Most companies know their turnover rate the way most people know their weight: one number, checked occasionally, argued with frequently. An employee turnover dashboard replaces that number with a picture — who is leaving, from which teams, how soon after joining, whether you wanted them to go, and what it costs. Done well, it also shows who is about to leave, while there is still time to act.
This guide covers the eight metrics a turnover dashboard should show, how to build one in Excel, how to lay the same thing out in Power BI, and the mistakes that make most turnover dashboards quietly wrong. If you just want the arithmetic, the free turnover rate calculator and cost of turnover calculator do it in seconds.
The 8 Metrics Every Employee Turnover Dashboard Needs
Each one answers a question the one before it raises, so build them in this order.
1. Monthly turnover rate
Formula: separations in the month ÷ average headcount for the month × 100. Average headcount is (headcount on day one + headcount on the last day) ÷ 2.
Why it matters: everything else on the dashboard is a slice of this number. The full walkthrough is in how to calculate employee turnover rate.
2. Annualized and rolling-12 turnover rate
Formula: separations in the last twelve months ÷ average headcount over those twelve months × 100. If you only have a few months of data, annualize by multiplying the period rate by 12 ÷ number of months — with the caveats covered in annualized turnover rate.
Why it matters: monthly rates are jumpy — a team of nine that loses one person posts an 11% month. The rolling-12 line is the one to compare against the benchmarks in what is a good employee turnover rate and turnover rates by industry.
3. Voluntary vs. involuntary split
Formula: the same rate calculation, run twice — once counting only resignations, once counting only terminations and layoffs.
Why it matters: they have opposite causes. Involuntary turnover is a hiring or performance-management problem. Voluntary turnover is a retention problem, and the Work Institute estimates roughly three in four voluntary departures are preventable. Blending the two tells you something is wrong without telling you what. More in voluntary vs. involuntary turnover.
4. Regrettable vs. non-regrettable
Formula: voluntary separations flagged "regrettable" by the manager at exit ÷ average headcount × 100.
Why it matters: losing someone already on a performance plan is not the same as losing your best engineer to a competitor. Regrettable turnover is the number leadership actually cares about, and it only exists if someone records the flag within a week of the exit.
5. Turnover by team and by manager
Formula: rolling-12 voluntary turnover, grouped by team and separately by manager, shown next to the headcount of each group.
Why it matters: Gallup attributes roughly 70% of the variance in team engagement to the manager. Company-wide turnover of 14% can hide one team at 40% and three at 5%. Show headcount beside the rate so nobody panics over a four-person team that lost one person.
6. Turnover by tenure band
Formula: separations bucketed by tenure at exit — 0–90 days, 3–12 months, 1–3 years, 3+ years — each divided by the employees who passed through that band.
Why it matters: the shape of this chart diagnoses the problem. A tall first bar means hiring or onboarding is broken. A spike at 1–3 years usually means a ceiling on pay or growth. A rise at 3+ years is often a culture or leadership shift. Same overall rate, three different fixes.
7. Cost of turnover
Formula: headcount × average salary × annual voluntary turnover rate × replacement cost multiplier. SHRM puts replacement at 50–60% of salary; the Work Institute's low estimate is about 33%, and Gallup's range runs from half to twice salary for senior roles.
Why it matters: 100 employees × $65,000 × 15% turnover × 50% = $487,500 a year. Put that tile at the top of the dashboard and budget conversations change tone. Try your own multiplier in the cost of turnover calculator.
8. Leading indicators
Formula: there is no single one. The panel shows signals that move before a resignation: engagement scores by team, the number of employees whose recognition received has dropped to zero for 60+ days, overdue one-on-ones.
Why it matters: every other metric on the dashboard is a post-mortem; this is the only panel you can act on. Recognition data is a good source because it is continuous and cheap — employees who feel inadequately recognized are roughly twice as likely to say they'll quit within a year (Gallup / Workhuman). How to wire it up is in building a recognition early-warning system for attrition and detecting quiet quitting with recognition data.
Building the Employee Turnover Dashboard in Excel
Three tabs. Resist adding a fourth until the first three work.
Tab 1: Data (one row per employee)
Columns: Employee ID, Hire Date, Exit Date, Exit Type (Voluntary / Involuntary), Regrettable (Y / N), Team, Manager, Salary. Leave Exit Date blank for current staff, and add a helper column, Exit or Future, so the headcount formulas have something to compare against:
=IF(C2="", DATE(9999,12,31), C2)
That is the whole data model. Export it from your HRIS monthly, paste it over the old rows, and the workbook recalculates.
Tab 2: Calc (one row per month)
Column A holds the first day of each month. The next columns are built from COUNTIFS against the Data tab:
- Separations — count rows whose Exit Date falls between the first and last day of the month:
=COUNTIFS(Data!$C:$C, ">="&A2, Data!$C:$C, "<="&EOMONTH(A2,0)) - Voluntary and regrettable separations — the same formula with an extra pair of arguments testing Exit Type for "Voluntary", then another testing Regrettable for "Y".
- Headcount at month end — count rows hired on or before the last day of the month whose Exit or Future column is after it:
=COUNTIFS(Data!$B:$B, "<="&EOMONTH(A2,0), Data!$I:$I, ">"&EOMONTH(A2,0)) - Average headcount — this month's ending headcount averaged with last month's.
- Monthly rate — separations divided by average headcount.
- Rolling-12 rate — the sum of the last twelve months' separations divided by the average of the last twelve months' average headcount. Fill it down from the thirteenth row.
For team and tenure breakdowns, add two small blocks on the same tab: a team list with COUNTIFS filtered by Team, and a tenure block counting rows by days between hire and exit. The Excel traps — blank dates, text that looks like dates, month boundaries — are covered in employee turnover rate formula in Excel.
Tab 3: Dashboard (four charts and three tiles)
- Tiles: rolling-12 voluntary rate, regrettable rate, estimated annual cost.
- Line chart: monthly rate (thin) and rolling-12 rate (bold) on one axis.
- Stacked columns: voluntary vs. involuntary separations by month.
- Horizontal bars: rolling-12 voluntary rate by team, sorted, headcount in the label.
- Columns: separations by tenure band.
Point every chart at the Calc tab, never at Data, so a refresh carries straight through.
A Turnover Dashboard in Power BI
Power BI earns its keep once you want drill-through, multi-year history, or a leading-indicator panel fed by a second system.
Data model
Three tables. Employees (one row per person: ID, team, manager, salary). Events (one row per hire or exit: employee ID, date, event type, voluntary flag, regrettable flag). Calendar, a date table marked as such, related to Events on date. Storing hires and exits as events keeps the headcount measure honest when people leave and are rehired.
Measures
Separations is a count of exit events in the current filter context. Headcount at a date is cumulative hires minus cumulative exits. Average headcount is the mean of the start- and end-of-period headcounts. Turnover rate is separations over average headcount, guarded against division by zero. Rolling-12 wraps the rate in a twelve-month window. Two examples:
Separations = CALCULATE(COUNTROWS(Events), Events[EventType] = "Exit")
Turnover Rate = DIVIDE([Separations], [Average Headcount], 0)
Voluntary and regrettable versions add one more filter on the flag column. Cost of turnover multiplies rolling-12 voluntary separations by average salary and a multiplier held in a what-if parameter, so finance can slide between 33% and 60% live in the meeting.
Page layout
- Headline tiles across the top: rolling-12 voluntary rate, regrettable rate, annual cost, and a count of employees with no recognition in 60+ days.
- Trend line of rolling-12 rate with a reference line at your target or industry benchmark.
- Team heatmap: a matrix of team by month, coloured by voluntary rate, with headcount in the tooltip.
- Tenure bars: separations by tenure band, filterable by team.
- Drill-through page: click any team or manager to see the individual exits with tenure, flags, and reasons.
If you already run a recognition dashboard, its per-person "recognition received" measure drops straight into the leading-indicator tile.
Mistakes Turnover Dashboards Make
- Averaging averages. The mean of twelve monthly rates is not the annual rate. Sum the separations, average the headcount, divide once.
- Mixing headcount definitions. If January counts contractors and February doesn't, the trend line is measuring your spreadsheet, not your company. Write the definition on the dashboard.
- No voluntary split. A layoff month looks identical to a resignation wave. One is a decision; the other is a warning.
- No leading indicators. A dashboard made entirely of exits is a well-formatted obituary. Add at least one signal that moves before the resignation letter does.
How Often to Review It
- Weekly, two minutes: the leading-indicator tiles only. Who went quiet?
- Monthly, in the leadership meeting: headline tiles and the trend line. Did anything move, and do we know why?
- Quarterly, with managers: team and tenure views, each manager's numbers next to the company's.
- Annually: re-check the benchmark, the cost multiplier, and the headcount definition.
The dashboard does not reduce turnover by itself. What reduces it is a manager noticing in week two, not month six, that a good person has gone quiet — and doing something. Deloitte found that strong recognition cultures see up to 31% lower voluntary turnover, which is why recognition data belongs on the same screen as the exits.
That leading-indicator tile is the part we care about most — yes, Propsly is ours. It is a Slack bot: type /props, tag a coworker, add a hashtag, and the recognition posts to a feed channel with points attached. Every give is timestamped per person, which is exactly what a "no recognition in 60+ days" tile needs. The free tier covers unlimited users with 200 props per person per month; Pro adds the analytics that surface who is being overlooked.