Every dealer group we work with can tell you their turnover rate. Very few can tell youwho is about to leave.

That is the gap. Turnover rate is a scoreboard number. It tells you what happened last year, it makes for an uncomfortable slide in the board deck, and by the time it moves there is nothing left to do about it. People analytics is only useful if it gives you something you can act on while the person is still in the building.

The good news is that you do not need a data science team or a data lake to get there.

Most of what predicts an exit is already sitting in your HR system. It is just sitting in five different places and nobody is looking at it together.

First, the number that does not predict anything

Annual turnover rate is a lagging indicator, and treating it as a management tool is the single most common mistake we see.

Two things make it worse in automotive. It is reported at the group level, which hides the one store or one department driving most of it. And it is usually reported annually, which means the signal arrives nine months after the decision it should have influenced.

Keep tracking it. Report it to the board. Just stop expecting it to change anything.

The five metrics that do

1. Onboarding checkpoint completion at day 30 and day 60

Our State of the Automotive Workforce study, built from data across more than 800 dealerships in North America, found that 27.6% of all exits happen before day 90, and 58% happen within the first year. The decision to leave is usually made well before the notice is given.

The measurable version is simple: what percentage of your new hires had a documented 30-day check-in and a documented 60-day check-in? Not "did the manager mean to." Did it happen, and is it recorded. In most groups the honest answer is under half, and the stores where it is lowest are the stores with the worst 90-day retention. That correlation shows up almost every time we look at it.

2. Ramp velocity - time to first sale or first billed hour

Ask any dealer principal what keeps them up at night and this is near the top of the list. It is also the earliest hard signal you have.

Track how long it takes each new hire to reach their first sale, or a technician to reach their first billed hour, and compare it to the cohort average for that store and role.

Someone tracking two weeks behind pace at week six is not just slow. They are usually frustrated, usually not getting the support they need, and usually gone by month four.

This is a metric you can act on the week you see it.

3. Unplanned absence clustering

One missed shift is life. A cluster of last-minute call-offs and shift swaps in the same three-week window is a pattern, and it very often precedes a resignation by a few weeks.

The reason this gets missed is that absence data lives in time and attendance, resignations live in HR, and nobody puts the two on the same timeline. When you do, the shape is usually obvious in hindsight — which means it was available in advance.

4. Training completion drift

Overdue courses are one of the most honest engagement signals you have, and one of the least used. An employee who completed everything assigned to them for eight months and has three courses sitting overdue is telling you something. So is a whole department thay suddenly stops finishing anything.

Track completion rate by employee and by manager, and watch for the drop rather than the absolute number.

5. Exits by manager, not by store

This is the one that changes conversations.

Roll your last two years of exits up by the manager the person reported to instead of by rooftop. Turnover almost never distributes evenly. It concentrates around a handful of managers, and store-level reporting averages that concentration into invisibility.

Fair warning: this metric surfaces uncomfortable answers, and it needs to be handled as a coaching input rather than a scorecard. Most managers with high exit rates were promoted for being good at the job, not for being trained to lead people. That is fixable. But you cannot fix it if the data never separates them from the store average.

Where to start if you are starting from zero

Pick two. Onboarding checkpoint completion and ramp velocity are the easiest to stand up and the fastest to act on, and both are already in your hiring and onboarding records.

Report them monthly, by store and by manager. Set one threshold each — for example, any new hire without a documented 30-day check-in gets flagged to the GM, and any new hire more than two weeks behind cohort ramp pace gets a conversation that week.

That is a people analytics program. It does not need a dashboard project to begin.

How HR4 helps

HR4 holds hiring, onboarding, time and attendance, training, and performance data in one place across every rooftop, which is what makes these five metrics possible to look at together instead of exporting five spreadsheets and hoping the employee IDs match.

Frequently asked questions

What is people analytics in a dealership context?

Using the workforce data you already collect — hiring, onboarding, attendance, training, and performance records — to spot patterns early enough to act on them, rather than reporting on outcomes after the fact.

What is the best leading indicator of employee turnover?

There is no single one, but the earliest reliable signals are onboarding checkpoint completion, ramp velocity against a cohort average, and clustered unplanned absences.

Why is turnover rate a bad management metric?

It is a lagging indicator reported at too high a level and too late. It tells you what already happened across the whole group, so it cannot tell you where to intervene.

How much data do you need to start?

Less than most people think. Two metrics tracked consistently by store and by manager will outperform a dashboard project that never ships.