Methodology

We do not test machines

We have not brewed, tasted, dismantled or physically tested any machine in this database, and no page here claims otherwise. Every specification is transcribed from the manufacturer's own published product page, and every machine page links back to the exact source it came from. Every score is arithmetic applied to those published specifications by the formula below.

This is a deliberate constraint. A site that claims to have tested sixty machines has usually tested three and paraphrased the rest. We would rather tell you what the manufacturer actually publishes, show you what it does and does not mean, and be checkable.

Where a manufacturer does not publish a figure, we store it as null and render an em dash. We never estimate a spec, never take one from a retailer listing or a review, and never fill a gap by analogy with a similar model.

Pump pressure is not brew pressure

This is the single most misleading number in the espresso market, and it is printed on nearly every box: 15 bar, 19 bar, 20 bar.

That figure describes the maximum pressure the pump can generate against a closed outlet. It is not the pressure at the coffee. Espresso extracts at roughly 9 bar, and a well-designed machine regulates its pump down to about that using an over-pressure valve, which bleeds the excess back to the water tank. A machine advertising 20 bar and a machine advertising 9 bar can pull an identical shot, because both are brewing at 9.

A higher pump rating is not a better machine, and a machine that genuinely delivered 20 bar to the puck would make worse espresso, not better — excessive pressure compacts the coffee bed and promotes channelling.

So we show both numbers, on every pump machine, everywhere on the site: the figure the manufacturer advertises, labelled as marketing, and the ~9 bar the coffee actually sees. Several manufacturers — Breville and Gaggia among them — publish both themselves. On bean-to-cup and capsule machines, where the pressure is internal and not user-adjustable, we say that rather than presenting a number that means nothing to the owner.

The scoring formula

Scores are deterministic: the same inputs always produce the same output, and no number is hand-adjusted. Each of six components is scored from published specifications, converted to a percentile within the machine's own category, and combined with fixed weights.

The six components

Two rules that matter more than the weights

Percentile within category. A super-automatic is ranked against other super-automatics, a semi-automatic against other semi-automatics. This is why a great bean-to-cup machine and a great prosumer machine can both score in the eighties: each is the best answer to a different question. Scores are not comparable across categories, and should not be cited as if they were.

Missing inputs are dropped, not zeroed. If a manufacturer does not publish the inputs for a component, that component is removed and the remaining weights are renormalised. A machine is never penalised for its manufacturer's silence, and a gap is never filled with an assumed value.

Note also that ease of use and espresso potential deliberately pull against each other. A machine that removes every decision scores well on one and poorly on the other, and that is the honest description of the trade-off — not a flaw in the scoring.

True cost of ownership

The purchase price is the largest single number but not the whole cost. On each machine page we show a first-year total built from the published list price plus four running costs: beans at roughly $260 a year (about a pound a fortnight, at typical US specialty pricing), descaler at about $30, water filters at about $40, and electricity computed from the machine's own published wattage at the US average residential rate.

Consumable figures are assumptions for roughly two drinks a day and are stated as such. Electricity is computed from a published spec, so machines that do not publish wattage get no electricity line rather than an invented one. Where a manufacturer publishes no list price, we do not compute a first-year cost at all — the biggest term would be a guess.

Where specifications come from

Specifications come only from official manufacturer sources: breville.com, delonghi.com, us.jura.com, gaggia-na.com, ranciliogroup.com, philips.com and its published product leaflets, lelit.com, rocket-espresso.com, profitec-espresso.com and ninjakitchen.com.

Authorised specialist retailers may be used only to confirm a figure the manufacturer sheet omits, never as a primary source and never for a verdict. Review sites, aggregators, video channels and blogs are excluded entirely as sources of record for specifications, prices or claims. Prices shown are the manufacturer's published list price plus any retail snapshots we have tracked ourselves.

How we make money

Affiliate links, including Amazon Associates. As an Amazon Associate we earn from qualifying purchases. Buying through a link on this site costs you nothing extra.

Rankings are computed from specifications by the formula above and are never sold, negotiated or adjusted for commercial reasons. No manufacturer has seen this database before publication, and no placement on any list is or has been paid for. There is no advertising on this site. Full disclosure.