You cannot fix what you do not measure, but you can also drown in dashboards that do not predict anything. For Canadian SMBs, a retention measurement stack should be small enough to fit on one page and predictive enough to drive actual decisions. This guide covers the five metrics worth tracking, how to calculate each, and the leading indicators that quietly predict who is about to leave.
The 5 metrics worth tracking
More than five and you stop looking. Fewer than five and you miss the signal. The right stack for a Canadian SMB:
- Voluntary turnover rate (annual): the rate of employees who chose to leave. The most common headline number; useful with caveats.
- First-year survival rate (cohort): the share of new hires still with you 12 months after start. Predicts onboarding quality.
- Average tenure (current employees): the average time current employees have been with you. Trends matter more than the number.
- Regrettable turnover rate (annual): the rate of departures that you would not replace with the same person. Catches the quality of who is leaving, not just the volume.
- Manager 1-on-1 cadence (rolling 30 days): a process metric, not an outcome metric. Predicts the next 90 days more reliably than any of the four above.
How to calculate each
1. Voluntary turnover rate
Formula: (voluntary departures in period / average headcount during period) × 100.
Example: 12-person team, 2 voluntary departures over the year, average headcount 13. Voluntary turnover = (2 / 13) × 100 = 15.4%.
Watch for: distinguish voluntary from involuntary (layoffs, terminations). Lumping them together makes the metric unusable for retention decisions. Statistics Canada separation rates do this distinction, your internal number should too.
2. First-year survival rate
Formula: (employees from a hiring cohort still active 12 months later / total cohort size) × 100.
Example: 8 hires made in Q1 2025; in Q1 2026, 6 are still active. First-year survival = (6 / 8) × 100 = 75%.
A Canadian SMB benchmark: 80% to 90% is healthy. Below 75% is a flag that onboarding, role clarity, or hiring fit is breaking down. Cohort analysis (instead of pooled annual numbers) is more honest because it controls for hiring burst timing.
3. Average tenure
Formula: sum of years employed (current employees only) / headcount.
Watch for: this number is biased by hiring growth. A team that doubled in the last year will have a low average tenure even if no one ever leaves. Track the trend, not the absolute number, and compare cohorts of similar size.
4. Regrettable turnover rate
Formula: (departures classified as regrettable / total voluntary departures) × 100.
At each departure, the manager and one other leader independently answer: would we re-hire this person for the same role today? Both yes = regrettable. Both no = not regrettable. Split = noted as ambiguous, discussed.
A healthy Canadian SMB target: under 50% regrettable. Above 50% means you are losing the people you want to keep. Below 20% may mean you are not hiring ambitiously enough, or that you are too quick to call someone "not regrettable" after the fact.
5. Manager 1-on-1 cadence
Formula: for each manager, the percentage of their direct reports with whom they had a 1-on-1 in the last 30 days.
The threshold that matters is 100%. A manager who is at 80% is signalling that two of their five reports are not being heard from. Those two are statistically the next ones to leave.
Leading indicators worth watching
Turnover rate is a lagging indicator. By the time it moves, the resignations have happened. The signals that predict departures before they happen are usually behavioural:
- A drop in 1-on-1 attendance or engagement. Cancelled meetings, shorter conversations, vague answers to "what are you working on next quarter?"
- Withdrawal from longer-term projects. Someone who used to volunteer for cross-team work stops.
- A pattern of vague PTO requests. Multiple short, unspecified days off in a quarter often correlates with interviewing.
- Sudden visibility on LinkedIn. Profile updates, new skills, "open to work" signal flips. A blunt instrument, but worth noticing.
- Promotion or compensation conversation that did not happen. A high performer who has not had a raise or promotion conversation in 18+ months is materially more likely to leave.
The trap is using these as surveillance signals. Use them as conversation prompts. A weekly 1-on-1 where the manager asks "what is going well and what would make you consider leaving?" surfaces most of the above before they harden into a decision.
Exit interviews: do them, but do not trust them blindly
Exit interviews are useful for surfacing themes across departures, but a single exit interview is often half-true. Departing employees soften reasons to avoid bridge-burning, and managers conducting their own exit interviews get sanitized answers.
Three rules that make exit interviews more honest:
- Have someone other than the direct manager conduct it. HR, the founder, or a peer leader. Removes the immediate relationship pressure.
- Ask for the second reason, not the first. "Pay" or "new opportunity" is almost always a surface answer. Ask what made them open to looking in the first place.
- Do a 6-month follow-up. A short note 6 months after departure asking what they would tell their past self. People are dramatically more honest once they are settled elsewhere.
A one-page monthly retention dashboard
The dashboard a Canadian SMB founder should be looking at monthly fits on one page. Suggested layout:
- Top of page: voluntary turnover (rolling 12 months), regrettable turnover (rolling 12 months), first-year survival (rolling cohort).
- Middle: manager 1-on-1 coverage (last 30 days, by manager), engagement pulse score (quarterly).
- Bottom: open roles aging, average days-to-fill, and the names of any departures in the last month with the reason category.
The discipline is to look at it monthly, even when nothing looks alarming. The patterns that hurt are the ones that drift quietly over two or three months and then show up as a wave.
Where to act on the data
Measuring is half the job; acting is the other half. Once the data tells you where the leaks are, the next moves depend on what is broken. For pay and market-fit issues, see our 2026 SMB salary benchmarks. For benefit-side gaps, see benefits that retain talent beyond salary. For onboarding, career paths, and manager training, see building retention culture as your SMB grows. And for the dollar case on why this work matters, see why employee turnover costs Canadian SMBs more.
Frequently asked questions
What is a good voluntary turnover rate for a Canadian SMB?
It depends on sector. As rough Canadian baselines: professional services and skilled trades 10% to 15% annually is healthy; retail and warehousing 20% to 30% is the working range; accommodation and food services often 35% to 50% is closer to industry norm. The actionable comparison is to your own sector median (Statistics Canada or trade association data), not a universal target.
How often should I run an engagement survey?
Quarterly pulse surveys (5 to 10 questions) plus one annual deeper survey works well for SMBs. The trap is surveying often without acting; trust drops faster from unactioned surveys than from no survey at all. If you cannot commit to closing the loop on findings within 30 days, run the survey less often.
Do we need HR software to track these metrics?
Under 50 employees, a well-built spreadsheet is enough. Track hire dates, departure dates, reason category, and regrettable flag in one tab. Most Canadian SMB payroll providers (Wagepoint, Knit, Humi, Rise) include basic turnover reports that satisfy the headline metrics. Reach for dedicated HRIS software when you are running multiple managers and want their 1-on-1 cadence tracked automatically.
Should we share retention metrics with the whole team?
Share the aggregate, not the individual. Voluntary turnover trends, first-year survival, and engagement pulse scores can be shared openly and build trust. Manager-specific metrics (1-on-1 cadence, team retention) should stay between the manager and leadership. The principle is transparency about the company, accountability for individuals.
What if our retention metrics are bad?
Two priorities, in order. First, identify whether the problem is concentrated (one manager, one team, one role) or distributed (the whole company). Concentrated problems usually have a specific cause and a specific fix. Distributed problems are usually about pay, benefits, growth, or leadership tone, and require a longer arc. Second, do not try to fix everything at once. Pick the one lever with the highest leverage for your specific data (often onboarding for first-year survival, or manager 1-on-1 cadence for everything else) and run it for a full quarter before adding the next.