Project management KPIs are the handful of measures that tell you whether a project will deliver what it promised: on time, within budget, at the right quality and with the outcome the business wanted. The useful ones answer a specific question and lead to a decision. The rest are decoration.
Most teams do not suffer from too few numbers. They suffer from dashboards full of task counts, hours logged and percent-complete bars that look reassuring right up until a milestone slips. This guide covers the KPIs worth tracking across six areas (schedule, cost, flow, quality, capacity and outcomes), shows the exact formulas behind earned value, and gives you a worked dashboard example you can adapt.
If you are new to the discipline, start with our guide to what project management is, then come back here when you are ready to measure it.
What makes a KPI useful (and what makes it noise)
A metric becomes a key performance indicator when someone would act differently depending on its value. Before adding anything to a dashboard, check it against five tests:
- It answers a question a stakeholder actually asks. “Will we hit the March release?” is a question. “How many tasks closed this week?” usually is not.
- It has one agreed calculation. If two people can compute it differently, it will be argued about instead of acted on.
- It comes from data the team already produces. A KPI that needs a weekly manual spreadsheet will quietly stop being updated.
- It has an owner and an action threshold. Someone is responsible for noticing when it crosses a line, and the line is written down.
- It is hard to game without improving reality. Counting closed tickets rewards splitting tickets. Measuring cycle time on consistently sized work is harder to fake.
Leading indicators (forecasts, work item age, risk exposure) warn you while there is still time to act. Lagging indicators (final cost, escaped defects, adoption) confirm what happened. A good dashboard mixes both, with more weight on the leading ones during delivery.
Schedule KPIs: milestones and forecasts
Schedule health is the KPI area executives care about most, and the one most often reported badly. A count of completed tasks says little, because tasks differ in size and importance. Track these instead:
- Milestone forecast vs baseline. For each key milestone, show the baseline date, the current forecast and the variance in days. Update the forecast from evidence, not optimism.
- Milestone hit rate. The share of milestones delivered on or before their committed date over the last few months. Useful for spotting whether a team’s planning is systematically optimistic.
- Critical dependency status. Whether the handful of dependencies that sit on the critical path are on track, at risk or late.
- Schedule performance index (SPI), if you use earned value. Covered in detail below.
When a milestone is at risk, the KPI alone is not enough. Pair it with the cause, the dependencies affected, the recovery options and the decision needed. That turns a red cell into an agenda item.
Cost and effort KPIs: earned value, explained properly
Comparing actual spend with budget tells you how much money has gone out. It does not tell you whether you got the expected amount of work for it. Earned value management (EVM) solves that by comparing three numbers in the same unit, usually money. PMI defines them as follows:
- Planned value (PV): the budgeted cost of the work scheduled to be done by now.
- Earned value (EV): the budgeted value of the work actually completed, not hours logged or money spent.
- Actual cost (AC): what has actually been spent to complete that work.
- Budget at completion (BAC): the total approved budget.
Every earned value KPI is a combination of those four inputs:
Schedule variance (SV)
- Formula
- EV − PV
- How to read it
- Positive = ahead of plan; negative = behind (in money, not days)
Cost variance (CV)
- Formula
- EV − AC
- How to read it
- Positive = under budget; negative = over budget
Schedule performance index (SPI)
- Formula
- EV ÷ PV
- How to read it
- 1.0 = on pace; 0.90 = working at 90% of planned pace
Cost performance index (CPI)
- Formula
- EV ÷ AC
- How to read it
- 1.0 = on budget; 0.80 = 80 cents of value per dollar spent
Estimate at completion (EAC)
- Formula
- BAC ÷ CPI
- How to read it
- Forecast final cost if current cost efficiency continues
Variance at completion (VAC)
- Formula
- BAC − EAC
- How to read it
- Expected overrun (negative) or underrun at the finish
To-complete performance index (TCPI)
- Formula
- (BAC − EV) ÷ (BAC − AC)
- How to read it
- Efficiency needed from now on to finish on budget
| KPI | Formula | How to read it |
|---|---|---|
| Schedule variance (SV) | EV − PV | Positive = ahead of plan; negative = behind (in money, not days) |
| Cost variance (CV) | EV − AC | Positive = under budget; negative = over budget |
| Schedule performance index (SPI) | EV ÷ PV | 1.0 = on pace; 0.90 = working at 90% of planned pace |
| Cost performance index (CPI) | EV ÷ AC | 1.0 = on budget; 0.80 = 80 cents of value per dollar spent |
| Estimate at completion (EAC) | BAC ÷ CPI | Forecast final cost if current cost efficiency continues |
| Variance at completion (VAC) | BAC − EAC | Expected overrun (negative) or underrun at the finish |
| To-complete performance index (TCPI) | (BAC − EV) ÷ (BAC − AC) | Efficiency needed from now on to finish on budget |
Formulas as published by PMI. Note that SV is expressed in money, not time: an SV of −$9,000 means you have completed $9,000 less work than planned by this date, not that you are a certain number of days late.
A worked example
An IT services firm is rebuilding a customer portal for a client. The approved budget (BAC) is $120,000 over six months. At the end of month three, the plan said $60,000 of work would be complete (PV). Accepted deliverables are worth $51,000 at their budgeted value (EV), and the firm has spent $57,000 in billable effort and licenses (AC).
SV
- Calculation
- 51,000 − 60,000
- Result
- −$9,000
CV
- Calculation
- 51,000 − 57,000
- Result
- −$6,000
SPI
- Calculation
- 51,000 ÷ 60,000
- Result
- 0.85
CPI
- Calculation
- 51,000 ÷ 57,000
- Result
- 0.89
EAC
- Calculation
- 120,000 ÷ 0.8947
- Result
- ≈ $134,100
VAC
- Calculation
- 120,000 − 134,118
- Result
- ≈ −$14,100
TCPI
- Calculation
- 69,000 ÷ 63,000
- Result
- 1.10
| Measure | Calculation | Result |
|---|---|---|
| SV | 51,000 − 60,000 | −$9,000 |
| CV | 51,000 − 57,000 | −$6,000 |
| SPI | 51,000 ÷ 60,000 | 0.85 |
| CPI | 51,000 ÷ 57,000 | 0.89 |
| EAC | 120,000 ÷ 0.8947 | ≈ $134,100 |
| VAC | 120,000 − 134,118 | ≈ −$14,100 |
| TCPI | 69,000 ÷ 63,000 | 1.10 |
Read together, these say the project is behind (SPI 0.85) and somewhat over budget (CPI 0.89). If nothing changes, it will finish roughly $14,000 over. To land on budget, the team would need to deliver $1.10 of planned work for every remaining dollar, noticeably better than its current 0.89. That gap is the real finding: it suggests a scope, staffing or budget conversation now, in month three, rather than an awkward surprise in month six.
Making earned value honest
- Earn value only for objectively complete work. PMI warns that subjective percent-complete figures (“90% done” for three weeks) produce misleading indices. Many teams earn nothing until a deliverable is accepted.
- On agile teams, measure in stories, story points or releases, and count value only when an increment meets the Definition of Done and is accepted.
- Read the trend, not a single data point. One period below 1.0 is noise; three in a row is a pattern.
- Effort can stand in for money. If your budget is in hours, PV, EV and AC can all be in hours, as long as you stay consistent.
If full EVM is too heavy for a small project, track at least budget burn vs work accepted, and planned vs actual hours from approved timesheets. Our guide to time tracking and resource planning covers how to get reliable effort data in the first place.
Flow KPIs: cycle time, throughput, WIP and age
For teams running continuous or iterative delivery, flow metrics are often more useful than earned value. The Kanban Guide defines four:
Work in progress (WIP)
- Kanban Guide definition
- Work items started but not finished
- What it reveals
- Overload and context switching
Throughput
- Kanban Guide definition
- Work items finished per unit of time
- What it reveals
- Delivery rate, and the basis for forecasts
Work item age
- Kanban Guide definition
- Time from start to today, for unfinished items
- What it reveals
- Stuck work, before it becomes late work
Cycle time
- Kanban Guide definition
- Time from start to finish, for finished items
- What it reveals
- Predictability; how long a typical item takes
| Measure | Kanban Guide definition | What it reveals |
|---|---|---|
| Work in progress (WIP) | Work items started but not finished | Overload and context switching |
| Throughput | Work items finished per unit of time | Delivery rate, and the basis for forecasts |
| Work item age | Time from start to today, for unfinished items | Stuck work, before it becomes late work |
| Cycle time | Time from start to finish, for finished items | Predictability; how long a typical item takes |
Work item age is the leading indicator many teams ignore. If most items finish within six days and one has been in review for fifteen, that item needs attention today, not at the retrospective.
Report cycle time as a percentile rather than an average, for example “85% of items finish within 7 days”. Averages hide the long tail that customers and stakeholders actually feel. And only compare like with like: a two-hour bug fix and a three-week integration should not share one cycle time figure. Scrum teams also track velocity, but treat it as a planning input, not a performance score; our comparison of Kanban and Scrum explains how each approach uses these measures.
Quality KPIs: defects, rework and test coverage
- Escaped defects by severity: bugs found after release or by the customer, rather than during testing.
- Reopen rate: the share of items marked done that come back. A rising rate usually points to unclear acceptance criteria or rushed testing.
- Rework effort: hours spent fixing delivered work, as a share of total effort.
- Test pass rate and coverage of priority requirements: whether the work items that matter most have test cases, and whether those tests pass.
- Acceptance cycle: how long deliverables wait for client or stakeholder sign-off, a common hidden delay in agency and services work.
Avoid optimizing a single quality number. A team told to raise test count will write many trivial tests. Coverage is meaningful only when it is measured against important risks and requirements. Our test management guide goes deeper on linking tests, runs and defects to the work they validate.
Capacity and workload KPIs
Capacity KPIs show whether the plan is realistic for the people available. Track planned work against available hours per person or role, utilization trends and the share of time going to unplanned work such as support and incidents.
Two cautions. First, logged hours measure effort, not productivity, so never rank people by them. Second, sustained utilization near or above 100% is a warning, not an achievement: it leaves no slack for the unexpected and usually shows up later as missed dates or quality problems. Use workload views to start a conversation about priorities, and remember to subtract meetings, support rotations and leave before you calculate capacity.
Outcome and stakeholder KPIs
Projects exist to change something. A project that ships on time and budget but delivers no benefit has not succeeded. Agree outcome measures during planning, along with when and by whom they will be measured, because benefits often appear after the project closes. Examples by project type:
SaaS onboarding redesign
- Outcome KPI
- Trial-to-paid conversion; time to first value
- When to measure
- 4–8 weeks after release
IT service desk migration
- Outcome KPI
- Ticket resolution time; first-contact resolution
- When to measure
- Monthly for a quarter after cut-over
Marketing campaign
- Outcome KPI
- Qualified leads; cost per lead
- When to measure
- During and 30 days after the campaign
Internal ERP rollout
- Outcome KPI
- Adoption rate; month-end close time
- When to measure
- After the first two close cycles
Agency client website
- Outcome KPI
- Client acceptance; post-launch issue count
- When to measure
- At handover and after 30 days
| Project | Outcome KPI | When to measure |
|---|---|---|
| SaaS onboarding redesign | Trial-to-paid conversion; time to first value | 4–8 weeks after release |
| IT service desk migration | Ticket resolution time; first-contact resolution | Monthly for a quarter after cut-over |
| Marketing campaign | Qualified leads; cost per lead | During and 30 days after the campaign |
| Internal ERP rollout | Adoption rate; month-end close time | After the first two close cycles |
| Agency client website | Client acceptance; post-launch issue count | At handover and after 30 days |
Stakeholder satisfaction can be measured with a two-question pulse at each milestone: “How confident are you that this project will deliver what you need?” and “What would increase that confidence?” The second answer is usually more useful than the score.
Example KPI dashboard for a client project
Here is a compact dashboard for the customer portal project from the earned value example. Eight measures, each with a current value, a threshold and an action. This is the level of detail most steering groups need.
Release milestone forecast
- Current
- 14 days late
- Threshold
- > 5 days late
- Status and action
- Red: scope trade-off decision needed from client
SPI (EV ÷ PV)
- Current
- 0.85
- Threshold
- < 0.95
- Status and action
- Red: add QA capacity for two sprints
CPI (EV ÷ AC)
- Current
- 0.89
- Threshold
- < 0.95
- Status and action
- Amber: review license and rework costs
EAC (BAC ÷ CPI)
- Current
- ≈ $134,100
- Threshold
- > BAC + 5%
- Status and action
- Red: budget conversation with sponsor
Work item age, oldest in review
- Current
- 11 days
- Threshold
- > 7 days
- Status and action
- Amber: pair reviewer on stuck items
Throughput (items/week, 4-wk avg)
- Current
- 14
- Threshold
- < 10
- Status and action
- Green
Escaped defects (sev 1–2, last 30 days)
- Current
- 1
- Threshold
- > 2
- Status and action
- Green
Team utilization (avg)
- Current
- 96%
- Threshold
- > 90% for 2 weeks
- Status and action
- Amber: protect focus time; defer non-critical work
| KPI | Current | Threshold | Status and action |
|---|---|---|---|
| Release milestone forecast | 14 days late | > 5 days late | Red: scope trade-off decision needed from client |
| SPI (EV ÷ PV) | 0.85 | < 0.95 | Red: add QA capacity for two sprints |
| CPI (EV ÷ AC) | 0.89 | < 0.95 | Amber: review license and rework costs |
| EAC (BAC ÷ CPI) | ≈ $134,100 | > BAC + 5% | Red: budget conversation with sponsor |
| Work item age, oldest in review | 11 days | > 7 days | Amber: pair reviewer on stuck items |
| Throughput (items/week, 4-wk avg) | 14 | < 10 | Green |
| Escaped defects (sev 1–2, last 30 days) | 1 | > 2 | Green |
| Team utilization (avg) | 96% | > 90% for 2 weeks | Amber: protect focus time; defer non-critical work |
Every KPI on the dashboard should also have a short definition record so the numbers mean the same thing to everyone:
Milestone forecast
- Question it answers
- Will we hit the committed date?
- Data source
- Project timeline
- Owner
- Project manager
- Review
- Weekly
SPI / CPI / EAC
- Question it answers
- Are we getting planned value for money and time?
- Data source
- Budget sheet + accepted deliverables
- Owner
- Project manager
- Review
- Monthly
Work item age
- Question it answers
- What is stuck right now?
- Data source
- Board / workflow history
- Owner
- Delivery lead
- Review
- Daily
Escaped defects
- Question it answers
- Is quality holding as we speed up?
- Data source
- Bug tracker
- Owner
- QA lead
- Review
- Per release
Utilization
- Question it answers
- Is the plan realistic for the people we have?
- Data source
- Approved timesheets
- Owner
- Resource manager
- Review
- Weekly
| KPI | Question it answers | Data source | Owner | Review |
|---|---|---|---|---|
| Milestone forecast | Will we hit the committed date? | Project timeline | Project manager | Weekly |
| SPI / CPI / EAC | Are we getting planned value for money and time? | Budget sheet + accepted deliverables | Project manager | Monthly |
| Work item age | What is stuck right now? | Board / workflow history | Delivery lead | Daily |
| Escaped defects | Is quality holding as we speed up? | Bug tracker | QA lead | Per release |
| Utilization | Is the plan realistic for the people we have? | Approved timesheets | Resource manager | Weekly |
Notice what is missing: total tasks closed, hours logged per person and lines of code. None of them would change a decision on this project. When you share the dashboard, the project status report is where you add the narrative: what changed, why, and what you need from stakeholders.
How to choose your first five KPIs
- 1.List the three to five questions your sponsor or client asks most often. Each KPI must answer one of them.
- 2.Pick at most one or two measures per question. Prefer forecasts and trends over snapshots.
- 3.Write the calculation, data source, owner, review cadence and threshold for each one before showing any numbers.
- 4.Baseline the plan. Variance means nothing without an agreed baseline, and baselines should change only through approved change control.
- 5.Run the dashboard for four to six weeks, then remove any metric that did not trigger a conversation or decision.
- 6.Revisit at each phase gate. The KPIs that matter in planning (estimate confidence, open risks) differ from those that matter near launch (defect trend, readiness).
Risk exposure deserves a place on most dashboards too, typically as the number of high-rated risks without an active response. Our guide to project risk management shows how to score and track them.
Common ways KPIs mislead
- Ranking individuals by task counts or hours. It destroys trust and rewards ticket splitting and padded timesheets.
- Silently resetting baselines. If scope changes, record the approved change and show both the original and current baseline. Otherwise variance disappears and nobody learns.
- Watermelon reporting: green on the outside, red inside. Many low-risk tasks complete can hide one critical item that is badly late. Weight status by milestone and risk, not task volume.
- Comparing teams on velocity. Story points are relative to each team; comparing them across teams invites inflation.
- Measuring without definitions. If “done” means merged for one team and released for another, cycle time comparisons are meaningless.
- Tracking what is easy instead of what matters. Hours are easy to collect; outcomes are harder. Do the harder work.
Where QuikProjects fits
Most KPI pain comes from data scattered across tools: tasks in one system, hours in another, test results in a third, then copied into a spreadsheet each week. QuikProjects keeps sprints, timesheets, test runs and docs in one workspace, so several of the measures above come straight from the work itself.
- The executive report shows completion, on-time delivery, slipping work and hours variance team by team, with weekly scorecards and week-over-week change.
- Planned vs actual hours come from approved timesheets, and the resource report shows expected, spent and estimated hours per person, with overload highlighted.
- Velocity reports, test coverage per work item and JUnit results matched to test cases support the flow and quality KPIs.
- Saved views and CSV export let you combine project data with your budget figures if you calculate earned value in a spreadsheet.
- On Premium and Pro, the AI Project Assistant summarizes sprints and flags risks, which helps you spot at-risk work between reviews.
Earned value itself needs budget and cost data, so most teams pair project data with their finance figures rather than expecting any project tool to know their full cost base.
Frequently asked questions
How many project management KPIs should a team track?
For a single project, five to eight is usually enough: one or two each for schedule, cost or effort, flow, quality and outcome. A portfolio dashboard can roll up a smaller set across projects. If a metric has not changed a decision in a month or two, remove it.
What is the difference between a KPI and a metric?
A metric is any measurement. A KPI is a metric tied to an objective, with a target or threshold and an owner who acts on it. Throughput is a metric; “throughput below 10 items a week for two weeks triggers a scope review” makes it a KPI.
Is earned value management useful for agile projects?
Yes, with care. PMI notes that EVM can be applied using stories, story points, iterations or releases, as long as value is earned only when an increment is genuinely done and accepted. Many agile teams combine a light release-level EVM view for sponsors with flow metrics for day-to-day management.
Should schedule variance be reported in days or money?
In earned value, SV is a money figure (EV − PV). Stakeholders usually also want to know how many days late a milestone is, so report the milestone forecast in days alongside SPI. Do not convert SV to days by a simple formula; it can mislead when spending is uneven across the schedule.
The bottom line
Good project KPIs are few, defined and connected to decisions. Start with the questions your stakeholders ask, use earned value or flow measures to answer them with evidence, and add outcome measures so the project is judged on what it changed, not just what it shipped. Set thresholds in advance, read trends rather than snapshots, and keep the numbers honest by tying them to accepted work and an agreed baseline. If you are still setting up the plan those KPIs will measure, start with how to create a project plan.
Want the numbers to come from the work instead of a weekly spreadsheet? Try QuikProjects free for up to 10 users, no credit card needed.
Sources
See it running on your business.
A walkthrough built around how your team works today.


