The onboarding metrics that actually matter
Completion rates tell you someone finished the paperwork. These metrics tell you whether they can do the job — and whether they want to stay.
Most onboarding metrics measure the wrong thing. Completion rate tells you whether someone ticked boxes. It doesn't tell you whether they understood what they read, whether they know what's expected of them, or whether they feel prepared to do their job. Yet completion rate is what most HR teams report upward, because it's what the system automatically tracks.
The companies that get onboarding measurement right have accepted something uncomfortable: the metrics that are easiest to collect are often the least useful, and the metrics that actually predict long-term success require more deliberate effort to generate.
The metrics worth your attention
Forget completion rates as a primary signal. Track these instead.
90-day retention. If someone leaves in their first three months, onboarding failed — either the role was misrepresented, the integration didn't happen, or they weren't set up to contribute. SHRM estimates that replacing an employee costs between 6–9 months of their salary. A 90-day departure isn't just a hiring problem; it's a process problem. Tracking who leaves before the 90-day mark, and why, is one of the highest-ROI diagnostic tools available to a People team.
Time to first meaningful contribution. Not "time to productivity" in the abstract — specifically, when did this person do something that mattered to a project, a colleague, or a client? This is a qualitative call from the manager, which makes it useful precisely because it forces a conversation. If a manager can't answer this question at the 60-day mark, that's a signal in itself. Either the milestones weren't clear or the manager hasn't been close enough to the work to know.
Role clarity at 30 days. A simple, direct survey question: "Do you understand what's expected of you in the first 90 days?" It takes 30 seconds to answer. The score tells you whether your onboarding is actually communicating expectations or just completing tasks. If it's consistently low across new hires, the structure of your onboarding isn't landing — regardless of what your completion rate says.
Manager confidence at 60 days. Ask the manager: "Do you feel confident this person will succeed in this role?" A low score, caught early, gives you a six-week runway to course-correct. The same conversation held at an exit interview six months later gives you nothing actionable. It's retrospective. By that point, the person's already decided.
What you can't measure with a survey
Some of the most important onboarding signals are structural, not attitudinal. They show up in data before anyone puts them into words.
New hires who are struggling with belonging tend to go quiet. They stop contributing in optional meetings. Their async output drops. In teams with visibility into peer recognition — who's giving kudos, who's receiving them — a clearer picture emerges. Someone who hasn't received any peer recognition in their first 45 days is a very different situation from someone who's been called out publicly three times. Both might answer "fine" on a pulse survey.
The gap between what someone says they're experiencing and what the patterns show is where most retention failures live. People are reluctant to flag problems early, especially when they're new and still trying to make a good impression. By the time they'll say it plainly, they've already made up their mind.
Managers who rely exclusively on self-reported signals — surveys, check-in conversations — are working with incomplete information. The data that makes early intervention possible is structural: milestone completion, peer interaction, document activity. It's available if the system surfaces it. Most systems don't.
When to intervene
The practical reality is that most teams don't act on onboarding signals until something's already broken. A manager notices a performance dip, has a difficult conversation, and it's already month four. The person has mentally moved on. What felt like a performance problem was actually a belonging and clarity problem that had three months to compound.
Useful measurement creates intervention windows, not post-mortems. That means a check-in at week two — not just week one — and a structured 30-day survey before the initial energy has faded. It means reviewing manager confidence scores before the 90-day mark, not after. And it means treating at-risk signals as signals, not as noise to be explained away.
None of this requires sophisticated tooling. It requires a consistent process and someone who reviews the data with enough frequency to act on it. Weekly is better than monthly. Monthly is better than quarterly. Quarterly means the only thing left to track is whether they're still employed.
Closing the feedback loop
Measurement without action is overhead. The point of tracking 90-day retention and role clarity scores isn't to produce a report — it's to find the specific places your onboarding breaks down and fix them.
That means reviewing your data by cohort: do hires in engineering perform differently than hires in sales? Do Q1 hires integrate faster than Q3 hires, when the team is heads-down on planning? Do hires who had structured 1:1s in their first two weeks stay longer than those who didn't? These aren't hard questions to answer if you're collecting the right data. They're almost impossible to answer if you're only tracking completion rates.
The signal you're looking for isn't "did onboarding go well." It's "what specifically went wrong, and for whom, and when." That's the version of measurement that actually improves things.
The On&On Advantage
On&On's AI onboarding generates a milestone-based plan for every new hire — built from the job description and CV, with specific goals per phase rather than a flat task list. That structure makes the "role clarity at 30 days" question easier to answer before you even ask it, because the expectations are already visible to both the hire and the manager.
The Analytics & Visibility Dashboard surfaces the signals that matter: where each hire sits in their milestone plan, whether tasks are overdue, and at-risk flags before they become a resignation. You don't need to run a custom survey to know something's off. The pattern's already in the data.
Ready to get real visibility into your onboarding? Request a demo to see how On&On tracks what actually matters.