Process metrics and KPIs explained
Table of contents
Workflow metrics measure how work moves through a process: how long requests take, how much work gets completed, and how often tasks need correction. A key performance indicator (KPI) is a measure selected to track an important business objective. Choose the objective first. Then define the calculation, target, owner, and review period.
What is the difference between a process metric and a KPI?
A process metric describes performance. A KPI identifies which measure matters most for a particular objective. For example, a team might track request volume, waiting time, and rework, but use on-time completion as its KPI for improving service reliability.
APQC’s process measurement guidance(opens in a new tab) recommends selecting measures that fit the process and its strategic purpose. A target and deadline belong to the improvement goal; a metric can still be measured before either is set.
Which workflow metrics should you track?
Start with measures that help the process owner make a decision. The definitions below are examples you can adapt. Record the reporting period, eligible requests, and calculation rules so the numbers remain comparable.
| Metric | Example calculation or definition | Decision it supports |
|---|---|---|
| Cycle time | Completion timestamp minus submission timestamp for each request | Where to investigate slow delivery |
| Throughput | Number of requests completed in a defined period | Whether completed work is keeping pace with incoming work |
| Rework rate | Completed requests that required correction divided by all completed requests, multiplied by 100 | Whether forms or instructions need improvement |
| On-time completion rate | Requests completed by their due dates divided by completed requests with due dates, multiplied by 100 | Whether the workflow meets its service target |
| Backlog | Number of open requests at a specified point in time | Where additional capacity or prioritization may be needed |
| Cost per completed request | Allocated process cost divided by requests completed in the period | Whether an improvement reduces operating cost |
Cycle time includes waiting as well as active work; APQC defines it as elapsed time from the beginning to the end of a process(opens in a new tab). Decide whether to report calendar or business hours, and keep that choice consistent.
Report a median or percentile alongside the average when a few unusually slow requests could hide the typical experience. For rate measures, state the denominator explicitly. A rate based only on completed requests won’t show overdue requests that are still open, so track those separately.
How do you turn a metric into an improvement goal?
Attach a target, deadline, and accountable owner to a measure tied to the desired outcome. Here’s an illustrative goal, not a customer result: Reduce median invoice approval time from five business days to three by the end of the next quarter while keeping the rework rate at or below its current level.
Record the baseline before streamlining the workflow. Compare equivalent request types after the change so a shift in workload doesn’t look like a performance improvement.
Who will use these process metrics?
Before a report is created or a KPI is decided on, it’s important to ask “Who will use this information, and what action will be taken based on it?” This can reduce “report creep” drastically, as information is bucketed into categories such as “actionable” and “nice to know.”
To do this, start by determining who needs the information:
- Process owners
- Executives
- Process stakeholders
- Process users
Each group might need different information presented in different ways to make it more effective. Once you decide who has access to specific reports and dashboards, agree on how the information will be used.
What you can measure
- Efficiency/process cycle-time improvement
- Effectiveness/increased customer satisfaction
- Capacity/producing more
- Productivity/cost reduction
- Quality
- Profitability/revenue improvement
Establish tolerance levels
You’ll need to establish the tolerance level of your KPIs. In other words, ask yourself: “What values outside the expected range are critical, requiring corrective action?” These are often displayed as green, yellow, and red zones in a gauge. They might also be called RAG (red, amber, green) statuses.
Discover ready-to-use workflow automation templates that streamline your core processes with end-to-end automation.
Set thresholds according to the direction of improvement. Lower cycle time is usually preferable, while a higher on-time completion rate is usually preferable. Applying the same percentage bands to both can reverse the meaning of a dashboard’s colors.
For an illustrative response-time target of four business hours, a team could mark responses within four hours as green, more than four through five hours as amber, and more than five hours as red. These are example thresholds, not an industry benchmark. Agree on thresholds using service commitments, observed variation, and the consequences of a delay.
Assign someone to investigate when a threshold is crossed, and agree on the action they can take and when the team will review whether it worked.
What are leading and lagging indicators?
A leading indicator can signal a future outcome. For example, a growing queue of unresolved requests may signal longer completion times ahead. That relationship needs to be checked against the workflow’s actual results.
A lagging indicator describes an outcome that has already occurred, such as last month’s on-time completion rate. Use the two together: Queue size helps identify work that may need attention, while completion rates show whether the service objective was met.
Be SMART when you choose metrics
When using any methodology for defining metrics, keep the popular SMART model in mind:
- Goals should be specific and tied to business goals.
- Goals should be meaningful and measurable. The underlying data elements must be captured accurately and completely, and the calculation must be correct. If you can’t measure it or report on it from your current enterprise systems, consider the tradeoff in business value vs. manual effort, or report development before including it in your list of metrics.
- The target values for your metric must be realistically achievable. Do the right people have the ability and authority to drive the metric toward the target value?
- Metrics should be realistic, relevant, and results-oriented. Are there identified actions that can drive the metric toward its target value?
- Reports need to be timely, especially if they’re leading indicator metrics. Can reports be generated, or can dashboards be updated in time to allow appropriate interventions to drive the metric toward its goal?
More process KPI examples
- Average process time
- Average time per task
- Average time to complete a task
- Number accepted/rejected
- Number in queue in stages
- Number of complaints
- Number of errors
- Number of escalations
- Number of late tasks
- Number of requests
- Number of tasks sent for rework
- Percentage of overdue tasks
- Percentage of processes below/above goal
- Requested vs. budgeted
- Requests by type
- Time from request to fulfillment
- The volume of tasks per staff
Conclusion
Choose a small set of measures that lead to a decision. Define each calculation, establish a baseline, and assign an owner who can act on the result. Review speed alongside quality so a faster workflow doesn’t simply pass more errors to the next team.
