Service levels and remedies for poor performance

A service level should measure an outcome that matters to the customer. A response target does not necessarily promise a resolution, and an availability percentage may exclude the hours when the service is most important. Read definitions, measurement methods and exclusions alongside the headline commitment.

Decide whether service credits are the sole remedy or one part of a wider response. Check the claim process and time limit for requesting a credit. Repeated poor performance may need escalation, remediation or an exit route even when individual credits have been issued.

Define the service boundary before choosing a percentage

Identify the systems, locations, users and hours included in the commitment. A service described as available all day may exclude planned maintenance, third-party networks or customer equipment. Those exclusions can remove periods that matter commercially. Ask the supplier to show how the measurement works during the customer's busiest operating hours, using the actual architecture and support arrangement.

Separate availability from successful business processing. A website can respond while checkout fails, and a helpdesk can answer while the underlying problem remains unresolved. Select measures that reflect the outcome the customer needs. If several suppliers contribute to delivery, decide who investigates incidents and how responsibility is assessed when the source of failure is initially unclear.

Make severity levels usable during an incident

Describe critical, major and lower-priority incidents in operational terms. Identify who can raise a priority case and what information is required. A severity definition that depends on every user being affected may miss a failure that stops the only team handling orders. Test the categories against realistic incidents rather than adopting labels with no shared meaning.

Distinguish acknowledgement, diagnosis, temporary restoration and permanent correction. Each can need a different target and update frequency. Agree how a priority classification is disputed or changed, and preserve the incident record. The team should know what the supplier must do next rather than repeatedly asking whether an automated ticket response counts as satisfactory performance.

Agree measurement and exclusions with evidence

Specify the clock start, permitted pauses and source of performance data. If customer delay pauses a target, require a clear request identifying the missing input. An indefinite pause caused by a vague request for more information can make the target difficult to enforce or even measure. Keep a record of when the necessary response was provided.

Review planned maintenance notice, frequency and permitted duration. A maintenance exception should not silently cover repeated emergency outages. Agree how disputed measurements are checked and how reports are retained. For an outsourced support service, the customer may need access to ticket history after termination to understand recurring faults and manage the replacement provider.

Design remedies for both isolated and repeated failure

A service credit can provide a defined response to a missed target, but it may be small compared with disruption. Check its calculation, claim deadline and any monthly ceiling. Decide whether credits are automatic or must be requested and which evidence is needed. A remedy that the team never claims because the process is unclear offers little practical value.

Read any exclusive-remedy language and liability restrictions together. Applicable business exclusions may require assessment under the Unfair Contract Terms Act. ^[1] Repeated failure may justify escalation, a corrective plan or an agreed exit route beyond individual credits. Define the trigger and process so the customer does not have to renegotiate the response after every incident.

Connect performance management with payment and handover

Clarify how credits affect invoices and whether disputed performance permits withholding any amount. Do not assume a missed target automatically removes the obligation to pay every fee; the contract and applicable late-payment framework need separate consideration. ^[2] Give finance the agreed credit process so it does not make deductions that the operational team cannot support.

Assign a service owner to review reports and recurring failures. Retain the incident evidence, agreed corrective actions and outcome. A quarterly report should lead to a decision where service is deteriorating, not simply be filed away. Read Reviewing a supplier contract before signing for the wider supplier review. Supply agreement drafting support can help align service levels, remedies, payment and exit provisions with the actual dependency on the supplier.

Reconcile the supplier's report with user experience

Review a sample incident against the report rather than relying only on the monthly headline percentage. Check the start time, severity, pauses and closure reason. If the service was technically restored but users could not complete essential work, identify whether the agreed metric captures that outcome. The review may reveal a measurement gap rather than deliberate misreporting. Use the evidence to correct definitions or operational procedures through the agreed change process. Keep the resulting decision with the service record so a later account manager understands why the measurement changed and which incidents should be assessed under the revised definition.

Illustrative scenario

A support contract promises a one-hour response. The supplier sends an acknowledgement within minutes but takes days to restore the system. The customer separates acknowledgement, diagnosis and restoration targets, identifies severity levels and agrees evidence for each. The review then checks how planned maintenance and customer-caused delays affect those targets.

Preparation checklist

• Define the business hours and systems covered.
• Separate response, restoration and permanent resolution.
• Record exclusions and the evidence used to measure them.
• Check credits, escalation and repeated failure provisions together.