Site mobilisation: the IT checklist your project managers wish you had

A typical SME IT partner is good at offices. They install switches in a comms cabinet, configure Wi-Fi for a fixed building, and patch a server quarterly. A construction site is none of those things. The 'office' is a portacabin, the comms cabinet is a plastic box on a desk, and the building moves to a different postcode in eight months.

Worse, the site has its own pace and culture. A delay of three days to get connectivity working is treated as catastrophic, because every other workstream is queued behind it.

A workable mobilisation pack contains: a router with dual SIMs (different carriers), a small managed switch, a couple of access points, a multifunction printer, a wired phone if the project still needs one, and printed laminated cards with the network details. It's pre-configured, tested, and sealed in a flight case.

The pack ships to the site the day before mobilisation. The first person on site plugs it in, connects to the named Wi-Fi, signs in with their existing work account, and within an hour the site is operational.

Behind the scenes, the router phones home to a central management platform. The IT team can see every site, push configuration changes, and notice when connectivity degrades.

Most construction sites can't wait for a BT install. 4G and 5G have closed most of the gap, and a dual-SIM router using two different carriers (typically one EE and one Vodafone or Three) covers the vast majority of UK sites with usable bandwidth.

Where the coverage is genuinely bad, a roof-mounted directional antenna or a Starlink Mini fills the gap. Starlink in particular has become a useful escalation: when nothing else works, it works. The unit cost is small compared to a day of site delay.

Site teams are mobile-first. Phones and tablets do most of the work, with the occasional laptop. Intune-managed devices with conditional access, Autopilot for laptop enrolment, and app protection policies for phones are the baseline. A new site engineer should be productive without anyone shipping a laptop.

Document control belongs in SharePoint or the project's CDE, accessed from whatever device the engineer happens to be holding. The days of 'come back to the cabin to print the drawing' are mostly over, but the IT setup has to actually support that.

What happens when a site closes? In a mature setup, the kit comes back to a depot, gets wiped, refreshed and pre-staged for the next mobilisation. Documents are archived or transferred to the next phase. Accounts for subcontractors who only worked on that project are tidied up.

In an immature setup, the router stays plugged in for another six months because nobody is responsible for retrieving it, and the SIM continues to be billed. The aggregate cost of unmanaged demobilisation is surprisingly large across an estate of sites.

Connectivity hardware, device enrolment, and the first-day SharePoint or CDE folder structure. Three things, decided once, repeated every time.

Connectivity: one router model, one SIM strategy, one management platform. Devices: one enrolment process, one MDM policy, one set of apps. Document control: a template site or CDE project that gets cloned per new project, with the same metadata, retention and permissions every time.

A new site is operational on day one, including connectivity, devices, accounts and the document control structure. Mobilisation is a known sequence of steps, not a series of small dramas. Site managers stop ringing IT in the first week because there's nothing to ring about.

The firms that get there have usually had a couple of bad mobilisations and decided not to repeat them. The investment is modest. The payoff is every subsequent site.

Every construction project pulls in sub-contractors, designers and consultants who need access to project information for a finite period. Treating them like permanent staff produces a mess at demobilisation. Treating them as if they don't exist produces shadow IT.

External user invites in Entra ID, scoped to specific SharePoint sites or CDE projects, with an automatic expiry, are the right pattern. Set the expiry at instruction; extend it if the engagement extends; let it lapse otherwise. The principle is to make 'they shouldn't have access any more' the default state rather than a manual cleanup task.