A connectivity failure at 10:30 on a Monday is not just an IT issue. It can stop cloud applications, card payments, customer calls and access to shared files across every site. Business internet SLAs set out what a provider is committed to deliver when that happens - but the detail behind the headline matters as much as the uptime percentage.
What is a business internet SLA?
A service level agreement is the contractual commitment attached to a connectivity service. It defines measurable standards for availability, support and fault restoration, along with the process for reporting an issue and any service credits available when the provider does not meet its target.
On a dedicated internet access service or leased line, the SLA is usually a significant part of the value. Unlike a consumer-grade or best-efforts broadband product, the connection is designed for business use with committed performance, defined support arrangements and a route to escalation.
That does not mean every business needs the most stringent SLA available. A small office with a reliable 4G or 5G backup may make a different commercial decision from a legal firm, healthcare provider or contact centre where even a short outage has an immediate operational cost. The right level depends on the impact of downtime, not simply the bandwidth requirement.
For organisations relying on cloud platforms, hosted telephony and multi-site networks, an SLA should provide clear operational accountability. It should explain what is covered, how quickly faults are handled, who owns the incident and what happens if agreed service levels are missed.
The SLA measures that deserve close attention
Availability is the figure most buyers see first. A 99.9% annual availability target permits considerably more downtime than 99.99%, even though the numbers appear close. At 99.9%, the allowance is roughly 8 hours and 46 minutes over a year. At 99.99%, it is about 53 minutes.
However, availability should never be assessed in isolation. Check how it is measured and which events are excluded. Planned maintenance windows, customer-site equipment faults, third-party damage and events outside the provider's reasonable control may not count against the target. These exclusions can be legitimate, but they need to be understood before signing.
Fault response and restoration are different commitments
A provider may promise to acknowledge a fault within a set period, such as 30 minutes, but that is not the same as restoring the service. Response time confirms that the incident has been accepted and is being investigated. Restoration time sets the target for returning the service to normal operation.
For a business-critical circuit, look for a clear restoration target, the support hours that apply and the escalation path if progress stalls. A 24/7 helpdesk is valuable when staff are working outside standard office hours, but it should be backed by ownership of the incident rather than an automated acknowledgement alone.
It is also worth asking when the restoration clock starts and stops. Does it begin when your team logs the fault, when the provider verifies it, or when an engineer attends? Can the clock be paused while access to the premises is arranged? Clear definitions prevent confusion during an already disruptive event.
Service scope must match the actual network
An SLA may cover the access circuit but not the router, firewall, internal LAN or Wi-Fi estate. If the issue sits beyond the provider's service boundary, the circuit can technically be meeting its SLA while users remain unable to work.
This is not necessarily a problem, provided responsibilities are explicit. Businesses with limited in-house IT resource often benefit from a managed service where the provider monitors the connection, manages the edge equipment and coordinates fault diagnosis. That reduces the risk of several suppliers passing responsibility between one another.
For multi-site organisations, the scope should also cover the wider design. An SLA on each individual circuit does not automatically guarantee application performance between locations or to cloud services. SD-WAN, traffic prioritisation and resilient routing may be needed where voice, business systems and large data transfers share the same network.
Why service credits are not the main protection
Service credits are a standard part of many business internet SLAs. They offer a reduction on monthly charges when a provider misses an agreed target. They are useful because they create contractual accountability, but they rarely compensate for the real cost of an outage.
Lost trading, delayed work, idle staff, missed calls and reputational damage can quickly exceed the value of a credit. Treat credits as a remedy of last resort, not as the continuity plan.
The stronger approach is to reduce the chance that a single fault stops the organisation. This may mean a secondary internet connection using a different network path, a 4G or 5G failover service, or dual diverse leased lines for sites with very low tolerance for disruption. The design should be proportionate: full physical diversity is not required for every office, but a critical site should not depend on a single point of failure without consciously accepting that risk.
Questions to ask before agreeing an SLA
Before comparing providers, establish what downtime means in practical terms for your organisation. Does a loss of connectivity prevent staff from using cloud applications? Will your hosted PBX fail over to mobiles? Can remote sites continue operating if the head office connection is unavailable? The answers will shape the required service and resilience level.
Then ask the provider to explain the SLA in operational language. Four areas are particularly useful to clarify:
Also examine the installation commitment. A strong in-life SLA is of limited comfort if the service delivery timescale is vague or dependent on unconfirmed civil works. For a relocation, new office opening or network refresh, ask which milestones are committed, what dependencies sit with the landlord or carrier, and how delays will be communicated.
- The availability target, measurement period and stated exclusions.
- The target for fault response, restoration and escalation.
- The precise service boundary, including responsibility for routers, firewalls and on-site equipment.
- The resilience options available if the primary circuit fails.
Choosing the right connectivity service
The underlying service affects the SLA that can reasonably be offered. Dedicated internet access and leased lines are generally suited to organisations that need uncontended bandwidth, predictable performance and stronger fault commitments. They are commonly used for cloud-first operations, multi-site networks and sites carrying large volumes of voice or transactional traffic.
FTTP can be an excellent option where availability, budget and operational risk align. It may provide substantial bandwidth at a lower cost, but the support model and restoration commitments can vary. It is essential to compare the actual contractual terms rather than assuming all full-fibre services offer the same protection.
For businesses with several locations, a combination often makes sense. A primary dedicated circuit at a head office, suitable full-fibre services at lower-risk sites and centrally managed backup connectivity can provide a sensible balance between cost and continuity. The goal is not to buy the highest specification everywhere. It is to ensure each site has connectivity appropriate to its role.
Business internet SLAs should support accountability
The best SLA is one your team can use confidently during an incident. It should identify one accountable provider contact, give your IT team access to meaningful updates and make it clear when an issue has been resolved rather than merely closed.
Ask how faults are communicated, whether proactive monitoring is included and how major incidents are escalated. In a multi-supplier environment, a provider that can take ownership of diagnosis across the connection and managed edge can save valuable time. Clear accountability is especially important where connectivity, voice, security and cloud services depend on one another.
A well-written SLA cannot prevent every outage. Fibre can be damaged, power can fail and third-party networks can experience disruption. What it can do is set realistic expectations and ensure the right response is triggered quickly when business operations are affected.
When reviewing business internet SLAs, focus less on the headline percentage and more on the complete service model: restoration commitments, exclusions, support coverage, network design and ownership during a fault. That is where dependable connectivity becomes a practical part of business continuity, rather than a promise buried in a contract.
