KPI and SLA Definition: Commitments vs Measures

KPI and SLA definitions differ: an SLA states a service commitment, while a KPI measures performance against an important objective. SLA attainment can be a KPI, but a KPI such as average resolution time does not become a customer promise simply because it appears on a report.

The distinction matters when a support team reports good average response times while some customers receive service outside their agreed limits. Start with the ticket rows and the applicable agreement, then calculate the performance measures and commitment results separately.

What each one tells you

A service level agreement defines the service expected between a provider and a customer. It can specify response and resolution targets, responsibilities, measurement rules and consequences if expectations are missed; these are elements described in the service agreement guide.

A key performance indicator is a measure selected because it matters to an objective. The service team might track average response time to manage workload, resolution time to identify difficult issue types, and SLA attainment to monitor customer commitments.

An internal target is useful, but label it as internal. Reporting it as a contractual SLA can misstate what customers were promised. For the broader service scorecard, use the B2B customer service KPI guide.

Bring the ticket rows and the terms together

Use a support export with one record per ticket, including ticket ID, account ID, priority, creation time, first qualifying response, latest resolution and current status. Add the agreement version, applicable calendar, response limit and resolution limit from the contract register.

The agreement must match the account, service, priority and effective date. Keep the original timestamps alongside any approved elapsed-time calculations. An agreement measured in business hours needs the correct service calendar; subtracting two timestamps gives calendar elapsed time instead.

Before calculation, settle whether the agreement measures response and resolution independently or requires both on the same case. Also settle which tickets enter the reporting period and how overdue unresolved cases are handled. Those choices can change attainment without changing a single response.

The same six cases measured both ways

Consider a US business-services company's illustrative review. Every case below is eligible and has a completed response and resolution measurement. All times use calendar elapsed time; the limits are example contract terms, not industry benchmarks.

First calculate the operating KPIs. An arithmetic mean adds the measurements and divides by their count, as the AVERAGE function documentation explains.

Average response time = Total response minutes / Responded cases

Average resolution time = Total resolution hours / Resolved cases

SLA attainment = Eligible cases meeting the required conditions / Eligible cases

Percentages are rounded to one decimal place. The means describe the experience of these cases. The attainment figures compare each case with its own limits. Comparing the overall mean with a single convenient limit would erase the different commitments.

If the illustrative agreement requires both-target attainment of at least 90.0% for this review, the measured gap is 90.0% minus 50.0%, or 40.0 percentage points. The report identifies that shortfall; any remedy still follows the agreement's actual wording.

Prove the attainment result before reporting it

The case-level checks should reconcile to the summary. Three cases meet both targets and three fail at least one, giving all six eligible cases. Five pass response and one fails it. Four pass resolution and two fail it.

Do not derive both-target attainment by averaging the separate response and resolution percentages. A case must satisfy the required conditions together. Keep the case flags so the service owner can inspect exactly which condition each failed.

For account-specific terms and the fuller contract comparison, follow SLA attainment per account. Carry the definition, source period and accountable owner into the KPI reporting process, so next month's result uses the same rules.

Where it goes wrong

• A good mean presented as compliance. An average cannot show which individual cases missed their own limits. Report the mean and attainment separately.
• One default limit applied to every customer. Match each case to the agreement effective for that account and service. Keep unmatched cases visible for review.
• Calendar hours mixed with business hours. Use the contractual clock, including its holidays and approved pauses. Record the clock in the report definition.
• Unresolved cases dropped from a due-case measure. A completed-only export can omit cases that are already overdue. Reconcile eligible cases to the required reporting population.
• Every miss described as a financial penalty. A performance shortfall does not establish a credit amount. The commercial owner must apply the relevant remedy conditions.

Covirage is not affiliated with Atlassian or Microsoft.

Compare your own service results in Covirage

Bring a ticket export and the applicable service terms: Covirage maps the columns, and its tools compute and check the response, resolution and attainment figures. The external AI model explains the comparisons with every figure traced to its rows. Create a free Covirage account and bring a sample of your own data; creating the account takes about a minute.

See metrics governance for the supporting approach to keeping definitions consistent.