Who is already serving coffee on this street?

The unit of analysis

What coffee shop location analysis actually examines

General business location analysis asks what is true around any address. A coffee read narrows that to a format whose economics run on a clock rather than a meal plan: the walk-in at 7:40am is not the customer at 3pm, and the population figure underneath both of them counts residents, not whoever is standing at the window at that hour. So the coffee version of the question is not “is this street busy” but “who is on it, when, and what is already selling them the same cup.”

That is where a coffee page earns its own existence rather than borrowing a restaurant one. Two things separate them concretely: the place types that define the category are different — a bakery competes for the pastry basket in a way a sit-down kitchen never does — and the time of day decides which anchors matter, in a way that does not apply to lunch traffic. Everything on this page is filtered through those two facts.

It is still not a shortlist tool — comparing candidate addresses is site selection — and not a score. It is what a careful operator checks about one address, one category, before anyone signs.

The coffee-side evidence

The cafes, coffee shops and bakeries around your point

The analyzer's activity controls carry a Coffee Shops preset. Selecting it does two things across the analysis surface: the competitive view orders the coffee-matched places by distance — its own definition, in the interface's words, is “a view over the same nearby places data; facts from the place records, no scoring” — and the category mix shows how the coffee supply sits against everything else loaded.

“Matched” is honest about its mechanism. On the live Places path, matching uses documented place types: cafe, coffee shop, pastry shop, dessert restaurant and bakery — which is why a bakery or a pastry shop appears in a coffee read where it competes for the same morning basket. On the open-data path, places from Overture and OpenStreetMap fall into a Food & Drink bucket through the platform's category grouping, and every record keeps its original source label viewable underneath, because grouping is a reading aid, not data. Density is computed as unique loaded place records per square kilometer of the study radius: a raw count with arithmetic attached, not an index of how hard your opening will be.

What a café listing can show, and where it comes from:

Beyond the direct competitors

What the rest of these streets hold

A café trades on anchors rather than on the street as a whole. The categories worth counting are the ones that put a person at a window at a specific hour: offices and coworking floors for the 7–9am run, education for the gap between classes, hotels for guests who have nowhere else to sit, gyms for the after-work cup, and residential blocks for the slow afternoon. Loaded places are grouped into sixteen readable categories so those counts can be asked for by name, and the inevitable “Other” bucket stays clickable down to the raw source tags behind each place, so no classification is taken on trust.

The two opt-in tools answer the two questions this produces. Co-location counts the categories you specify around the site — and the coffee-specific move is to run it per anchor rather than all at once, because “12 offices” and “12 offices inside 300 metres” answer different questions. White Space probes categories you suspect are missing from the loaded sources entirely. Both are explicit about their limit: “not observed” means no record was found in the loaded sources for this area — absence of evidence, kept clearly apart from evidence of absence, and never a claim that the office park is empty. The general mechanics belong to competitor analysis; what a coffee read does with them is apply them to the anchor-and-hour question rather than to the street as a whole.

Audience and access

Who lives within reach, and how minutes travel

Inside your study extent — a radius from roughly 100 metres to 20 kilometres, or a boundary you draw on the map yourself — population is summed from gridded datasets and shown beside the nearest city context: who is in the area, per the grid's own version and vintage. Read it for what it is: a residential headcount. It is not a count of commuters, students, shift workers or a morning queue, and the platform does not convert it into one. This matters more for coffee than for most formats precisely because the residential figure and the actual customer base diverge hardest there — an office district and a residential block print the same number here while behaving like two different businesses.

Reach is also the instrument where coffee is the odd case out, and the page says so rather than repeating the general pitch. The drive-time probe measures car travel over sampled points — average, fastest, longest, the share inside ten, fifteen and twenty minutes, with unroutable points excluded and counted — and none of that describes a customer who walks 300 metres. For a coffee trade the driving-distance rings are close to decoration: the useful question is not how far a car can get but what sits inside walking distance, and that is the radius you draw. The same measurement, framed for the right format, is worth keeping for drive-through and for the second shop in a catchment; it is just not the primary lens here. The general method is documented at trade area analysis and the drive-time methodology.

Where the evidence ends

What coffee shop location analysis here does not do

• No success arithmetic. Revenue, footfall, profitability and demand estimates are not computed by anything on the platform — the analysis note that ships with results reads “raw observations only — no ratings, no predictions, no advice,” and no feature quietly adds them.
• No verdict, no “best location.” There is no coffee score, no ranking of addresses, no recommended site. Density and competition are facts you weigh; deciding what a saturated café street means for your format and your daypart is yours.
• No daypart model. The platform does not know who drinks coffee at 7am, who takes one at 3pm, or who is in the area at all on a Tuesday. Population is residential; an office district and a residential block produce the same number here for very different businesses.
• No property economics. Rent, income and employment data are not in the catalog at all — a gap stated as a gap, not estimated around.
• No coverage pretense. Thin source coverage shows as thin or empty blocks with their reason attached; a quiet count over an under-covered area is exactly the silence the labeling exists to expose.

One pass in a larger program

Where a café read sits in the wider platform

The coffee pass sits on the general machinery: it is the category-tuned lens on the point dossier, asks its supply questions through competitive evidence, sizes audience inside the catchment you choose, and hands surviving candidates to the shortlist compare. For a single opening, that is the whole chain; for a group adding its fourth site in a year, it continues into expansion analysis, where the campaign keeps the evidence dated and re-checkable instead of left in a spreadsheet from March. Screening twenty candidate streets at once is a bulk job. And the method under all of it is location intelligence, with how each figure is computed and which sources feed it both published rather than implied.

Where coffee genuinely overlaps another format on this platform, the crossover is explicit: restaurant location analysis runs the same dossier through the wider Food & Drink category, and a mixed-format operator — a café that also serves lunch, or a group with both formats — should read the two together rather than assume the coffee preset speaks for the whole food trade.