Running one office network is a project. Running fifteen is an operating model, and most of the cost is not in the hardware. It is in the drive time, the spare switch in a closet at each location, the four sites configured slightly differently by four different people, and the fact that nobody can tell you what is happening at the branch until somebody there calls. This page covers where NaaS genuinely solves that, where it does not, which providers fit which multi-site shape, and how the per-site math actually adds up.
Usually yes, and the reason is specific: multi-site is where the operational cost of a network stops being proportional to the hardware and starts being proportional to the number of places you have to be. NaaS moves the hardware, the deployment, the monitoring, and the on-site response to a provider under one contract and one dashboard, which is exactly the cost that grows fastest as you add locations.
It is a poor fit in three cases. If you have one dominant campus and a couple of tiny satellites, the per-site economics rarely beat cloud-managed gear your team already runs. If your sites need genuinely different network designs, a standardized service model fights you rather than helping. And if you have deep in-house network engineering that is already covering all locations well, you are paying a provider to do something you do well, and the honest answer is to keep doing it yourself.
The rest of this page is about telling those situations apart. If you want the general background first, start with what NaaS is and the full NaaS guide.
An access point dies at a branch three hours away. Someone drives, or you pay a local contractor who has never seen your network, or the site limps until the next scheduled visit. Multiply that by every site and every year, and the travel and contractor line is often larger than the hardware line nobody questions.
Keeping a network up across many locations means keeping spares somewhere. Either you stock every site, which is capital sitting in closets, or you stock centrally and accept shipping delays during an outage. Provider-owned models make spares and replacement the provider's problem, which is one of the cleaner wins in the model.
Site 3 was set up in 2021, site 9 by a contractor, site 12 during an acquisition. VLANs, SSIDs, firewall rules, and firmware all diverge quietly. The drift is invisible until an incident or an audit, and then it is a project. Standardized deployment across sites is the point of the service model, not a bonus feature.
When the branch calls to say the internet is slow, the honest answer is often that you cannot see it from here. Cloud-managed platforms fix visibility; managed NaaS goes further and puts someone else on the hook for noticing before the call comes.
Opening a location and waiting weeks on circuits, gear, and an engineer visit puts the network on the critical path of a business decision. Providers that pre-stage hardware and bundle connectivity compress that, which matters most if you open sites on a predictable cadence.
Sites bought gear in different years, so hardware refresh arrives as a rolling series of unbudgeted capital requests rather than one planned event. Subscription models flatten that into an operating line, which is often the reason finance gets interested before IT does.
Multi-site is not one profile. What fits ten similar offices is different from what fits three campuses plus twenty small branches.
| Your footprint | Worth a close look | Why it fits | Watchouts |
|---|---|---|---|
| Many similar offices, lean IT team | Meter | Fully managed and billed per square foot, so cost scales with the footprint rather than with device counts you have to inventory. Provider owns the hardware, which removes both the spares problem and the refresh cycle across every location at once. | Less engineer-level control than a platform you operate. Newer entrant, so check reference sites at your scale. |
| Regulated or security-led multi-site | Nile | Zero-trust segmentation is built into the fabric rather than assembled per site, which is the difference between one security posture and fifteen slightly different ones. Performance SLAs on the Advanced tier give you something to hold across locations. | US-primary footprint, so international sites need checking. No hardware buyback at exit. |
| Existing cloud-managed estate you want to keep | Cisco Meraki | Central visibility across sites without changing operating model, and the skills are easy to hire for. The most common incumbent path when the real problem is visibility rather than staffing. | Still hardware CapEx plus per-device licenses at every site, and features stop when licenses lapse. You or a partner still do the operating. |
| Large campuses plus branch sites | Juniper Mist | Handles high-density campus and distributed branch in one platform, and Marvis AIOps is aimed squarely at troubleshooting sites you are not standing in. | Per-device subscription SKUs and real engineering depth to run well. Partner-led deployment adds timeline per site. |
| Security-driven, firewall-first across sites | Fortinet | SD-WAN, firewall, and switching under one policy stack, which is a genuine multi-site strength when site-to-site connectivity and policy consistency are the actual problem. | A security platform first, a network platform second. You or an MSSP operate it, so the staffing question does not go away. |
| Cost-sensitive, simple sites, skills in-house | Ubiquiti UniFi | No recurring license fees at any site and a free self-hosted controller, which is a material difference once you multiply licensing by location count. | Not a managed service and no SLA. Every truck roll and every spare is still yours, which is the cost multi-site buyers usually came here to solve. |
Each option links to its full independent review. For two providers weighed head to head, see the provider comparisons, or work through the evaluation framework.
Multi-site pricing is where NaaS quotes get misread most often, because the headline rate is only part of it. The common structures, and the ranges we see, are per square foot at roughly $0.08 to $0.25 per sq ft per month, per user at roughly $10 to $80 per user per month, and per access point at roughly $15 to $120 per AP per month. On top of any of those, multi-location deals frequently add a per-site base fee, commonly $0 to $1,000 per site per month. These are estimate ranges, not quotes; your actual number depends on density, service tier, security scope, and term.
That base fee is the line that decides multi-site economics. A per-square-foot rate looks excellent until you apply a base fee to twenty small branches, at which point a footprint-based model can cost more than a per-AP model for the same estate. Run both structures against your own site list before you form an opinion about which provider is expensive.
The second thing to model is the cost you are currently not counting. Travel and contractor time for site visits, spares sitting in closets, the staggered refresh across locations, and the engineering hours spent reconciling configurations are all real costs that sit outside the hardware line. They are also the costs a managed model actually removes, so leaving them out of the comparison makes NaaS look worse than it is.
Model your own footprint with the NaaS pricing calculator, then read the NaaS pricing guide for how each model behaves as you add sites. For the buy-versus-subscribe question specifically, the NaaS vs SD-WAN comparison is a useful companion if connectivity between sites is part of the scope.
Ask every provider on the shortlist the same questions, and get the site-level answers in writing rather than the corporate ones.
The longer, provider-agnostic version is the NaaS evaluation checklist.
NaaSAdvisor helps multi-site buyers compare providers, pricing structures, and per-site terms side by side, free and vendor-neutral. We can model a footprint-based quote against a per-AP quote for your actual site list, including the base fees that decide which one wins. Bring your locations and we will help you shortlist without the sales pressure.
An independent advisor competes every relevant provider across your full site list, base fees included, at no cost to you.