie cola, half the sugar of regular Pepsi, aimed at people who'd left the cola category for water or energy drinks over health concerns but never warmed to the taste of diet soda. What the case mentions almost in passing is that this was Pepsi's fifth attempt at a mid-calorie cola since the 1970s. The first four all failed within a couple of years.

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 already made peace with diet cola's taste. A mid-calorie product sits in the gap between them, wanted by neither. That was the whole thesis, built before seeing a single number.

The case pushed back with something specific, a third group I hadn't accounted for: people who care about sugar but specifically dislike diet cola's taste, and left the category entirely rather than switch to Diet Pepsi. That's not the same group as either of my two. It's a real gap a genuinely better-tasting reduced-sugar product might fill.

Then came the actual numbers, 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 hit almost 1% market share within weeks, unusually fast for any new product, according to Pepsi's own CEO. I said, honestly, that I didn't see a reason it might fail.

My synthesis

That was the wrong move, and I'd flagged the reason myself two exchanges earlier without noticing I'd flagged it: trial and repeat purchase are not the same thing. Free sampling drives trial. It always does. None of the strong early numbers said anything about whether someone who tried it once came back and paid for it again.

Once I applied my own distinction back to the data, the read flipped. A product built on curiosity generates exactly this kind of early spike regardless of whether it's actually 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, likely trying to reenergize a brand that was already losing steam. It was discontinued in the US by 2015, the fifth failure in the same category, for close to the same reason as the first four.

Questions I still have

How do you 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 rather than a fixable execution problem?

Final perspective

The part of this case worth keeping isn't the final call. It's the mid-case wobble, getting talked out of a sound framework by numbers that looked convincing, and catching it before the wobble became the answer. Being right the first time is useful. Staying right when the data starts arguing with you is the harder skill, and the one that actually gets tested.