A Glassdoor for Office Coffee: Why Workplace Coffee Deserves Its Own Ratings

Deskbrew is essentially a Glassdoor for office coffee. Instead of rating salary, management or culture, people rate the coffee at individual office locations from 1–10 and compare those locations publicly.

Every city has guides to its cafes. Roasters are reviewed, espresso bars are scored, and the flat white two streets away has a rating attached to it. Meanwhile the coffee that people actually drink most — five days a week, at their desk, for years — has never been written down anywhere comparable.

That coffee gets discussed constantly, in the kitchen queue and in the lift, and the discussion evaporates every time. Nobody knows whether the machine on the fourth floor is genuinely worse than the one on the second, whether the company's other site does it better, or whether the building across the street is quietly excellent. The information exists in thousands of heads and nowhere else.

Deskbrew is the boring fix for that: write it down, in the same shape every time, so it can be compared.

The closest familiar model is an employer-review site. The mechanics match: ratings come from the people who are actually there, nobody needs the company's permission to leave one, and the result is public whether the employer likes it or not.

The difference is scope. Employer reviews collapse salary, management, workload, and the commute into one star, which makes the star useless for any specific question. Deskbrew asks a single thing — how good did the coffee taste, on a scale of 1 to 10 — and a narrow question is what makes answers comparable across thousands of offices.

It also means a score is not a verdict on an employer. An office with mediocre coffee can be a superb place to work, and a 9 out of 10 says nothing about anything except the cup.

The scope rule is strict: one physical office location of one company is one entry. Public coffee bars, roasteries, and cafes are out, because those are already well covered by other guides and because they answer a different question.

The unit matters just as much. Coffee is decided building by building — a different machine, a different supplier, a different cleaning rota, sometimes a different budget holder. A company with six sites therefore shows six scores rather than one average, because the average would be true of nowhere.

That gives two useful views. The company page ranks a company against itself, which is how you find out whether your building drew the short straw. The city ranking puts every mapped workplace in a city in order of taste, which is how you find out whether the office across the road really does have better coffee, or whether that is just a rumour that survived because nobody checked.

• Company page — every mapped office of one employer, each with its own score
• City ranking — every ranking-eligible office in one city, ordered by taste
• Map — the same data geographically, for when the question is about a neighbourhood

Alongside the score, each office carries facts: whether coffee is free, subsidised, or paid; which milks are stocked; whether it comes from a portafilter, a bean-to-cup machine, capsules, filter, or a staffed barista; and whether visitors can get a cup.

None of those move the score, and that restraint is what makes them useful. Because cost is a filter rather than an ingredient, 'is paid coffee actually better?' becomes a question you can answer by comparing the two groups' taste ratings instead of assuming an answer from a price tag. The same goes for machine type: the specification sheet has an opinion, but the ratings have evidence.

None of this is solemn. It started as a joke that turned out to be answerable, and the right way to take a coffee leaderboard is lightly.

But the data underneath is real, and what people do with it is their own business. Some settle an argument. Some check a prospective employer's building before an interview, on the theory that an office which maintains its machine maintains other things too. Some show a slipping score to whoever signs the coffee contract — an anonymous, structured, dated signal that is considerably easier to act on than a shrug in the kitchen.

All of those are legitimate uses. The only claim Deskbrew makes is that the comparison is now something you can look at, rather than something everybody guesses at.