The process of negotiating commercial agreements is at least as important as the end product documentation. By working through a comprehensive list of outstanding issues the parties will, in almost all cases, shed light on previously unexamined aspects of their nascent relationship and, in so doing, will enhance the probability of its success. Equally, it seems fair to say that in most cases the negotiated agreement will simply “sit on the shelf”; however, to the extent that a dispute does arise, it frequently will have been caused by the parties’ failure to properly define their respective performance obligations.

In the particular context of “traditional” software as a service models (leaving aside particular issues associated with cloud computing), via which data may be dispersed across and stored in multiple data centers all over the world), it is common practice for the parties to in effect define the services to be provided at a granular level via a Service Level Agreement (“SLA”) attachment to the overall Services Agreement, although the length and complexity of the SLA will be varied to fit the circumstances.

By defining a suite of both Critical Performance Indicators (CPIs) and Key Performance Indicators (KPIs), the former having associated fee claw-backs, the customer will also cause the service provider to put “skin in the game” as a means for incenting it to provide the agreed upon quality of service(s) – the maximum credit available against the upcoming monthly fee(s) is almost always defined as a percentage of such fee(s). The fee credit / refund regime is generally positioned as a sole remedy in the event of such relatively modest failures.

Additionally, by detailing in the SLA the point at which very poor service will be considered to be a material breach of the Services Agreement, the parties will have agreed upon the threshold beyond which the customer will be entitled to elect to seek additional remedies up to and including termination and recovery of its damages arising from such breach – such damages amount usually mediated by reference to the agreed upon cap on direct damages set out in the limitation of liability provision.


At Risk Amount / Percentage of Fees: a defined percentage of monthly service fees paid by the customer to the service provider.  This defined number (usually in the 10% to 12% range, but can go all the way up to 50%) is used in calculating the amount of the refund/credit the customer will be entitled to if the service provider fails to meet a CPI.

Critical Performance Indicators (“CPIs”): performance obligations with minimum service levels attached.  If these targets are not met, the customer will have remedies against the service provider.

Increased Impact Service Level Default: where the provider fails to achieve the service level multiple times during a specified time period, it may be liable to increased penalties (e.g., failures in two consecutive months, or failures in any four months during a twelve-month period will double the severity weight used in calculating credits owing to the customer).

Key Performance Indicators (“KPIs”): KPIs are items that are monitored using a range of quantitative and qualitative metrics but a failure to achieve a KPI will not trigger the remedies associated with a breach of a CPI.  The contract may provide the customer with an option to convert a KPI to a CPI, exercisable, for example, twice per year.

Minimum Service Level: the minimum quantitative level or degree of performance that must be met to satisfy a CPI.  In complex transactions there will generally be a suite of CPIs, each with a specific target level (e.g., 99.90% for one, 97% for another etc.).

Service Level Credit: credit available to the customer should the service provider fail to meet minimum service levels for a CPI.

Service Level Default: a defined state where the provider has fallen short of meeting a CPI.  These events of default are used to calculate credits owing to the customer and may count towards enabling a trigger a termination right.

Severity Weight: the failure to meet a particular CPI may be more, or less serious to the customer.  In more complex SLAs, each CPI is assigned an individual severity weight to account for its unique impact to the customer.  The total will generally exceed 100%. For example:

Click here to see table.


Common service levels found in Hosted Services Agreements

For very small-scale projects, all services to be performed under the contract may be defined by a single CPI:

e.g.: "Services will be available to the customer for not less than 99.75% of the time on a 24 hours a day, 7 days a week basis each calendar month, minus any Excused Down Time Events."

More often, a hosted services provider will offer the type of service levels described below. The example below distinguishes between managed services and mere co-location services.

Managed services

Server Hardware & Infrastructure

  • 99.7% Uptime SLA.
  • Compliant when devices and components (physical hardware for servers including firewalls, load balancers, VPNs and switches) are powered on and functioning properly, meaning performance of all essential functions including remote access.

Managed Operating Systems

  • 99.7% Uptime SLA.
  • Compliant when the OS is working properly meaning the computer has network access, can ability to boot, login and run all critical services and/or daemons. Network Availability

­Managed Applications

  • 99.7% Uptime SLA.
  • ­Compliant when the administrator can login to the application, all related services / daemons start, the application(s) have external network connectivity. 

Co-location services

Facility Power & Environmental Control

  • ­100% Uptime SLA.
  • ­Compliant when datacentre power and environmental control systems are operational.

Network Availability

-      100% Uptime SLA. ­    -      Compliant when the network connections for the cage, cabinet, half cabinet or managed environment are able to connect to the data centre. 

Single or multiple service levels For more complex arrangements, the customer will bear the negotiating burden of obtaining a more fulsome CPI regime. Below is an example of the kind of detail one might find in such arrangements.

Click here to see table.

Calculation of service credits

If the provider fails to meet specified service levels, the SLA will describe a formula for determining the credit amount owing to the customer.  The SLA may also provide for increasing credits according to the number of times per given time period (e.g., one month) that the services are not performed at an acceptable level.  Credits are set by a formula that computes variables such as time, severity of the failure and a percentage of fees paid.