A Cold Root Beer and a Small System
It’s starting to get warmer out.
Which, at least for me, means two things:
- more time outside
- and the return of cold root beer
There’s something about a really good root beer that feels… complete. Not just sweet, but balanced. A little bite, a clean finish, maybe just enough carbonation to keep things interesting.
Naturally, I had a normal, reasonable thought:
I should build a system for this.
The Problem With Taste
Taste is subjective.
Everyone knows this.
But it’s also surprisingly inconsistent—even for the same person. One day something feels perfectly balanced. The next, it’s too sweet or not sharp enough.
That makes comparisons difficult.
Which, of course, makes it interesting.
Turning Root Beer Into Data
So I built a small system to try and bring a bit of structure to something inherently subjective.
Nothing overly complicated. Just a few attributes:
- sweetness
- bite
- aftertaste
- carbonation
- overall balance
Each one gets a score, and those roll up into a simple overall rating.
The goal wasn’t to be perfect.
It was to be consistent.
First Entry
Here’s one of the entries:
https://root-beer-reviews.onrender.com/rootbeers/695fc98acb0e4e4826b8118f
It breaks the score down across each attribute and shows how they combine into the overall rating.
A visual breakdown of a single root beer’s profile across key taste attributes.
What Shows Up Quickly
The interesting part wasn’t ranking root beers.
It was how sensitive the results were to the model.
A few things became obvious pretty quickly:
- Weighting matters more than individual scores
- “Bite” can completely change the perception of sweetness
- A strong first impression doesn’t always translate to a good finish
In other words:
The system didn’t just rank root beer. It exposed how I evaluate it.
The Real Lesson
The moment you try to quantify something human, you’re making decisions about what matters.
Those decisions shape the outcome more than the data itself.
What Comes Next
I’ll probably add to this occasionally over time.
Not at any fixed cadence. Just whenever I come across something worth testing.
If nothing else, it’s a good excuse to try more root beer.
Purely for research purposes, of course.
And One Final Question
What’s the best root beer you’ve had?
I’m always looking for the next data point.
(And for the record, my kid recently asked what root beer was made of. I told him: beer squared.)
If you’re curious, the system itself is open source and available in this repo on GitHub as well.
Top comments (0)