The short answer is that three were good enough to sell and the others weren't.
The longer answer is more useful, because it explains how we make decisions and what happens to the ones that don't make it.
We asked 426 people
Not friends. Not family. Not the people who tell you it's lovely because they like you.
We sent samples to 426 people and built a questionnaire specific enough that "it's nice" wasn't an available answer. Sweetness on a sliding scale where the middle meant just right. Whether the ginger was too sharp, nicely warm, gentle, or undetectable. What they'd change. What they'd actually pay.
And the question that matters more than any other: would you buy this. Which is a completely different question from do you like it, and produces noticeably different answers.
Across three flavours we got 1,073 individual verdicts back.
What came back
Ginger Lemon won. 46% ranked it first, and it had the highest proportion saying they'd actually buy it. The most repeated request was more lemon — 106 people said the lemon was "there but gentle."
So the recipe you'll get has more lemon in it than the one they tasted. That's the entire point of asking.
Mango Blaze split the room. Roughly a third loved it. Roughly a third said it was fine. The mango strength was judged just right by a clear majority, and the mouthfeel came back as "light and clean" from 200 people, which is exactly what we were aiming for.
The third one was harder reading. A clear majority said it was "fine, but missing something."
That's a worse result than being disliked. Disliked means it made an impression and someone will love it. "Fine" means nobody will ever choose it on purpose.
What we did about it
We could have shipped it anyway. Three flavours looks better on a shelf than two, and "fine" isn't a failure — plenty of drinks sell perfectly well on fine.
Instead we went back to it. The panel also told us what direction to go: 170 said they'd prefer it with a warmer, rounder character over the version they tried. That's not a vague signal, it's a brief.
Why not launch with six
There's a real temptation to fill the shelf. More options look like more choice, they take more retail space, and they give people a reason to come back.
But three flavours somebody would reorder beats six where two exist to make up numbers. Every extra product is inventory, packaging design, a production run and shelf space, and if it's only there for optics it's costing you on all four.
A small range also means we can change things. When 106 people say the lemon is too quiet, that's a formulation change we can make before launch rather than a note for version two next year. That gets much harder at six products.
What happens next
More flavours will come. They'll come the same way these did — sampled widely, questioned specifically, and dropped if the answer comes back "fine."
If you want to be on that panel next time, the fastest route is to tell us what you think of these three. Every one of those 1,073 verdicts changed something.
What we'd do differently
Two things, honestly.
We'd have asked about price earlier. We got willingness-to-pay data at the end, and the modal answer came back at ₹60 a can. Useful to know, but better known before you've set your packaging and your production run.
And we'd have tested the third flavour on its own rather than in a flight of three. Tasting three drinks back to back changes how each one reads — the middle one always suffers, and there's decent research on this. A flavour that's fine in isolation can come across as forgettable simply because it followed something sharper.
We're not sure that's what happened. But we didn't design the test to rule it out, which is a flaw in the test rather than the drink.
On price
Since we mentioned it: our six-pack works out at ₹99 a can, against a panel that said ₹60. The twelve and twenty-four packs bring that down to about ₹83.
Panels almost always understate what they'll pay — it costs nothing to say a lower number in a survey. But the gap is worth naming rather than quietly ignoring, and it's part of why we launched with packs rather than single cans. The per-can maths looks different when you're buying six.
The wider point
Most brands run this kind of testing and then publish only the flattering parts.
We'd rather say that one of our three flavours came back as "fine, but missing something," because that's the number that actually shaped what we did next. A brand that only ever reports good results isn't testing anything. It's collecting compliments.
