How we rate bean-to-cup machines

Every bean-to-cup machine on this site is scored on the 8 criteria below, weighted to 100 points. This page is generated from the same rules that produce the numbers, so it cannot describe a system we are not using.

Each criterion says what it measures, why it carries the weight it does, and what the evidence actually is — which is often the more useful half. A score read without knowing that grind consistency comes from burr geometry rather than a measured particle distribution is a score taken on faith.

Drinks & Milk System

16 of 100 points

What one button actually makes, and what has to happen to the milk for it to happen. A machine that pours a flat white unattended and a machine that hands you a wand are sold as the same thing and are not.

What we can actually read: The drinks list and the milk system are both published, though makers count two sizes of the same coffee as two drinks and we count the maker's list.

Scored from figures that predict it, because the thing itself is not published.

The 3 rules below score this criterion out of 10, and they add up to 10 exactly. That result is then worth 16 of the category's 100 points.

  • Milk system3.5 of 10
    integrated carafe, lattego, latte go
    3.5 pts
    carafe, automatic, one-touch
    3 pts
    frother
    2 pts
    pannarello, panarello, steam wand, manual
    1.5 pts
    no milk system, none
    0 pts
  • One-touch drinks4 of 10

    Also read under Ease of Use, which asks a different question of the same fact.

    8 or more
    4 pts
  • Automatic milk2.5 of 10
    present
    2.5 pts

10 of 10

Grinder & Grind Control

15 of 100 points

The grinder is inside, so it is the one you get. Steps of adjustment decide whether the machine can be dialled in for a darker roast at all, and the burr material decides how long that stays true.

What we can actually read: Burr material and the number of grind steps are published by every maker in this category.

Scored from figures that predict it, because the thing itself is not published.

The 3 rules below score this criterion out of 10, and they add up to 10 exactly. That result is then worth 15 of the category's 100 points.

  • Burr material4 of 10
    titanium, stainless steel
    4 pts
    ceramic
    3 pts
    plastic
    0 pts
  • Grind steps4 of 10
    5 or less
    1.5 pts
    over 5, up to 8
    2.5 pts
    over 8, up to 12
    3.5 pts
    over 12
    4 pts
  • Built-in grinder2 of 10
    present
    2 pts

10 of 10

Cleaning & Upkeep

15 of 100 points

The heaviest cleaning criterion on the site, deliberately. On a machine you operate, cleaning is a chore. On a machine that operates itself, cleaning is the whole failure mode: a brew group that lifts out is rinsed under a tap, and a sealed one eventually becomes a service bill.

What we can actually read: Whether the brew group is removable is published, and it is the single most useful fact in this category.

Scored directly from published figures.

The 4 rules below score this criterion out of 10, and they add up to 10 exactly. That result is then worth 15 of the category's 100 points.

  • Removable brew group4 of 10
    present
    4 pts
  • Automatic rinse2.5 of 10
    present
    2.5 pts
  • Dishwasher-safe milk parts2 of 10
    present
    2 pts
  • Descaling indicator1.5 of 10

    Also read under Ease of Use, which asks a different question of the same fact.

    present
    1.5 pts

10 of 10

Ease of Use

12 of 100 points

These are bought to remove decisions, so the score is about how few are left: one-touch drinks, strength steps, and whether you can make one decaf without emptying a hopper of beans.

What we can actually read: The drinks list and strength settings are published. The pre-ground bypass often is not, even where the machine has one.

Scored from figures that predict it, because the thing itself is not published.

The 4 rules below score this criterion out of 10, and they add up to 10 exactly. That result is then worth 12 of the category's 100 points.

  • One-touch drinks3 of 10

    Also read under Drinks & Milk System, which asks a different question of the same fact.

    8 or more
    3 pts
  • Strength steps3 of 10
    3 or less
    1.5 pts
    over 3, up to 5
    2.5 pts
    over 5
    3 pts
  • Pre-ground bypass2 of 10
    present
    2 pts
  • Tells you when to descale2 of 10

    Also read under Cleaning & Upkeep, which asks a different question of the same fact.

    present
    2 pts

10 of 10

Capacity & Refills

12 of 100 points

How long it goes between refills, which is the difference between an appliance and an errand in an office or a large household.

What we can actually read: Hopper grams and tank litres are both published.

Scored directly from published figures.

The 2 rules below score this criterion out of 10, and they add up to 10 exactly. That result is then worth 12 of the category's 100 points.

  • Bean hopper5 of 10
    200 g or less
    1.5 pts
    over 200 g, up to 275 g
    3 pts
    over 275 g, up to 350 g
    4 pts
    over 350 g
    5 pts
  • Tank5 of 10
    1.2 L or less
    1.5 pts
    over 1.2 L, up to 1.8 L
    3.5 pts
    over 1.8 L
    5 pts

10 of 10

Value for Money

12 of 100 points

Points per pound against the category median.

What we can actually read: This is the most expensive category on the site, and the spread inside it is wide enough for the median to mean something.

Computed from price, not from a rule list

Value is the only criterion here with no thresholds of its own, because it cannot be read off one product. It asks what else the same money buys in this category, so it needs the category. That result is then worth 12 of the category's 100 points.

Points already earned on everything else
out of 10
Multiplied by the category median over this price
median ÷ price
Capped at
10
No price published, or fewer than 3 priced rivals
5 of 10

Half marks is not a verdict, it is a refusal to guess: with nothing to compare against, a product is neither good value nor bad, and scoring it either way would invent a finding. It corrects itself as the category fills.

Build Quality & Warranty

10 of 100 points

Casing, weight and warranty. The internals are sealed, so what a specification can honestly say about durability is mostly what is around them.

What we can actually read: Body material and warranty are published; country of manufacture usually is not, and the score says so rather than assuming.

Scored from figures that predict it, because the thing itself is not published.

The 4 rules below score this criterion out of 10, and they add up to 10 exactly. That result is then worth 10 of the category's 100 points.

  • Body material4 of 10
    stainless steel, titanium, brass
    4 pts
    aluminium, ceramic
    3 pts
    glass
    2 pts
    abs, plastic
    1 pt
  • Weight3 of 10
    7 kg or less
    1 pt
    over 7 kg, up to 10 kg
    2 pts
    over 10 kg
    3 pts
  • Warranty2 of 10
    none
    0 pts
    up to 1 yr
    0.8 pts
    over 1 yr, up to 2 yr
    1.4 pts
    over 2 yr
    2 pts
  • Manufacture location published1 of 10
    published at all
    1 pt

10 of 10

Kitchen Fit & Noise

8 of 100 points

Whether it fits under the cupboard, whether a travel mug fits under the spout, and how loud it is at seven in the morning.

What we can actually read: Cup clearance and noise are the two least-published figures here, so this criterion is often carried by the footprint alone.

Scored directly from published figures.

The 3 rules below score this criterion out of 10, and they add up to 10 exactly. That result is then worth 8 of the category's 100 points.

  • Footprint published3 of 10
    published at all
    3 pts
  • Cup clearance4 of 10
    110 mm or less
    1.5 pts
    over 110 mm, up to 135 mm
    3 pts
    over 135 mm
    4 pts
  • Noise3 of 10
    55 or less
    3 pts
    over 55, up to 65
    2 pts
    over 65, up to 75
    1 pt
    over 75
    0.5 pts

10 of 10

This category is too small to judge its own silence

We hold 7 bean-to-cup machines here. Everywhere else on this site, an input fewer than one maker in four publishes has its weight taken out of the total, because at that point we are measuring the market rather than the product. That rule needs at least eight products to mean anything — on a handful, one maker's silence looks like the whole market's. So nothing is struck out here, and scores in this category are out of the full 100.

One input is thin enough that it would have been struck out in a bigger category. Read the scores here knowing that:

  • noise db

How we rate everything else

The rules that apply to every category, where specifications come from and why blanks stay blank, are on how we score.