Opening a new location
IT Setup Checklist for Opening a New Business Location
In short
Opening a new business location needs eight things in place before it can trade: internet connectivity, site network, end-user devices, identity and email, physical security, customer-facing technology, compliance, and a defined support model. Of these, internet connectivity is the only one that should be ordered out of sequence with the fit-out, because circuit lead times are the longest single dependency and routinely run longer than the build itself.
This applies whether you are a growing business opening a second site, a franchise operator adding a location, a company relocating, or an international brand standing up a first presence in a new country. The technical work is largely the same. What changes is who owns the decisions and how far away they are.
Most of it is not difficult. It fails on sequencing, on the boundary between whoever owns configuration and whoever is on site, and on a small number of decisions that become expensive once the lease is signed.
The checklist#
- Order the internet circuit the day the lease is executed
- Order a secondary service on a different provider and a different physical path
- Confirm roof, riser and demarcation access during lease negotiation, not after
- Take the network pattern from your existing smallest site
- Source end-user devices locally, configure them remotely
- Confirm whether the new site's identity depends on a larger internal programme
- Define delegated permissions as actions, not as services
- Specify physical security and customer-facing technology against the fit-out programme
- Complete any compliance work with counsel before opening
- Agree the support and escalation model before anyone needs it
1. Connectivity is a schedule decision, not a procurement decision#
Quoted installation timelines for the same class of business internet service vary from two weeks to six months depending on the address and the provider. Where a carrier requires construction, a quote of 90 to 120 business days is four to six calendar months, which means a service ordered in early September may not be live until the following February.
Three consequences follow.
Order on lease execution day. At those lead times, a two week delay in placing the order becomes a two week delay in the site being able to trade. This is one of the very few items in a location launch that should jump the fit-out queue entirely.
Fixed wireless is your date insurance. Where fibre cannot be delivered before opening, fixed wireless is typically available within two to three weeks at any address, independently of which site is chosen. The workable pattern is to order fixed wireless as the primary service and place the fibre order in parallel. The site opens on wireless and cuts over when fibre arrives.
Check for municipal fibre. More cities operate publicly owned fibre networks than most businesses realise, and they do not appear in standard carrier comparisons. Where one exists, pricing can be less than half the cheapest commercial carrier at any speed, because the infrastructure was built for economic development rather than to recover a carrier's build cost. Two cautions: confirm the address actually falls inside the municipal boundary, and check whether construction cost is absorbed by the provider or recovered through the contract term. Where it is recovered, the quoted monthly figure is a floor rather than a price.
Speed, service levels and redundancy
Size against your existing estate rather than against the brochure. If your other sites run a 100/40 service, a 500 Mbps symmetric connection gives roughly five times the download and twelve times the upload, which is ample for six to ten staff with guest access, displays and video conferencing. Moving to a gigabit typically adds $125 to $200 a month for capacity a small site will not use, unless a municipal provider makes it cheaper than everyone's 500 Mbps.
Ask for written service levels at quote stage. In a recent five-address comparison across four providers, only one supplied written uptime, latency and repair commitments. For a site with no local IT staff, the mean time to repair matters more than the speed tier.
Take the secondary service, and do not let it share a provider or a last mile with the primary. The cheapest backup is almost always offered by the incumbent, which puts both services on the same physical path. Expect $50 to $150 a month for a genuine secondary.
One argument worth making internally: if your existing sites do not run redundant circuits, a new site may still warrant one, for reasons specific to it. If it takes payment on the floor, if it is your only presence in that market, or if technical support is a long way away, a same-day fix on a failed circuit becomes unlikely and the second circuit pays for itself the first time it is needed.
2. Some IT requirements are actually property requirements#
Fixed wireless services generally use a non-penetrative rooftop radio, which requires roof access and demarcation access. Those are landlord permissions. So are riser access and any cabling route through common parts.
Raise all of them during lease negotiation. After signature they become expensive, slow, or refused.
3. Take the network pattern from your existing estate#
A new site should default to the pattern your smallest existing site already runs, and justify every deviation from it. In practice that usually means a security appliance, a small number of access points, two wireless networks separating staff and guest traffic, and possibly no local switch.
The reason is not technical conservatism. It is that the estate stays supportable by the team that already supports the rest of it, and the deviations you do make are visible and deliberate.
If this is your first site, the equivalent move is to choose a pattern you can repeat, because the second site is considerably cheaper when it is a copy of the first.
4. Source devices locally, configure them remotely#
The instinct is to ship a known-good build from head office. Local sourcing with remote configuration achieves the same standardisation without shipping lead time, and across a border it also avoids customs exposure and warranty complications.
Size the order after recruitment completes, not before. Device procurement sized against a headcount that later changes is one of the more common avoidable costs in a launch.
5. Read the adjacent programme, not just the scope of work#
Scopes of work are usually written before the organisation's own internal projects are fully visible, and divergence between the two is common.
A pattern worth checking for. A scope of work described a new site's email as a child domain of the existing environment. The organisation's parallel identity programme was in fact migrating the entire business to a new primary domain, with the original retained as an alias. That is not an addition for one site, it is a global change, and it meant the new site's identity depended on a much larger programme with its own timeline and its own risk.
Finding a divergence like that early is worth considerably more than delivering the letter of the scope.
6. Define permissions as actions, not as services#
A support responsibility matrix that names which services each party touches looks like agreement. It is not one, until it says which actions each party may perform.
This is the most common place a managed service arrangement discovers, late and under pressure, that it cannot do what it promised. It also gates pricing, because a partner cannot commit to a service level for work it may not be permitted to carry out.
Where the site is in a different timezone from whoever owns the systems, the related design question is which routine tasks cannot afford to wait a sleep cycle. An administration model that lets on-site staff perform routine work under delegated permissions, without escalating overnight, is usually the difference between a support model that functions and one that loses a day on trivial requests.
7. Check the plan for reversed dependencies#
Worth a dedicated pass, because reversed dependencies never show as overdue and therefore never show as red.
Two seen on a recent launch plan: an insurance controls pack scheduled for delivery after the insurance was to be bound, and website content testing scheduled to finish after the website launched. Neither was late. Both were impossible in the order given.
The other register discipline worth adopting is separating who routes a question from who can answer it. Most dependency registers have a single owner column, which conflates the person chasing an answer with the person able to give one. Launches with several parties involved fail at the routing layer more often than at the technical layer.
8. State the boundary in one sentence#
Every arrangement involving more than one party needs a sentence a new person can read and correctly infer who does what. On a typical launch it reads something like: head office and its existing provider decide how things are configured, and the local partner makes them work on site, during local hours, and supports the people using them.
If it takes a paragraph, the boundary is not agreed yet.
Common questions#
How long does IT for a new location take end to end? The IT work itself is a matter of weeks. The binding constraint is the internet circuit, which can run two weeks or six months depending on address and provider, and the lease, because almost nothing site-specific can start before it is executed.
Can our existing IT team handle it instead of a local partner? For configuration and architecture, usually yes, and they should retain that ownership. What they cannot provide remotely is physical installation, device deployment, coordination with local suppliers, and support cover during the site's business hours.
What does IT for a new location cost? Connectivity runs roughly $400 to $1,200 a month for a dedicated business service, plus $50 to $150 for a secondary. Hardware and installation depend on site size. Managed support is typically structured as a one-off setup fee, a fixed monthly retainer, hourly on-site above an included baseline, and a per-seat charge above an included headcount.
Do we need a redundant circuit? If the site takes payment on the floor and technical support is not local, yes. The cost is small relative to a day of lost trading with no engineer available.
What is usually missed? Roof and demarcation access in the lease, reversed dependencies in the programme plan, and delegated permissions defined by service rather than by action.
Is this different for an international expansion? The technical work is largely the same. What differs is the distance between the site and whoever owns the systems, which makes the support model and the delegated permissions matter considerably more, and the addition of local compliance, payroll and tax requirements that do not carry over from the home market.