Site selection for a clinic: start from the patients, not the building

Choose a clinic site by working back from the patients it will serve. Count the people of the right age who live within the distance they will travel for that kind of care, see how many clinics of the same kind already serve them and how well, and make sure the patients you want can physically get in: parking and a short level walk for older people, a route past schools for children, a lunchtime slot near work for adults. A beautiful building in the wrong place stays half empty.

This article is for doctors opening a practice, clinic groups adding branches, and physiotherapy, dental, eye and diagnostic businesses. It is organised by patient group, because the right site for a paediatric clinic and for a memory clinic share almost nothing. A worked example in Hyderabad and an AI employee that screens addresses follow.

First, decide who the clinic is for

Write one sentence: "Most of our patients will be ...". Everything else follows from it. A general practice serves everyone near it; a specialist clinic serves a group from further away. The Demographics block gives nine age groups, so you can count exactly the group in that sentence. Population figures exist only in some countries, among them France, Canada, India, Thailand, Egypt and Australia; not the United Kingdom or the United States.

If the patients are children and families

• Count: residents aged 0 to 9 and 10 to 19 within a 10 to 15-minute drive, or a walk in dense cities.
• Look for: primary schools and nurseries nearby, which bring parents past twice a day. Nearby places counts them.
• Watch out for: a site on a road parents avoid at the school run. Traffic Flow shows the live speed and jam factor on the roads next to it; it measures speed, never the number of cars.

If the patients are older adults

• Count: residents aged 60 and over, and the share they make of the area, within a short drive.
• Look for: a ground-floor unit, parking at the door, a pharmacy close by, and a bus stop within a short level walk.
• Watch out for: hills, stairs and wide roads to cross. A site that is close on the map but across a busy road is far for someone who walks slowly.

If the patients are working adults

• Count: adults aged 20 to 59 who live nearby, but also how busy the area is in the day. Mobility Footfall gives people-hours of presence by hour, about 10 days behind; a strong midday share means office workers who could come at lunch.
• Look for: offices, stations and gyms nearby, and evening hours on the road outside that are not jammed.
• Watch out for: business districts that are empty at weekends, if weekend appointments are part of your plan.

Every clinic: competition and quality

Count clinics of your own kind inside the reach, not every medical place. A dental practice competes with dentists, not with a hospital emergency department. Then read their ratings: a crowd of clinics rated under 4 is a different market from a crowd rated 4.7.

For a sense of scale, xMap's places database lists 329 medical clinics within 10 km of the centre of Toronto (November 2026). Of them, 36 belong to a chain and 293 are independent, rated 4.1 out of 5 on average, and 37 of the 328 with an opening date opened in the twelve months before. A market where one in nine clinics opened in the last year is moving quickly; any site chosen there should be checked again a few months later.

329Medical clinics listed

293Independent

4.1Average rating

37 of 328Opened in the last year

A worked example in Hyderabad

A made-up case. A physiotherapy group in Hyderabad treats mostly people over 50 recovering from joint surgery. It has two candidate units. The figures below are invented, to show the reading.

Made-up figures, for illustration only.

North has more older residents but more than four times as many competitors around them, and a first-floor unit that its own patients would struggle to reach. South has fewer older residents but far fewer competitors, three hospitals that discharge exactly its patients, and step-free access. South wins clearly, and the group's next step is to ask the hospitals' discharge teams whether they would refer.

What data cannot tell you

• Whether the unit can be licensed for clinical use, or meets the rules for your kind of care. Ask the regulator and a planning adviser early.
• Where referrals will come from. Count the hospitals and general practices near you, then talk to them.
• What patients can pay, or which insurers cover them.
• How quickly you can hire clinicians in the area.

Screening addresses with an AI employee

1. My data: the candidate addresses from your property agent.
2. Travel area: a 10-minute drive from each.
3. Demographics: residents in the age groups your patients belong to.
4. Nearby places: clinics of your kind, plus pharmacies and hospitals, inside each reach.
5. xMap LLM: score each site out of 100 against your one-sentence patient profile, with a reason, using only the figures given.
6. Make a report: one page per site for the partners' meeting.

The partners receive a short report: each site's patients, competitors, referral anchors and score. To keep watching after you open, Rating scorecard for my clinics tracks how you and nearby clinics are rated; Where to open sample collection points applies the same thinking to small satellite sites.