The 9 HR Metrics a CEO Should See Every Week (With Dashboard Templates)

Most HR dashboards fail for the same reason most dashboards fail: they show everything that can be measured rather than the handful of things that would change a decision this week. A CEO looking at 24 tiles learns less than one looking at 9, because the signal is buried in metrics that only move quarterly.

The filter is simple. A metric earns a weekly slot if it satisfies two tests: it can move meaningfully in seven days, and a change in it would trigger an action. Engagement scores fail the first test. Training hours fail the second. Nine metrics pass both.

Table of contents

• The nine metrics
• Leading vs lagging: which ones predict
• Building the dashboard in an afternoon
• What to review quarterly instead
• FAQs

The nine metrics

#MetricFormulaWatch for 1Headcount vs planActual FTE − planned FTEDrift either direction beyond ±5% 2Unplanned attrition (rolling 12mo)Voluntary leavers ÷ average headcountAny month-over-month acceleration 3Open roles agingDays since each requisition openedAnything past 45 days 4Time to hireDays from posting to accepted offerTrend, not the single number 5Absence rateAbsent days ÷ scheduled daysConcentration in one team 6Overtime hoursTotal OT hours this weekSame names repeating 7Payroll as % of revenueTotal payroll cost ÷ revenueMovement without a headcount change 8Leave liabilityAccrued unused leave days × daily rateSteady growth = a future cash event 9Onboarding completionNew hires past day 30 with checklist completeAnything below 90%

1. Headcount vs plan

The base number everything else is normalised against. Track actual full-time equivalents against the hiring plan, split by department. The value is not the total — it is the variance. A department running three heads under plan is either a hiring problem or a budget the finance team is about to reallocate; either way it is a conversation this week, not next quarter.

2. Unplanned attrition rate

Voluntary departures divided by average headcount, on a rolling twelve-month basis so a single resignation in a small company does not create a false spike. Separate voluntary from involuntary — they have completely different causes and completely different responses. The signal worth acting on is acceleration: three months of rising voluntary attrition means something structural, and exit interviews are where you find out what. Our exit interview question set is built for exactly this.

3. Open roles aging

Not "how many roles are open" but "how long has each been open." A requisition that has been live for 70 days is telling you something specific: the compensation band is wrong, the job description is wrong, or the interview loop is losing candidates. All three are fixable, and all three get worse with time.

4. Time to hire

Days from posting to accepted offer. Measure the trend rather than the absolute number, because the absolute varies enormously by role type. If it is lengthening while open-role aging is also rising, the bottleneck is in your process, not the market.

5. Absence rate

Absent days as a share of scheduled days. The aggregate is nearly useless; the distribution is where the information is. A company-wide 3% absence rate that is actually 1% everywhere and 9% in one team is a management signal disguised as an HR statistic.

6. Overtime hours

The single most under-used early-warning metric in small companies. Sustained overtime concentrated on the same three names is the leading indicator of both burnout and a resignation you have not received yet. It is also a direct wage-and-hour exposure if the hours are not being recorded properly — see our guide to attendance records and wage-hour compliance.

7. Payroll cost as a percentage of revenue

The one metric on this list a CFO will already be watching, which is exactly why it belongs on the HR dashboard too — it is the shared language. Benchmarks vary widely by sector, so the useful comparison is against your own trailing twelve months. Movement without a corresponding headcount change means compensation drift, overtime, or contractor spend that has quietly become permanent.

8. Leave liability

Accrued unused leave multiplied by daily pay rate. In jurisdictions where accrued leave is payable on termination, this is a real balance-sheet liability that most small companies discover for the first time during a payout. Watching it weekly turns a surprise into a policy decision — and if it is growing steadily, the underlying issue is usually that people do not feel able to take time off. Our PTO policy guide covers the fix.

9. Onboarding completion

Share of new hires past day 30 with a complete onboarding checklist. It looks like an administrative metric and it is actually a retention metric — incomplete onboarding correlates strongly with early attrition, and early attrition is the most expensive kind because you have paid the full hiring cost and received none of the productivity. Below 90% means your onboarding process exists on paper only.

Leading vs lagging: which ones predict

Half the value of a dashboard is knowing which numbers are warnings and which are receipts.

Leading (predicts)Lagging (confirms) Overtime concentrationAttrition rate Absence rate by teamPayroll % of revenue Open roles agingTime to hire Onboarding completionLeave liability Headcount variance vs plan—

The practical rule: act on the left column, report the right column. A CEO who only reviews lagging metrics is always managing something that already happened. Attrition tells you people left. Overtime concentration told you they were going to, six weeks earlier.

Building the dashboard in an afternoon

Four steps, in this order. The order matters — most failed dashboards started at step three.

1. Define each metric in writing before you build anything. "Headcount" means something different to finance and HR. Write the formula, the data source, and the inclusion rules (contractors? part-time as FTE fractions? employees on notice?). Disagreements discovered at this stage cost an hour; discovered after launch they cost the dashboard's credibility.
2. Find where each number lives. If four of the nine require manual assembly from spreadsheets, the dashboard will be current for two weeks and then abandoned. Any metric that cannot be pulled automatically should be dropped or the underlying system fixed — those are the only two honest options.
3. Build the smallest version that works. One screen, nine numbers, each with its current value, the prior week, and the direction of travel. No pie charts. No drill-downs in version one.
4. Attach it to a recurring meeting. A dashboard nobody is scheduled to look at is a reporting artefact, not a management tool. Fifteen minutes, same slot every week, one owner per metric.

On tooling: if your HR data lives in one system, the dashboard should too — exporting to a BI tool adds a refresh step that eventually stops happening. If it is spread across four tools, the dashboard project is really a consolidation project, and you should read our comparison of HR analytics dashboards before building anything.

What to review quarterly instead

These matter, and they do not belong on a weekly screen — they move too slowly, and watching them weekly trains people to react to noise:

• Engagement score / eNPS — quarterly at most; weekly measurement changes the behaviour being measured
• Compensation ratio vs market — twice a year, tied to your benchmarking cycle (see salary benchmarking on a budget)
• Diversity representation — quarterly, with sufficient population size to be meaningful
• Training and development hours — quarterly
• Cost per hire — quarterly; too volatile weekly in small companies
• Span of control / org shape — annually or on reorganisation
• Internal mobility rate — annually

Key takeaways

• A metric earns a weekly slot only if it can move in seven days and would trigger an action.
• The nine: headcount vs plan, unplanned attrition, open roles aging, time to hire, absence rate, overtime hours, payroll % of revenue, leave liability, onboarding completion.
• Overtime concentration is the most under-used early-warning metric in small companies.
• Act on leading indicators; report lagging ones.
• Define every formula in writing before building — definitional disagreement is what kills dashboards.
• Any metric requiring manual assembly will be stale within a month.
• Engagement, eNPS, training hours and cost per hire are quarterly metrics, not weekly ones.

Frequently asked questions

What HR metrics should a CEO look at weekly?

Nine: headcount versus plan, unplanned attrition rate, open roles aging, time to hire, absence rate, overtime hours, payroll cost as a percentage of revenue, leave liability, and onboarding completion. Each can move within a week and each implies a specific action when it does.

How do you calculate employee attrition rate?

Divide the number of employees who left during a period by the average headcount for that period, then express it as a percentage. Use a rolling twelve-month window in companies under about 100 people, and always separate voluntary from involuntary departures — they have different causes and different fixes.

What is a good absence rate?

Rates vary substantially by industry and country, so external benchmarks are of limited use. Compare against your own trailing twelve months and look at the distribution by team rather than the company average — concentration in a single team is the actionable signal.

What is leave liability and why does it matter?

Leave liability is accrued unused leave multiplied by employees' daily pay rate. In jurisdictions where accrued leave is paid out on termination it is a genuine financial obligation, and it can grow quietly for years. Rising liability usually also indicates that employees do not feel able to take time off.

Should small companies track HR metrics at all?

Yes, but fewer of them and with less precision than large companies. In a 30-person company the useful metrics are the ones that identify a specific person or team needing attention this week — overtime concentration, absence by team, onboarding completion — not statistical aggregates that need a large population to be meaningful.

What is the difference between leading and lagging HR indicators?

Leading indicators predict future outcomes and can be acted on — overtime concentration, absence by team, open roles aging, onboarding completion. Lagging indicators confirm what already happened — attrition rate, time to hire, payroll as a percentage of revenue. A dashboard of only lagging indicators reports history rather than enabling intervention.

Get these nine numbers without building anything

TracefyHR's budget and analytics module reports headcount, attrition, absence, overtime, leave liability and hiring metrics from live data — no exports, no spreadsheet assembly. If you want a metric it does not ship with, describe it to Forge AI in plain English and it builds the tracker and the dashboard in about a minute. All on a flat $49/month. See analytics in TracefyHR or start a 30-day free trial.