Take a per-advisor baseline before the training. At 30 days, check observed behavior: write-ups scored in the lane and inspections presented. At 60 days, read hours per customer-pay repair order and the close rate on recommended work. At 90 days, read CSI and the money. Adjust for car count, work mix and season throughout.
Start With a Baseline, Per Advisor
Pull the 90 days before the training by advisor and by month, then lock the file so nobody revises history. Define each measure once and keep the definition fixed:
- Hours per repair order, customer-pay: customer-pay labor hours sold divided by customer-pay ROs. Leave warranty and internal out; the advisor does not sell those.
- Effective labor rate: customer-pay labor sales divided by customer-pay hours sold. It shows whether hours were added by discounting.
- Inspection completion rate: inspections completed over eligible ROs. This one belongs mostly to the shop.
- Inspection presentation rate: inspections reviewed with the customer over inspections completed. This one belongs to the advisor.
- Close rate on recommended work: dollars or lines sold over dollars or lines recommended. Pick one basis and keep it.
- CSI: your manufacturer's survey score, by advisor if reported that way.
Add a behavior baseline. Have the manager score five write-ups per advisor on the lane observation scorecard you will use afterward. Our library's service advisor KPI guide covers the wider set of numbers.
Four Levels, in the Order They Show Up
Donald Kirkpatrick's four-level model, first published in 1959, is the usual frame for judging training. The quoted definitions are from Kirkpatrick Partners; the service examples are ours. Level 1, Reaction, is the degree to which people find the training "favorable, engaging, relevant, and supportive." Did the advisors think it was worth their time? Level 2, Learning, is whether they acquired "the intended knowledge, skills, attitude, confidence, and commitment." Can each advisor present an inspection in role-play without prompting? Level 3, Behavior, is the degree to which they perform "the critical behaviors" on the job and are "supported in and accountable for" doing so. Is the walkaround happening on live cars? Level 4, Results, is "the degree to which targeted organizational outcomes occur as a result": hours per RO, close rate, CSI, retention.
Kirkpatrick Partners advises planning from Level 4 backward. Choose the result you want before the visit, then name the behaviors that produce it, and measure those.
Expect each level to be smaller than the one before. A meta-analysis of 117 studies of behavior modeling training found effects were largest on learning, smaller on job behavior and smaller still on business results (Taylor, Russ-Eft and Chan, Journal of Applied Psychology, 2005). The studies covered workplace training in general, not service departments. Advisors who loved the trainer tell you about Level 1 only.
What to Read at 30, 60 and 90 Days
| Measure | Type | First readable | What distorts it |
|---|---|---|---|
| Observed write-up score | Leading | Day 30 | A different observer; scoring only quiet hours |
| Inspection presentation rate | Leading | Day 30 | Technicians completing fewer inspections |
| Close rate on recommended work | Lagging | Day 60 | Changes in what technicians recommend; price changes |
| Hours per customer-pay RO | Lagging | Day 60 | Work mix, an oil change special, season |
| Effective labor rate | Lagging | Day 60 | Discounting, a door rate change |
| CSI | Lagging | Day 90 or later | Survey timing, small counts, parts delays, comebacks |
| Retention | Lagging | Beyond 90 days | Visits are months apart |
Day 30: is the behavior in the lane?
Score the same number of live write-ups per advisor as in the baseline, on the same scorecard, at the same time of day. Check the two or three behaviors the training targeted. Then pull the inspection presentation rate.
Count the manager's work too: how many observations and one-on-ones actually happened. If the behaviors are not visible at 30 days, the financial numbers are unlikely to follow. Call the trainer, and look hard at the coaching cadence.
Day 60: the numbers closest to the advisor
With a full month of post-training ROs, read hours per customer-pay RO and close rate per advisor against that advisor's own baseline. Look at effective labor rate beside them. Hours that rose while the rate fell were bought with discounts.
Small counts mislead. A labeled illustration: an advisor who writes 250 customer-pay ROs in a month at 1.4 hours each has sold 350 hours. One week with four large repairs can add 30 hours and lift the figure to about 1.5 with no change in skill. So read two months together, and trust the number most when the observed behavior moved with it.
Day 90: CSI and the money
CSI lags for a plain reason: the survey goes out after the visit, the customer answers when they get to it, and the manufacturer reports on its own schedule. A score you see this month largely describes visits from weeks before.
CSI is also not purely an advisor score. The J.D. Power 2026 U.S. Customer Service Index Study, based on 51,228 owners of one- to three-year-old vehicles, combines five measures in order of importance: service quality, vehicle pick-up, service advisor, service facility and service initiation. The advisor is third. A comeback lowers the score no matter how good the write-up was, so read the advisor questions and comments separately where you can.
For the money, apply the change in hours per RO to your customer-pay RO count, then set the added labor and parts gross against the full cost: the fee, travel, advisor hours off the drive and the manager's time. The advisor training ROI calculator does that arithmetic and notes that sold hours only become gross if the shop can produce them.
Controlling for Car Count, Mix and Season
These habits keep the obvious errors out.
- Use ratios. Total labor sales rise with car count. Hours per RO and close rate do not depend on how many cars came in.
- Watch the mix. A coupon mailer or an oil change special fills the lane with one-line ROs and pulls hours per RO down even if every advisor improved. Split customer-pay ROs into maintenance and repair, or express and main shop, and compare like with like.
- Compare the same season. Set the post-training months beside the same months last year as well as beside the prior quarter. Tire, battery, air conditioning and brake work all move with the weather.
- Find a comparison group. An advisor who missed the training, an express lane that was not included, a sister store.
- Keep a change log. Date every other change: a new advisor, a technician who left, a labor rate increase, a new inspection tool, a revised menu, a pay plan change.
- Read each advisor. A department average can rise because the weakest advisor quit.
Use your own history as the yardstick. NADA Data 2025 reports total service and parts sales of $494 per customer repair order and an average customer mechanical labor rate of $186 an hour across franchised dealerships in 2025. Those are national averages for context, not a target for your store.
When to Conclude It Did Not Work
Give it 90 days, unless behavior is flat at 30. Then read the pattern before blaming the training.
- Advisors can do it in role-play, not on live cars. Nobody is observing. The fix is the manager's routine.
- Presentation rate rose, close rate fell. Advisors are presenting more and explaining less, or the findings arrive without measurements or photos.
- Close rate rose, hours per RO did not. Fewer or smaller recommendations are coming from the shop, or the mix changed.
- Hours per RO rose, CSI fell. Look at promise times and at how recommendations are being presented. More sold work in a shop without capacity means late cars, and pressure shows up in survey comments.
- Nothing moved, and advisors are writing more ROs a day than they can handle. That is a staffing problem. How many repair orders one advisor should write is the next read.
Trainer Claims, Including Ours
We have not found an independent study that measures what advisor training does to a dealership's service results. What exists is vendors' own case studies and ranges. Pinnacle's Service Advisor Training page prints result ranges too. Treat ours like anyone's: a claim to test against your baseline, not a forecast. The same page says advisors learn to manage their own numbers (RO average, labor hours per RO and CSI), which is where measurement should end up with any trainer. More on choosing and using training is in the Service Advisor Training Resource Center, and the library post on how advisor training affects CSI and retention covers the customer side.
Frequently Asked Questions
How soon should advisor training show in hours per repair order?
Give it a full month of repair orders after the visit before reading it, and two months before trusting it. One advisor's monthly count is small enough for a few large repairs to move the figure. Observed behavior in the lane should change sooner, within the first few weeks.
Why did CSI not move after the training?
It may be too early, since surveys describe visits from weeks ago. It may also be something outside the advisor's control: comebacks, long waits for parts, the condition of the car at pick-up. Read the advisor-specific questions and the written comments before deciding the training missed.
Who should score the write-ups after training?
The service manager, using the same scorecard the trainer used. Before the trainer leaves, the two should score a few write-ups together until their marks agree. A second observer, such as a fixed ops director, should be calibrated the same way.
Can a higher average repair order hurt customer satisfaction?
It can, if it came from pressure or from selling work the shop could not finish on time. It should not when recommendations are presented by evidence and need, declines are respected and promise times hold. Watch CSI comments and comebacks alongside hours per RO for exactly that reason.
Summary
Measure advisor training in sequence. Lock a per-advisor baseline of observed write-ups, inspection presentation, close rate, hours per customer-pay RO, effective labor rate and CSI. Check behavior at 30 days, the advisor's ratios at 60, and CSI and the money at 90. Use ratios, split the work mix, compare the same season and log every other change. Read the pattern before you call it a failure, and test every trainer's claims, ours included.
