A (Customer Satisfaction Index) score is your 's measure of how satisfied your customers are with a sales or service visit, built from a survey the manufacturer or its research firm sends after the visit. Scales and what counts as "good" differ by brand and region. Your store's job is the experience the survey measures. The survey itself belongs to the OEM.
What is a CSI score?
is the manufacturer's report card on your store, written by your customers. After a customer buys a vehicle or has it serviced, the (usually through a research firm it hires) sends that customer a survey. The answers roll up into your CSI score for the period.
The dealer doesn't send the CSI survey, doesn't write the questions and doesn't pick who gets it.
OEMs usually survey sales and service separately:
- Sales CSI covers the buying experience: the salesperson, the negotiation, finance, delivery of the vehicle.
- Service CSI covers a repair or maintenance visit. This is the score a service manager and director live with month to month.
Service CSI questions usually touch the same parts of the visit, whatever the brand:
- Scheduling: how easy it was to get an appointment that suited the customer.
- The advisor: courtesy, listening, explaining the work and the cost.
- The facility: waiting area, cleanliness, shuttle or loaner.
- The repair: whether it was fixed right and ready when promised.
- Pickup: whether someone explained what was done and what was charged.
The scale is where brands part ways. Some OEMs report an average of all ratings. Others report the share of customers who gave the top rating, often called "top box". A raw score from a given brand tells you little about a store selling another brand, so comparing them side by side misleads.
Takeaway
Match each question on your last service report to a part of the visit. Each question then has an owner in the store.
What is a good CSI score?
Your defines it. In many programs, "good" is measured against your region or the national average for your brand, and the bar can move every period as other stores' scores move. A score that cleared the line last period may not clear it this period.
Your district or field rep can tell you the thresholds your program uses for incentives and recognition. Ask for the threshold, the comparison group it is measured against (region or national) and the period it covers, in writing, and ask how often they change.
An outside number is worth knowing. J.D. Power runs an industry study of service satisfaction, its U.S. Customer Service Index () Study. It shares the initials but is not your OEM's CSI. In that study, among owners of 1- to 3-year-old vehicles, when dealer service satisfaction is 950 or higher on a 1,000-point scale, 86% of mass-market and 88% of premium customers say they "definitely will" return to the dealer for paid service (J.D. Power 2026 U.S. Customer Service Index Study).
J.D. Power's study is an independent survey of owners across brands, published as an industry benchmark. Your OEM's CSI is the manufacturer's own survey of your own customers, and it is the one tied to your store's standing with the brand. The J.D. Power figure says nothing about where your OEM sets its line.
Takeaway
Email your field rep this week for the threshold, comparison group and period. The written reply is your program's definition of good.
How is a CSI score calculated?
The does the math, and the details are in your program's rules. Programs commonly use some version of these approaches:
- Average of ratings. Every answer is turned into points and averaged across questions and customers. Average score = sum of all ratings ÷ number of ratings.
- Top-box share. The top rating alone counts as a "hit." Top-box share = top ratings ÷ all responses. A customer who gives the second-highest rating counts the same as a customer who gives the lowest.
- Weighted questions. Some programs give certain questions, such as "fixed right" or "overall satisfaction," extra weight, so the total is not a plain average of every question.
Say a month's customers all give either the top rating or the next one down. Averaged, the score sits near the ceiling of the scale. As a top-box share, every next-one-down answer is a miss, exactly as the lowest rating would be. Same customers and same answers, yet the totals look nothing alike. That's why you need to know which method your program reports.
Sample size matters too. Your score for the period is built from the customers who answered. In a small store each response weighs heavily in that total, so a single unhappy customer can swing the month. In a large store the same response is diluted across many others. That is arithmetic, and it's why a small store's score can jump around from month to month while nothing changes in the drive.
Read the question-level results as well as the total. The total tells you how the month went. The questions point at the part of the visit that's costing you: an advisor who doesn't explain charges, cars that aren't ready when promised, a waiting area nobody has looked at in a year.
Takeaway
Before you compare months, learn which method your program reports. That tells you whether a swing came from the drive or from the arithmetic.
Why does CSI matter to the store?
At many , can affect:
- Incentive money. Some programs pay on CSI directly or make it a condition for other payments.
- Allocation. At some brands, standing on CSI can affect how much inventory or which models a store receives.
- Recognition and standing. Brand awards, dealer council seats and the store's reputation with the district team often lean on CSI.
The details vary by brand and program, and they change. Your dealer agreement and your OEM's program guide are the sources that count, and your field rep can walk you through them.
There is a business reason under the incentives too. According to Cox Automotive, 74% of buyers who returned to the dealer for service in the past 12 months are likely to buy from it again, against 44% of those who didn't. The J.D. Power finding above, for scores of 950 or higher, points the same way on paid service.
Where your CSI actually comes from
The score starts on the drive, in moments the customer remembers and the store often doesn't notice. In the 2025 Cox Automotive Service Industry Study, 45% of dealership service customers reported a frustration during their visit. Some of what frustrates customers shows up in the store's own records.
How to read it: find the stage of the visit, then read across from what the customer judges to the signal in your own systems, who owns it and the fix.
| Stage of the visit | What the customer judges | In-store signal it went wrong | Owner | The fix |
|---|---|---|---|---|
| Appointment | Could I get a time that suits me? Did anyone confirm it? | Customer arrives with no appointment on file; long hold times on the service line; repeated reschedules | or scheduling lead | Check the appointment log daily for no-shows and overbooked slots; confirm the day before; call back missed calls the same day |
| Write-up | Did the advisor listen? Did I understand the estimate? | Advisor note of a complaint at write-up; customer concern written vaguely on the ; estimate not signed | Write the customer's concern in their words on the RO; walk through the estimate before work starts | |
| Waiting | Was I told how long? Was the loaner or shuttle ready? | Loaner not ready at drop-off; shuttle missed; waiter in the lounge past the quoted time | Service manager (loaner fleet), porter or shuttle driver | Reserve loaners with the appointment; give waiters a time and update them before it passes |
| The repair | Was it fixed right? | RO on the same concern; repeat warranty work on the same issue; parts delay not communicated | Shop foreman, technician | Review every comeback weekly; road-test on drivability concerns; tell the customer about parts delays the moment you know |
| Pickup and explanation | Was it ready when promised? Did someone explain the work and the bill? | Promised-time overrun; final price above the estimate; customer waits at the cashier with no advisor present | Service advisor, cashier | Call before the promised time if it will slip; approve any overage with the customer before the work; advisor walks the customer through the RO at pickup |
| After the visit | Did anyone care how it went? Did the problem get fixed? | Complaint to the receptionist or BDC; negative comment in the store's own follow-up; unresolved concern noted on the RO | Service manager or director | Manager with authority calls back the same day; log the issue and the outcome in the ; close the loop |
What it tells you: every stage leaves a signal in your own records, and every signal has an owner in your store.
Text only customers who agreed to texts. Which customers your store may call, and how, is a question for your store's own counsel.
Each signal in the "In-store signal" column is a customer at risk, whatever any survey says. Treat it as a problem to fix for that customer this week.
Takeaway
Give each row's owner a signal to watch and a day to report on it, so a customer at risk reaches someone that week.
How to improve CSI scores
improves when customers stop leaving the drive with unfixed problems, and when the ones who do get a real response from someone who can help.
Fix the causes in the map
Pick the row in the map where your question-level results are weakest and work it for a month. If "ready when promised" is your weak question, the work is in promised-time discipline: realistic times at write-up, a midday check on every open , and a call to the customer before the time slips. If "explained the work" is weak, the work is at pickup.
Hear from unhappy customers and fix the problem
Some unhappy customers never say so at the counter. You find them through your own follow-up after the visit and a habit of reading the signals in the map. Start with the "After the visit" row. Keep the conversation on the car and the visit, and on what you'll do about it.
A manager with authority calls back the same day
When a customer is unhappy, the person who calls back should be able to fix it: the service manager, or the director for anything serious. Don't hand that call to the advisor the customer is upset with.
Open with something that shows you read the RO: "I see your car came back for the same noise you brought it in for last week. I'm sorry. I want to get that fixed." Then agree on what happens next.
Decide the escalation path before you need it: who takes the call, what they can approve (a loaner, a slot tomorrow, a goodwill adjustment), and where the outcome is logged in the so the loop is closed. The full version is in our service recovery playbook.
Coach advisors on patterns, not scores
Advisor coaching works when it is about what happened on the drive: estimates that weren't explained, promised times that slipped, customers who left without a walk-through. It stops working when it becomes a monthly argument about whose survey came back low. Look for patterns across many ROs, coach on the behavior, and leave blame out of it.
Loaner, shuttle and promised-time discipline
Transportation and time are where good repairs still turn into bad visits. Reserve loaners with the appointment, check the next morning's loaner list the afternoon before, and keep the shuttle schedule honest. Promise a time you can keep. If it slips, call the customer before they have to ask.
A weekly look at complaints and comebacks
Monthly CSI reports arrive after the fact. Get the service manager, shop foreman and lead advisor together every week to look at that week's comebacks, complaints and promised-time misses. You'll catch the pattern while the customers involved can still be helped. Keep it short and end every item with an owner.
Takeaway
Put the weakest row of the map on the weekly meeting's agenda with a name next to it. By month end the meeting can tell whether the fix held.
What not to do about the survey
The owns the survey. Your dealer agreement or program guide may set rules against influencing it. Check yours, and ask your counsel or field rep where the lines are.
How to read it: each row pairs something to do for the customer with a survey habit to drop.
| Do | Don't |
|---|---|
| Fix the customer's problem | Ask a customer for a particular score, or tell them what a "good" answer is |
| Call back unhappy customers through a manager with authority | Time any call, text or email around when the OEM survey might arrive |
| Coach advisors on patterns in the work | Coach customers, or advisors, on how to answer survey questions |
| Read the question-level results to find what to fix | Ask a customer to delay, skip or change a survey answer |
| Keep accurate customer contact details in the because it's good record-keeping | Change customer contact details to steer who receives the survey |
| Treat every customer the same, whatever you guess they'll say | Put employees, their families or staff vehicles through the survey |
| Offer goodwill when the store got it wrong | Offer a discount, gift or anything else tied to the survey |
What it tells you: anything in the right column is a habit to drop this month.
Those habits put the store's standing with the OEM at risk.
Takeaway
Walk your advisors and through this table at the next drive meeting, so everyone knows where the lines are.
Read your OEM's CSI report
These are the questions to ask of your own report, with the answer to look for. Where your program doesn't publish the answer, ask your district or field rep.
- Which scale does my program use? An average of ratings, a top-box share, or something else. It changes how to read your total: under top-box, a customer who was mostly satisfied counts the same as one who was not. If you don't know, ask your rep before you compare months.
- Which questions carry extra weight in my program? If "fixed right" or "overall" carries extra weight, that is where your effort pays. Ask your rep for the weighting.
- How am I doing against region and national? Look at the gap, and at whether the region itself moved this period.
- How many responses are in this period? In a small sample, each customer can move the month noticeably. Read a small month with care before acting on it.
- Which advisors' customers are in the sample? Many reports break results down by advisor. Use that to find patterns to coach. Never use it to single someone out over a single survey.
- Which questions moved since last period? Line the movers up against the map above and your and complaint log for the same weeks.
- Can I see individual responses and comments? It depends on your program. Ask your rep what you can see and what you are allowed to do with it, and don't guess.
Takeaway
Bring your answers for this month's report to the meeting, along with the next row of the map to work.
CSI, Google reviews and your own follow-up: how they differ
These get lumped together, but they measure different things.
- is the manufacturer's survey of a sample of your customers, with questions the OEM chose, scored on the OEM's scale. You don't run it.
- Google reviews are public comments from anyone who chooses to leave one, for any reason, on any part of the visit. You can respond, and you should respond well. If you want to understand what drives them, read bad Google reviews in the service department.
- Your own follow-up is the store's contact with customers after a visit: a call, a text, an email. This is what you design yourself: who gets it, what it asks, and what happens when a customer says something went wrong. If your own surveys come back empty, start with why customers don't answer surveys.
Your own follow-up is the piece that can help an individual customer this week. Build it to find and fix problems, and keep it separate from the OEM survey: it should not mention, preview or refer to the survey. For keeping customers coming back to the drive over time, see car dealership service retention.
Takeaway
Write down who owns your own follow-up and what happens after a bad answer. If nobody can say, give someone that job on Monday.
Where Lokam fits
Free your team from the follow-up list: they call back the customers who need it. Lokam AI is a post-visit follow-up product for US franchise dealerships. Your store's follow-up calls are made with Lokam's AI assistant, which asks how the visit went and flags unhappy customers from what they say. The people on your store's escalation list get the customer, a short summary and the main issue. It is not a replacement for your survey, not an inbound answering bot, not a , and not a replacement for your people.
FAQ
What is the Customer Satisfaction Index (CSI)?
The Customer Satisfaction Index is a manufacturer's measure of how satisfied customers are with a dealer's sales or service visit. The , or a research firm it hires, surveys customers after the visit and rolls their answers into a score for each store. Each brand sets its own questions, scale and thresholds. Some brands use other names for their sales or service survey.
What is CSI in a car dealership?
In a dealership, is the OEM survey score that tracks customer satisfaction with your sales and service departments. Service CSI covers scheduling, the advisor, the facility, the repair and pickup. The store doesn't send the survey, but the visit it measures is entirely the store's.
How can I improve my customer satisfaction scores?
Improve the visit. Fix the causes your question-level results point to, have a manager with authority call back unhappy customers the same day, and look at complaints and every week. Leave the survey itself alone.
Does CSI affect dealer incentives?
At many OEMs it can, through incentive payments, eligibility for other programs, allocation and brand recognition. The rules vary by brand and change over time, so your program guide and your field rep are the sources to trust.
Can a dealership see who filled out the survey?
It depends on the OEM program. Some programs share individual responses or comments with the store, and others report totals alone. Ask your field rep what your program shows and what you may do with it, and don't guess from the timing or the comment.
Is CSI the same as NPS?
No. CSI is your OEM's survey score built from questions about the visit, on the OEM's scale. (Net Promoter Score) is a single-question measure of how likely a customer is to recommend a business: the share of promoters minus the share of . Some OEM surveys include a recommend question, but they are different measures.
Before you go
What you can do with this
- Know where your store stands, from your own program's threshold.
- See whether the weakest part of the visit got fixed this month.
- Give each unhappy customer a manager who can fix the problem.
Where Lokam helps
- Free your team from the follow-up list: they call back the customers who need it.
- Every open issue has an owner, and a reminder when it sits.
- See feedback and open issues by advisor, so coaching starts from facts.