Last Tuesday, your store may have texted an equity offer to a customer who sold the vehicle 14 months ago, emailed his duplicate profile at an old address, and missed his wife—who still services a three-year-old SUV in your drive. The campaign report called two messages delivered and one bounced. Management saw weak engagement. What actually failed was the data.
Dealers have plenty of records. That gets confused with having reliable customer intelligence. Years of repair orders, sales history, web leads, email activity, declined-service records, and appraisal data can create a database with hundreds of thousands of rows. A large percentage may be technically real and operationally useless.
The Database Is Not the Customer
Most stores still treat the DMS customer number as a person. It isn't. It is an account created during a transaction, often under time pressure and by an employee trying to get to the next screen. The same customer may appear under a maiden name, a nickname, a spouse's phone number, and two email addresses. Another profile may combine a father and son because they share an address.
Vehicle ownership is even messier. A sold unit remains attached to the previous owner. A customer services a vehicle purchased elsewhere, but nobody connects that vehicle to the household's prior sales record. One spouse buys while the other schedules service. Lease maturity data sits in one system, appraisal history in another, and actual service behavior somewhere else.
I've seen this play out at stores from Phoenix to Pittsburgh. Marketing blames the offer. The BDC blames lead quality. Sales says service customers never respond. Meanwhile, the store cannot answer a basic question with confidence: Which customers can we reach today, and what vehicles do they actually own?
Five Fields Determine Whether a Record Can Make Money
I'd argue that record count is close to meaningless without a confidence score. A practical audit should grade each customer across five fields:
- Identity: Can duplicate profiles be resolved to one actual person?
- Reachability: Is there a current, permissioned phone number or email address?
- Vehicle: Does the customer still own the VIN associated with the campaign?
- Household: Are spouses, co-buyers, shared addresses, and multiple vehicles connected correctly?
- Timing: Is there a current event—service visit, mileage threshold, lease maturity, appraisal, or ownership change—that makes the message relevant?
A record scoring five out of five is marketable. A record with identity and an email address but no verified vehicle is merely contactable. Those are not the same thing, yet most campaign dashboards put them in the same denominator.
Bad Data Distorts More Than Marketing Metrics
Suppose a service-lane campaign identifies 144 apparent acquisition candidates and generates six conversations. The dashboard reports a 4.2% conversation rate. An audit then finds that only 100 records had a valid mobile number and credible current ownership. The actionable response rate was 6%, while 44 records never represented a fair opportunity.
That distinction matters because operators make budget and staffing decisions from the wrong denominator. A campaign may be killed despite producing solid engagement among valid records. Or a vendor may claim strong delivery while messages are reaching people who no longer own the target vehicle.
The same gap reaches the used-car department. If 30 legitimate service-lane owners per month are missed because of stale or disconnected data, apply your store's expected acquisition rate and replacement-channel premium. At an 8% close rate, that is 2.4 lost units. If buying comparable inventory elsewhere adds $1,200 per unit in fees, transportation, or additional exposure, the data leakage tax is roughly $2,880 a month—before counting weaker recon visibility or extra days-to-sale.
Use the Data Leakage Tax
The back-of-napkin formula is simple: invalid or missed candidates × expected conversion rate × replacement-channel premium. It will not satisfy an accountant down to the dollar, but it forces the conversation away from vague database hygiene and toward inventory cost.
Look, cleaning records once will not fix this. Customer data starts decaying again the moment the project ends. Every new RO, online lead, trade appraisal, returned email, and employee-created profile can either improve the record or split it further.
Build a Feedback Loop, Not Another Export
Start by choosing which system owns identity, which owns vehicle status, and how corrections flow back. Returned messages should update reachability. A customer saying, "I traded that truck," should trigger an ownership-status change rather than live forever in a BDC note. Household relationships need their own fields, not tribal knowledge held by one advisor.
Automation raises the stakes. A human caller may spot that a record looks wrong. An automated workflow will confidently send the wrong message at scale. Dealers using platforms like AutoRelay for service-lane SMS and vehicle acquisition still need suppression rules, ownership verification, duplicate handling, and a process for writing customer responses back into the operating record.
Pull a random sample of 200 customers contacted during the past 30 days. Check the five fields above and calculate the percentage scoring four or five. Then segment campaign results by confidence score. If high-confidence records materially outperform the full file, you do not have a creative problem. You have a data leakage problem—and now you can put a monthly dollar figure on it.
See how AutoRelay helps dealers acquire inventory from their own service drive → getautorelay.com