Recalls are your cheapest retention campaign. Prove it.

A recall repair is free to the customer, funded by the manufacturer, and scheduled by a person who is already in your database. It is the highest-intent service appointment in the industry, and most service directors cannot tell you what their last recall campaign actually produced beyond a percentage on an OEM portal.

Start with the uncomfortable numbers. NHTSA's April 2026 report to Congress puts the weighted average completion rate for vehicle-only recalls initiated between 2012 and 2022 at 65.8 percent. In NHTSA's analysis of 2,515 recalls, 22 percent finished below 75 percent completion. Roughly a third of the vehicles in a typical campaign never come back. Every one of those is a customer you already paid to acquire, standing in someone else's service lane.

What the manufacturer hands you, and what it keeps

The OEM owns the campaign, the mailed notice, and the completion report. You own the relationship, the schedule and the bay. That split is why recall measurement inside a dealership is usually thin: the number you are judged on is the manufacturer's completion percentage, while the number you control is appointments.

Two constraints worth knowing before anyone builds creative. NHTSA's guidance to manufacturers states that the DOT and NHTSA marks are reserved for the recalling manufacturer and should not be used by dealers; manufacturer logos are encouraged instead. And it explicitly warns against coupons attached to a free recall repair that read as solicitation for paid work. Free has to look free. Any paid-service conversation belongs after the recall repair is closed, not stapled to the notice that brings the car in.

The five numbers that replace a vague campaign review

Most stores report two things: the OEM completion percentage and a manager's impression of how it went. Neither tells you which touch produced which appointment. Measure in this order and the argument ends.

1. List quality. How many VINs in the campaign match a live customer record with a current phone, email and address. If owner match sits at 60 percent, you have a data problem disguised as a marketing problem, and no messaging strategy will fix it.
2. Contact-to-appointment. Booked appointments divided by owners reached, per channel: mail, email, text, outbound call, site banner. This is the only number that tells you where to spend.
3. Show rate. Booked against arrived. A campaign that books well and shows poorly is a scheduling and reminder failure, not a messaging one.
4. Close rate. A recall operation code on a closed repair order. The completion credit belongs to the manufacturer; the RO belongs to you, and so does the paid work it exposes.
5. Incremental lift. The gap between the contacted group and a holdout you deliberately left alone. Without a holdout of roughly ten percent, you can report a total but never a lift, which means you cannot defend the budget next quarter.

What today's tools actually give you

VIN lookup on your own site. NHTSA publishes a free public recalls API keyed by make, model and model year. Put a lookup on the service page, log the VINs that return open campaigns, and you are capturing first-party intent at the exact moment a driver is worried about their car. Caveat: federal data can lag a manufacturer campaign, and it does not know whether remedy parts are in stock. Treat a hit as a lead to verify, not a guarantee.

The OEM recall feed into the DMS. Useful only if the recall lands as an operation code on the repair order rather than a line in a note. If your advisors write recalls into free text, your campaign reporting is a survey of memory.

A campaign code per touch, and one landing page per code. Every channel gets its own code, its own tracking number, and its own URL parameter, all joined back to the appointment record. This is unglamorous and it is the entire difference between attribution and storytelling.

A contact preference field at write-up. NHTSA's guidance recommends asking at the time of service how an owner wants to be contacted in the event of a safety recall. Almost no store has this field, which is why recall lists get worked with a channel the owner never agreed to.

Consent discipline. A safety recall notice is not marketing. Your text reminder program still runs under TCPA consent rules, and a recall list is not a marketing blast list. Mixing the two is how a clean safety campaign becomes a class action.

Sequencing, not newsletters. NHTSA's guidance is that consecutive touches outperform touches at fixed intervals. Your automation should fire a sequence over days, not a monthly send that everyone learns to ignore.

Where recalls and CSI actually touch, and where they do not

Set expectations correctly with your service directors. JD Power's 2026 U.S. Customer Service Index measures satisfaction at franchised dealers among owners and lessees of one to three year old vehicles, fielded January through December 2025 with 51,228 responses. Industry average is 868, premium 886, mass market 865.

That population is warranty work. A recall campaign on a 2014 model year vehicle will not move your CSI score directly, so do not promise that it will. What it does is fill the pipeline of customers who later buy paid service, at a moment when they are standing in your drive and forming an opinion about your operation. The visit is still judged by the same behaviors the index measures.

Those behaviors are measurable, which is the useful part:

• The full script, not the hero behavior. Only 26 percent of customers report experiencing nine or ten of the top CSI key performance indicators. When all ten are met, satisfaction averages 979; when three are met, it is 632. Scores among customers who experienced seven or eight still sit 63 points below the full set. Compliance with the whole routine beats excellence at one step.
• Photo and video evidence. 64 percent of customers want photo or video documentation alongside inspection results, and only 26 percent of mass market customers and 44 percent of premium customers receive it. Advisor satisfaction runs highest where it is provided. This is the cheapest available CSI gain in the building.
• Time is the product. Mass market customers wait an average of 1.61 hours for maintenance work at a dealer, premium 2.46 hours, while 62 percent of comparable aftermarket visits finish under an hour. Appointment availability is a satisfaction lever, not a scheduling inconvenience.
• Direct-to-consumer expectations are already in your lane. Premium customers returning to a traditional dealer from a DTC brand score 855, 29 points below the mass market average. Mobile and valet service is becoming table stakes for that group.
• The retention economics justify all of it. At satisfaction of 950 or higher, 86 percent of mass market and 88 percent of premium customers say they definitely will return for paid service.

A thirty-day sequence a director can run

1. Pull open recall campaigns for VINs already in your service history. That subset converts best and it is the only list you fully control.
2. Audit owner match before you message anyone. Fix what is uncontactable first.
3. Attach one campaign code, one landing page and one tracking number to each touch.
4. Enforce recall operation codes on the RO. No code, no credit.
5. Hold out ten percent of the list so you have a lift number.
6. Add contact preference to the write-up process.
7. Report the five numbers weekly. Stop reporting the OEM percentage as if it were a marketing result.

On CSI, the short version

Send inspection photos and video while the job is running, not attached to the invoice. Update status at three fixed points. Explain the work performed at delivery, in writing. Return the vehicle cleaner than it arrived. Track the ten key performance indicators per repair order and coach the misses, because that is where the 979 and the 632 diverge.

Two warnings. Do not coach customers on how to answer the survey; that is a compliance exposure, not a strategy. And do not let your own follow-up survey stand in for the index. They measure different populations, and only one of them affects how the manufacturer funds you.

The short version

Recalls are paid work from customers you already own, and the industry leaves about a third of them on the table. Measurement is not exotic: a clean VIN list, current contact data, a campaign code per touch, a recall code on the RO, and a holdout. CSI is not a mystery either; it is a checklist executed consistently while the car is in the building. Neither one requires new technology. Both require the data to be joined, and joined data is where dealerships keep losing.

The report page applies that discipline to the rest of the in-house list, ranking sales and service campaigns by appointments per 1,000 delivered. A recall that books a bay and a mailer that books nothing should not share a row.