Every contractor measures something. The trouble is that most of what gets measured describes what happened after it became expensive to fix: a loss that surfaces in the final account, an injury logged after it occurred, a delay the programme admits three months after it started. The useful key performance indicators are the ones that move before the outcome, and are computed from a daily record rather than from the project manager’s estimate at the month-end meeting. This guide walks through cost, time, cash, quality, safety, documents and labour, computes each KPI with a worked example, and then suggests how many a project dashboard can carry before it turns into decoration.
Key takeaways
- A leading indicator moves before the outcome and leaves room to act; a lagging one confirms what happened. A dashboard needs both, but decisions are made on the first.
- A KPI is never more accurate than the record it is computed from: a CPI built on missing costs or eyeballed percent-complete is optimism with two decimal places.
- Around ten KPIs is enough for one project; every extra one has to justify itself with a decision that changes because of it.
- Green, amber and red thresholds are alarms, not verdicts: a three-month trend says more than one month’s colour.
- LTIFR is a lagging indicator; the near-miss reporting rate is what tells you the safety culture is working before anyone gets hurt.
Leading indicators versus lagging ones
The difference between leading and lagging is not academic. A lagging indicator measures an outcome that has already happened: final margin, injuries, actual handover date. It is essential for accountability and for comparing projects, but it gives you no chance to act — by the time it shows, the matter is closed. A leading indicator measures a condition that comes before the outcome: commitments placed against the budget before the money is spent, float eroding on the critical path before handover slips, near misses before they become injuries.
| Area | Leading indicator | Lagging indicator |
|---|---|---|
| Cost | Committed vs budget | CPI and cost variance |
| Time | Float burn on the critical path | SPI and % milestones on time |
| Cash | Days to certify | Billed vs earned, retention outstanding |
| Quality | First-time inspection pass rate | Open NCRs and their age |
| Safety | Near-miss reporting rate | LTIFR and TRIR |
| Documents and labour | RFI and submittal turnaround | Productivity: actual vs planned man-hours |
The split is not rigid. CPI is lagging because it measures money already spent, but it becomes leading the moment you use it to forecast the cost at completion. The test is always the same question: does this number buy me time to act? If it does not, it belongs in the monthly report to the board, not on the site manager’s screen.
A KPI is only as good as its record
A beautifully coloured dashboard can be entirely wrong. CPI divides earned value by actual cost; if paid supplier invoices or plant fuel are never booked as expenses against the project, the denominator is short and the index reads better than the truth. SPI depends on percent-complete; if those percentages are eyeballed in a month-end meeting, the index is the project manager’s opinion printed to two decimal places.
- Actual cost comes from real expense entries: paid invoices, approved subcontractor applications, payroll computed from attendance, fuel and plant. Any branch of spend that books no entry is invisible to every cost KPI at once.
- Percent-complete comes from quantities measured on site and recorded in daily site reports or task updates, not from a verbal estimate.
- Man-hours come from attendance stamped with time and place, not a timesheet filled in from memory on Thursday afternoon.
- Inspections, snags and near misses only count if they were logged when they happened; what was never recorded never enters any rate.
Cost and time: a worked example in SAR
An office building with a budget at completion (BAC) of SAR 48,000,000. At the end of month nine, planned value was SAR 22,000,000, earned value — the budgeted value of the work actually done — was SAR 20,000,000, and recorded actual cost was SAR 21,500,000.
| KPI | Formula | Calculation | Reading |
|---|---|---|---|
| CPI | EV ÷ AC | 20,000,000 ÷ 21,500,000 = 0.93 | Every riyal spent produces 93 halalas of work |
| Cost variance | EV − AC | 20,000,000 − 21,500,000 | SAR −1,500,000 |
| SPI | EV ÷ PV | 20,000,000 ÷ 22,000,000 = 0.91 | 91% of the work planned to date is done |
| Schedule variance | EV − PV | 20,000,000 − 22,000,000 | SAR −2,000,000 of work behind |
| EAC | BAC ÷ CPI | 48,000,000 × 21,500,000 ÷ 20,000,000 | SAR 51,600,000 if performance holds |
Both indices are lagging by nature, and reading them together matters more than either alone — covered in detail in how to read CPI and SPI together. The leading cost KPI is committed versus budget. The rebar line has a budget of SAR 6,200,000; purchase orders placed total SAR 5,400,000, but they cover only 70% of the tonnage required. Little of it has been spent yet, and still the projection is plain: 5,400,000 ÷ 0.70 = SAR 7,714,286, an expected overrun of about SAR 1,514,286 before a single bar reaches site.
Time: float erodes before handover slips
SPI hides something important: it can read 0.98 while the project is genuinely late, because progress on non-critical work offsets a slip on a critical activity. The leading time KPI is float burn on the critical and near-critical paths. The finishes chain in our example carried 20 working days of total float in the approved baseline; eight weeks later it has 9 left. That is 11 days in 8 weeks, roughly 1.4 days a week, and at that rate the remaining float is gone in about six and a half weeks and the chain becomes the new critical path. The mechanics are in our critical path scheduling guide.
The simplest time KPI everyone understands is the share of milestones hit on or before their date: 7 of the 10 falling due so far is 70%. It is crude, but it is hard to game, and it is the first thing a client asks.
Cash: earned, billed, certified, collected
A project that is profitable on paper can still sink its contractor. Cash KPIs measure the distance between work done and money actually in the bank, and every stage of that chain has its own indicator. Continuing the same example, earned value is SAR 20,000,000.
| Stage | Amount (SAR) | KPI |
|---|---|---|
| Earned value | 20,000,000 | — |
| Billed in payment applications | 18,400,000 | Billed ÷ earned = 92%: 1,600,000 of work done but not claimed |
| Certified by the consultant | 17,200,000 | Average 23 days from submission to certification |
| Retention at 10% of certified | 1,720,000 | Earned money that will not arrive before handover and defects liability |
| Net due after retention | 15,480,000 | 17,200,000 − 1,720,000 |
| Actually collected | 14,600,000 | Average 41 days from certification to payment; 880,000 certified but unpaid |
The leading indicator in this chain is days to certify. When it climbs from 15 to 23, it usually means applications are being rejected or returned for missing backup, and collection will slip a month later. Billed-to-earned under 100% means the site is doing work it is not claiming — usually unapproved variations or quantities nobody measured. See payment applications and cash flow for closing that gap, and watch retention: the balance that quietly accumulates until the end of the job.
Quality and safety: count it before the damage
First-time inspection pass rate is the best leading quality KPI: of 120 inspection requests this quarter, 93 passed first time, which is 77.5%. Every failed inspection means rework, a second visit from the consultant and a delay to the next activity. A falling pass rate shows up before rework cost does, and long before the snag list at handover.
- NCRs: the open count alone is not enough. Fourteen open, five of them older than 30 days, says corrective actions are not closing, which is more serious than the number itself. More in quality inspections and corrective actions.
- Snag close rate: 340 snags raised and 255 closed is 75%. What matters more is the weekly trend — 60 new and 45 closed a week means the list is growing and will not be clear by handover.
Safety: the lagging rates and what comes before them
LTIFR is lost-time injuries × 1,000,000 ÷ hours worked. Our project logged 850,000 hours and two lost-time injuries: 2 × 1,000,000 ÷ 850,000 = 2.35. Some clients and regimes use a 200,000-hour basis instead, which turns the same record into 0.47 — so confirm the basis before comparing anything. TRIR is normally on the 200,000-hour basis: 6 recordable injuries × 200,000 ÷ 850,000 = 1.41.
All of these are lagging: they tell you how many people were hurt. The leading KPI is the near-miss reporting rate: 85 reports × 200,000 ÷ 850,000 = 20 per 200,000 hours. A rising figure is usually good news — workers are reporting and not afraid of blame. A sudden drop while the workload stays the same usually means reporting stopped, not that the hazards went away. More in construction safety management.
Documents and labour: where time leaks quietly
An RFI that waits two weeks is either a stalled activity on site or work proceeding on an assumption that may be torn out later. RFI and submittal turnaround is a direct leading indicator of delay and of variations.
- The average turnaround misleads because one very late item distorts it; use the median plus the count open longer than 14 days. Example: 42 RFIs closed with a median of 6 days, and 9 still open after more than 14 days — those nine are the agenda for the next consultant meeting.
- The first-submission approval rate for shop drawings and material samples tells you how often paperwork is redone, and every resubmission delays procurement. See RFI and submittal turnaround.
Productivity: earned hours versus actual
Blockwork in the example is planned at 1.6 man-hours per square metre. The crew has laid 1,250 m² using 2,350 actual man-hours. Earned hours = 1,250 × 1.6 = 2,000, and the productivity factor = 2,000 ÷ 2,350 = 0.85. The crew is taking about 1.88 hours a square metre instead of 1.6, so every remaining metre will cost about 17.5% more hours than planned unless something changes. This KPI shows up before the extra labour cost reaches payroll — provided the hours come from verified attendance and measured output.
How many KPIs a dashboard can carry
A dashboard with thirty KPIs is unread by week two. The working rule: one or two per area, so about ten for a single project, and every one needs a named owner and a decision that changes when its colour does. If you cannot name the decision, delete the KPI.
| KPI | Green | Amber | Red |
|---|---|---|---|
| CPI | 0.98 or above | 0.95 to under 0.98 | Under 0.95 |
| SPI | 0.98 or above | 0.95 to under 0.98 | Under 0.95 |
| Projected commitment vs line budget | Within budget | Up to 5% over | More than 5% over |
| Days to certify | 14 days or fewer | 15 to 28 days | More than 28 days |
| First-time inspection pass rate | 90% or above | 80% to under 90% | Under 80% |
| NCRs older than 30 days | None | 1 to 3 | More than 3 |
| RFIs open more than 14 days | 0 to 2 | 3 to 5 | More than 5 |
Pick ten
One leading and one lagging KPI for each area that matters on this job, and add a new one only by removing another.
Tie each KPI to its source
Name the record each figure is computed from — expenses, attendance, inspections, reports — and make sure it is filled daily, not monthly.
Show the trend, not the snapshot
Put the last three to six readings beside every figure so improvement and decline are visible before the colour changes.
Review thresholds quarterly
A band that has not gone red in a year is either too wide or measuring something that no longer matters.
In muqawil, CPI, SPI and the S-curve are computed from the project’s own expenses and budget, the portfolio dashboard rolls up each project’s health in one place, and the quality, safety and RFI figures come from the same inspection, NCR, near-miss and RFI records the site fills in.
Frequently asked questions
What are the most important construction project KPIs?
About ten is enough: CPI and committed versus budget for cost, SPI and critical-path float burn for time, days to certify and billed versus earned for cash, first-time inspection pass rate and NCR ageing for quality, near-miss reporting rate and LTIFR for safety, plus RFI turnaround and a labour productivity factor.
What is the difference between leading and lagging indicators in construction?
A lagging indicator measures an outcome that has already happened, such as injuries or final margin, and leaves no room to act. A leading indicator measures a condition that comes before the outcome, such as near misses or commitments against budget, and gives you time to intervene before the loss lands.
How is LTIFR calculated?
Lost-time injuries multiplied by 1,000,000 and divided by total hours worked. Two lost-time injuries in 850,000 hours gives 2.35. Some organisations use a 200,000-hour basis instead, which gives 0.47 for the same record, so check the basis before comparing two projects.
Is a CPI above 1.0 always good news?
Not always. It can mean the work is coming in under budget, or it can mean real costs have not yet been booked to the project, such as paid supplier invoices, plant fuel or payroll. A sudden jump with no obvious cause is worth checking for missing expenses before anyone celebrates.
How many KPIs should a construction dashboard show?
Around ten for a single project, one or two per area. Each one needs a named owner and a decision that changes when its colour changes; otherwise it takes up space and pulls attention away from the indicators that actually matter.
KPIs computed from the site’s own record
muqawil computes CPI, SPI and the S-curve from approved expenses — paid supplier invoices and approved subcontract payment applications included — rolls project health into one portfolio dashboard, records inspections, NCRs and RFIs, and lets the site log near misses and daily reports even with no signal.