Site selection for a warehouse: what to measure before you sign

Choosing a warehouse site comes down to four questions: how quickly trucks get from the gate to the main roads, which customers or stores can be reached inside your delivery promise, whether you can hire and keep staff nearby, and what disrupts the roads around it. Shoppers walking past, the measure that decides a shop, hardly matters. Measure those four things the same way for every candidate, and the shortlist usually sorts itself.

This guide is for operations, logistics and real estate teams choosing a depot, a cross-dock, a micro-fulfilment hub or a regional distribution centre. It covers what to measure, what location data can and cannot tell you about each point, a worked example, the usual mistakes, and how an AI employee can keep checking the candidates while the lease is being negotiated.

The four things a warehouse site has to get right

• Road access. How long from the loading bay to a motorway or ring road, and how freely the roads next to the gate move at the hours your trucks leave and return. A site that looks close on a map can sit behind a congested junction.
• Reach. Which stores, customers or postcodes fall within a given drive. For a last-mile hub that drive is short and the count of addresses inside it is everything; for a regional centre it is the corridor that matters more than the radius.
• Labour. Warehouse work is staffed locally. A site in an area where many other operators are hiring the same roles will pay more and wait longer to fill shifts.
• Disruption. Recurring roadworks, closures and bottlenecks around the gate cost you every day, long after the rent is forgotten.

What location data can tell you about each one

Each of the four has a matching block in xMap Flows, and it helps to know exactly what each gives back.

• Traffic Flow reads the roads around each candidate right now: the current speed, the free-flow speed, the jam factor from 0 to 10 and the busiest road nearby. It gives speeds, never the number of vehicles. Run it on a timetable, say at 06:00 and 17:00 every weekday for a month, and you build up a picture of the gate at the hours that matter.
• Regional data has road traffic history for the United Arab Emirates: cars counted on the road nearest each place, hour by hour, since January 2021. Outside the UAE there are no car counts, only speeds.
• Travel area draws the area reachable from each candidate by road, in minutes or kilometres. For now it draws up to 15 minutes or 10 km, which suits last-mile hubs well; a regional centre's reach needs a few areas along its corridor, or a judgement on top.
• Nearby places counts what sits inside that reach: your own type of customer, other warehouses and logistics firms, truck workshops, fuel stations, places to eat for shift workers.
• Job postings shows who is hiring warehouse, forklift or driver roles near each candidate, which is the nearest thing to a live labour signal.
• Traffic incidents lists closures and roadworks in an area right now; checked every hour, it shows how often the roads round the gate are cut.

A worked example

A made-up example. A grocery delivery company wants a second hub in Dubai and has three units on its list: one near Jebel Ali, one in Al Quoz and one in Dubai Investments Park. Its promise is delivery within an hour, so it cares about a short drive and about the roads at the evening peak.

1. Put the three addresses in a table in My data, each with its position.
2. Draw a 15-minute drive around each with Travel area, then count the homes and shops it serves with Nearby places inside each area.
3. Ask Regional data for road traffic history on the road nearest each unit, month by month for the last year, to see how heavy and how slow the evening hours are.
4. Add Traffic Flow on a schedule at 18:00 every weekday for two weeks, to see today's speeds next to the history.
5. Count job postings for warehouse and delivery roles within 5 km of each unit.

Illustrative figures only, to show how the measures read side by side.

Al Quoz reaches the most customers but its gate road crawls at the hour the vans go out, and it competes hardest for staff. Investments Park is the easiest to staff and drive out of, but reaches fewer stores in 15 minutes. The decision is a trade-off between reach and reliability, and the table makes it visible instead of leaving it to whoever argues loudest.

Mistakes to avoid

• Measuring once. A single look at the roads on a Tuesday morning says little. Read speeds on a timetable across two or more weeks, at your own departure and return hours.
• Treating a circle as reach. Five kilometres across a river with one bridge is not five kilometres on open roads. Draw the area along the roads.
• Reading speeds as volumes. A slow road can be slow because of a junction, not because it is busy. Only the UAE history counts vehicles.
• Ignoring the neighbours. A cluster of other distribution centres means good truck services but also competition for drivers and pickers.
• Forgetting the night. If you load at 04:00, morning-peak congestion matters less than whether the access roads are closed for works overnight.

Doing it with an AI employee

You can set the checks up once and let an AI employee run them on the shortlist until the lease is signed. A flow that works well:

1. Schedule: every weekday at your departure hour.
2. My data: the table of candidate units.
3. Traffic Flow: roads within 500 m of each unit, for speed, free-flow speed and jam factor.
4. Traffic incidents: around each unit, for closures and roadworks.
5. xMap LLM: told to write one line per unit on how the gate roads behaved today compared with free flow.
6. Email or the Team inbox: a short note each morning, and a Make a report PDF on Fridays with the week's speeds and incidents.

A second flow, run weekly, can count warehouse job postings near each unit and new logistics businesses with Only new, so a new rival depot opening next door does not surprise you. When you are down to two or three units, Compare sites lays them out in one table with road speed, transport and the competition within 1 km.