Measuring value: KPIs, CSFs and OKRs

If value is perceived and co-created, how can anyone measure it? There is no single number for value, but organizations can measure the outcomes, costs, risks and experiences that make it up, and use those measures to make better decisions. This lesson introduces the basic vocabulary of measurement in service management: metrics, critical success factors (CSFs), key performance indicators (KPIs) and objectives and key results (OKRs). It explains the difference between leading and lagging measures, warns about common traps such as the "watermelon" service level agreement, and works through simple calculations. Later courses go deeper into measurement and reporting; this lesson gives the foundations.

Why measure?

Measurement supports four purposes:

• Validate: confirm that decisions and services are achieving what was intended.
• Direct: set targets and show people what matters.
• Justify: provide evidence for investment, change or continuation.
• Intervene: spot problems early and act before they grow.

Measurement is not free. Collecting, analyzing and reporting data takes effort, and every measure risks distorting behavior if people focus on the number instead of the purpose behind it. The aim is a small set of meaningful measures linked to value, not a dashboard of everything that can be counted.

Metrics, CSFs and KPIs

A metric is a measurement or calculation that is monitored or reported for management and improvement. Examples include the number of incidents per month, average response time and cost per transaction.

A critical success factor (CSF) is a necessary precondition for achieving intended results. It describes what must go right. For an online learning service, CSFs might be "learners can access courses whenever they need to" and "learners find the content useful".

A key performance indicator (KPI) is an important metric used to evaluate the success in meeting an objective, and is usually linked to a CSF. For the CSF "learners can access courses whenever they need to", a KPI could be "availability of at least 99.5 % during agreed hours". For "learners find the content useful", a KPI could be "average lesson rating of at least 4.2 out of 5".

Leading and lagging measures

Lagging measures report results after the fact: last quarter's availability, customer satisfaction, cost per user. Leading measures give early warning of future results: the number of unpatched servers, rising queue lengths at the service desk, a growing backlog of changes. A balanced set includes both. Lagging measures tell you whether you achieved your objectives; leading measures help you act in time.

Objectives and key results

Many organizations now use objectives and key results (OKRs) to set direction. An objective is a qualitative, inspiring statement of what you want to achieve. Key results are a few measurable results that show whether you achieved it. OKRs are usually set for a quarter or a year and reviewed often. They differ from KPIs in emphasis: KPIs monitor ongoing health; OKRs drive change toward a new level.

Service levels and the watermelon effect

Service level agreements (SLAs) record agreed targets for services, often for warranty measures such as availability and response times. A common problem is the watermelon SLA: green on the outside, red on the inside. Every target in the report is met, yet consumers are unhappy because the targets measure the wrong things, or measure them in a way that hides real problems. For example, an SLA may require incidents to be responded to within 15 minutes, which is met by sending an automatic acknowledgment, while the actual fix takes days.

To avoid watermelons, organizations:

• Link SLA targets to outcomes consumers care about, not only provider activity.
• Measure end to end, from the consumer's point of view, rather than component by component.
• Include experience measures alongside warranty measures. Some organizations use experience level agreements (XLAs) that set targets for experience, such as effort or sentiment.
• Review SLAs regularly with consumers, as part of relationship management.

Balancing measures

Measures can pull against each other. Pushing first-contact resolution up may lengthen call times. Cutting cost per ticket may reduce quality. A balanced set of measures across outcomes, costs, risks and experience helps prevent one improvement causing harm elsewhere. A simple weighted score can combine several KPIs into an overall view, as long as the underlying measures remain visible.

A well-known warning, often called Goodhart's law, says that when a measure becomes a target, it ceases to be a good measure. If agents are rewarded only for closing tickets quickly, some will close tickets before the problem is solved. Pair targets with checks on quality, and talk about the purpose behind each measure.

Reporting measures well

Measures only create value when people use them to decide something. Good reports are short, focused on the audience and tell a story: what happened, why it matters and what will be done. A service owner's monthly report to a customer might show three or four KPIs against targets, a trend over the last six months, a note on any target missed with its cause and fix, and one or two improvement actions. Long tables of numbers without interpretation are rarely read.

Trends are often more informative than single values. A satisfaction score of 82 % means little on its own; 82 % after six months of decline from 90 % tells a different story from 82 % after rising from 70 %. Showing targets, trends and short explanations together helps readers see whether action is needed.

Be careful to distinguish percentage points from percentages. A rise from 65 % to 70 % is 5 percentage points, but a 7.7 % relative increase. Reports that mix the two can mislead.

Finally, measurement should evolve. As services and objectives change, some measures stop being useful and new ones are needed. Reviewing the set of KPIs with consumers at least once a year, removing measures nobody uses and adding measures for new outcomes, keeps measurement focused on value rather than habit. A short list that people discuss and act on is worth far more than a long list that is collected and ignored. When a measure is retired, record why, so that the reasoning is not lost when people change roles.

It is also worth checking how each measure is calculated. Two teams can report the same KPI name, such as resolution time, using different start and stop points or excluding different cases. Agreeing clear definitions, writing them down and applying them consistently over time makes comparisons meaningful and prevents arguments about whose numbers are right. A short definition sheet for each KPI, stating its purpose, formula, data source and owner, is a simple way to do this.