Coffee performance
Taste, temperature, repeatability and category-appropriate drink quality.
CoffeeScore methodology · version 0.2
A score has to mean something.
CoffeeScore is being designed as a category-aware, evidence-backed assessment—not a repackaged customer rating. Scores remain unpublished until a machine completes the required tests.
Current statusFramework published0 machines scoredTest handbook in pilot-draft state
Milk performance
Texture, temperature, speed and repeatability where the machine includes a milk system.
Ease of use
Setup, controls, workflow clarity and consistency for the intended buyer.
Cleaning
Daily effort, descaling, access to removable parts and avoidable mess.
Noise
Repeatable sound measurements during grinding, brewing and milk preparation.
Value
Performance and expected ownership cost against machines in the same category.
Build quality
Materials, fit, stability, parts support and service-oriented design.
Reliability
Warranty, failure evidence, repair pathways and long-term ownership data.
Category-aware weighting
Different machines should not be judged as if they do the same job.
Every row totals 100%. A zero means the dimension is not part of that category’s score.
CategoryCoffeeMilkEase of useCleaningNoiseValueBuild qualityReliability
Espresso25%15%12%10%8%12%10%8%
Bean-to-cup20%20%15%12%8%10%7%8%
Pod25%0%20%10%10%18%9%8%
Filter35%0%15%10%8%15%9%8%
Publication gates
No score until the evidence clears every gate.
Passing these gates permits publication; it does not guarantee a high score.
Retail-representative machine
Testing must use a normal retail unit with its model and firmware recorded.
Repeatable preparation
Core drinks are repeated under a published recipe rather than judged from one successful cup.
Measured evidence
Temperature, time, noise and energy fields must retain raw observations and test conditions.
Ownership cycle
Daily cleaning and the manufacturer’s descale or deep-clean process must be completed.
Commercial separation
Retailer commission, samples and manufacturer access cannot change a score or its weighting.
Visible limitations
Missing long-term reliability evidence must stay marked provisional instead of becoming a guessed number.
Scoring rules
Claimed, measured, provisional and tested are different states.
- Manufacturer specifications remain labelled as claimed.
- Measurements retain the equipment, conditions, date and raw observation.
- Dimension scores use a 0–10 scale; the weighted overall score is rounded to one decimal place.
- A score can be provisional when long-term reliability evidence is incomplete.
- Material test-protocol changes create a new methodology version.
- Commercial relationships cannot alter results, weights or publication timing.
From framework to testing
The category test handbook is now public.
It defines shared controls, recipes, repetitions, measurements and the raw-data contract. Equipment and tolerances remain visibly draft until pilot testing validates them.