An offline sales match connects the people your marketing reached to the deals that closed in your DMS. You take a list of sold customers, match it to vendor lead, call, web and ad data, and see which sources touched buyers who actually bought. It is the only way to judge a vendor on sold units rather than clicks, form fills and cost per lead.
This guide explains how the match works, what data you need, where vendors overstate results, and how to read the output as a general manager or marketing director.
TL;DR
- An offline sales match joins sold-deal records (customer, VIN, date, gross) to vendor and ad data, so spend is judged against completed sales.
- A high match rate is not proof of ROI. Matching identifies who was exposed. It does not show the ad caused the sale.
- Lead-only reporting misses much of the picture, because many shoppers walk in without ever submitting a form.
- Ask every vendor to disclose match method, confidence, exclusions and how incrementality was tested.
- Run the match independently, on your own DMS data, over a 90 day window or longer.
Table of Contents
- What an offline sales match actually is
- Why lead counts understate and mislead
- What data you need before you start
- Match methods and what each one proves
- Match rate is not ROI
- Digital engagement is not the same as showroom results
- How to run a match you can defend
- A worked example with illustrative numbers
- Questions to put to every vendor
- Where the match does not reach
- Our perspective
- Get an independent read on your vendors
- FAQ
- Sources
What an offline sales match actually is
Most dealership marketing happens online, but the sale happens in the showroom, on the phone or at the F&I desk. The sale record lives in your DMS. The ad impression, click, VDP view or lead lives with a vendor, an ad platform or your CRM. Those systems rarely talk to each other.
An offline sales match closes that gap. In plain terms:
- You export closed deals from the DMS with customer identifiers, VIN, sale date, new or used, and front and back gross.
- You gather the activity data: leads, calls, chats, website sessions, ad interactions and third-party shopper data.
- You match records using email, phone, name and address, or hashed versions of them, and in some methods IP data.
- You report sold units, gross and cost per sale by source, not just lead counts.
The search term "VIN-level attribution" describes the same idea from the vehicle side. Each sale is tied to a specific VIN, so you can see which source touched the buyer and also which units got ad support, how long they aged, and what they grossed.
Why lead counts understate and mislead
Lead reports are easy to produce and easy to argue about. They also leave out a lot. An Auto Remarketing commentary on an on-the-lot study of close to 500 visitors across 30 dealerships found that nearly 70% of traffic was walk-ins who had never submitted an online lead. That research was done in Canada, so treat it as directional for U.S. stores, not a benchmark. The lesson holds: if your vendor scorecard counts only leads, it cannot see most of the shoppers your marketing influenced.
The Cox Automotive attribution white paper makes a similar point. It reports that, on average, 58% of dealership sales and service transactions in its study involved consumers who had visited Autotrader or KBB.com within four months of purchase. Those figures come from dealer data for July to December 2016, so they are historical context rather than a current standard. They do show why a form-fill count alone is a poor stand-in for influence.
The reverse problem exists too. Leads can be inflated by duplicates, bots, past customers and the same shopper submitting on three sites. A match against sold deals shows which lead sources produce buyers and which produce volume with no sales behind it.
What data you need before you start
A match is only as good as the records going in. Gather these before you talk to anyone:
- DMS sales export: deal number, buyer name, email, phone, address, VIN, stock number, sale date, deal type, gross.
- CRM export: lead ID, source, create date, rep, status, appointment and show data.
- Call tracking logs: caller number, tracking number, campaign, duration, and which vendor owns the tracking number.
- Vendor reports: each vendor's own claimed leads, sales and spend for the same period, so you can compare claims with the match.
- Spend by line item: invoices, not budgets. Include co-op and any bundled fees.
- Web and ad data: GA4 and ad platform exports, with UTM discipline noted where it exists.
Two hygiene checks save weeks. First, confirm that sales records carry a valid email or phone for a high share of deals. Second, confirm who owns the call tracking numbers. If a vendor owns them, the call data may not be yours to match.
Match methods and what each one proves
Not every match is the same. Vendors often use "match rate" to describe very different things, so ask which method sits behind the number.
| Method | How it links records | What it can show | Main weakness |
|---|---|---|---|
| Direct identity (email, phone, name, address) | Your own sold customers matched to leads, calls and CRM records | Which sources produced known buyers | Misses buyers who never gave contact data to a source |
| Hashed identifier match | Hashed email or phone compared across platforms | Ad exposure or interaction tied to a buyer | Depends on the platform's own user data coverage |
| IP or household match | Anonymized IP data linked to sold records | Site visits before a purchase | Shared networks and mobile IPs reduce confidence |
| Platform conversion import | Sales uploaded back into the ad platform | Which ad interactions led to real sales | Requires a captured click identifier at lead time |
| Store visit or geo methods | Device location near the store | Visit volume influenced by ads | Visits are not sales |
Google's own documentation describes the platform import route. According to the Google Ads API guide to offline conversions, you can import offline conversions to see which ads led to real-world sales, including sales completed by phone or through a sales representative. The workflow depends on consistent CRM capture and a usable ad-interaction identifier before upload. If your lead forms do not store that identifier, the sale cannot be sent back. Google also documents shop visit conversions, which cover auto dealerships among other physical locations. Visits are a useful middle signal, but they are not a sold unit.
For GA4, Google publishes guidance on growing offline sales with GA4 features, which helps you piece together on- and offline journeys. Use it as a way to improve your data capture, not as a replacement for a match against the DMS.
Match rate is not ROI
This is the point executives most often get wrong. A vendor says "we matched 76% of your sales" and the number sounds like proof. It is not.
The Cox Automotive study analyzed more than 2 million consumer transactions across 850 studies and reported a 76% average IP match rate. That figure describes how many records could be technically linked under that method. It does not show that every linked exposure caused a purchase. The method matched anonymized IP data, and it did not test whether the buyer would have bought anyway.
Keep three questions separate:
- Identity resolution: can the vendor connect the record to a real sold customer?
- Attribution: how does the vendor assign credit among the touches it found?
- Incrementality: would those sales have happened without the spend?
A vendor can be strong on the first and silent on the third. Retargeting is the classic case. It reaches people already shopping your inventory, so it matches to buyers at a high rate and looks excellent. Whether it created any of those sales is a different question. For more on credit assignment, see our practical guide to multi-touch attribution and our piece on dealership attribution modeling.
Digital engagement is not the same as showroom results
More digital activity does not automatically mean more sales. A 2022 SSRN working paper studied more than 20,000 U.S. dealerships over two years. It found that live showrooms increased monthly offline sales by 6.5%, while VR showrooms reduced automobile sales by 6.6%. Effects varied by vehicle type: live showrooms raised used-car sales by 8.6% and lowered new-car sales by 13% in that study. It is a working paper and not a final peer-reviewed article, so read it as evidence, not as a promised outcome for your store.
There is also peer-reviewed support for the broader idea. The study Online to Offline: The Impact of Social Media on Offline Sales in the Automobile Industry in Information Systems Research gives academic evidence that online activity can influence offline car purchases. The practical takeaway is to judge channels against completed sales, and to compare them by sold units, gross, inventory movement and customer experience, not by engagement volume.
How to run a match you can defend
A match is defensible when someone outside the vendor relationship can reproduce it. Here is a sequence that works for a franchise store.
Step 1: Fix the window
Pick a period long enough to cover your shopping cycle. We recommend 90 days of sales plus a lookback on lead and ad activity. A shorter window favors fast-closing used units and understates longer new-vehicle consideration. Our write-up on a 90 day vendor-agnostic sales match analysis walks through the timeline.
Step 2: Clean the sold file
Remove wholesale, fleet and employee deals unless you are measuring them on purpose. Standardize phone and email formats. Flag deals with no usable contact data and report them as a separate unmatched bucket, so nobody quietly drops them.
Step 3: Match in layers
Start with direct identity matching, then add hashed and platform methods. Record which method produced each match. A sale matched by a form submission and a sale matched by an IP guess should not carry the same weight in your scorecard.
Step 4: Deduplicate credit
One buyer often shows up in four vendors' reports. If every vendor claims the sale, the claimed total will exceed what you sold. Add up vendor-claimed sales and compare them with actual units. The gap is the first number to bring to your next vendor meeting.
Step 5: Calculate cost per sale and gross per dollar
For each source, divide invoiced spend by matched sold units. Then compare matched gross to spend. Show it next to the vendor's cost per lead so the difference is plain.
Step 6: Test, do not assume
Where you can, pause or hold out a channel in one market segment or time block and watch sold units. It will not be perfect, but it tests causation in a way that no match rate can.
A worked example with illustrative numbers
The numbers below are invented to show the arithmetic. They are not benchmarks and not client results.
Suppose a store spends $12,000 a month with a third-party lead vendor, which reports 300 leads, or $40 per lead. The match shows 18 sold deals tied to those leads by direct identity. Cost per matched sale is $12,000 divided by 18, or about $667. Now suppose another source costs $6,000, reports only 60 leads at $100 each, and ties to 14 sold deals. Its cost per sale is about $429.
By cost per lead, the first vendor looks 2.5 times better. By cost per sale, the second is cheaper. Add gross and the ranking can shift again. That reversal is the reason to run the match. Our Facebook ads benchmarks for dealers apply the same logic, comparing cost per lead with cost per sale on one channel.
Questions to put to every vendor
Use these in your next review. A vendor with a sound method answers them without drama.
- Which match method produced the figure, and what share of matches come from each?
- What is the confidence level, and how are low-confidence matches treated?
- Which sales are excluded, such as service, past customers or fleet?
- Did you test incrementality, and how?
- Who owns the tracking numbers, pixels and raw data?
- Can you share the matched records so our team can verify a sample?
If the answer to the last question is no, treat the report as a claim and not as evidence.
Where the match does not reach
A sales match has limits, and a good analyst says so. Buyers who use a different email or phone than the one in the sold file will not match. Cash-in-hand walk-ins with no prior digital footprint will not match. Brand and upper-funnel work, such as TV, shows up in lagged and indirect ways. We cover that problem in can you prove TV ads drive dealership sales. Service is a separate matching problem too: repair order value, not sale gross, is the outcome, as in our guide to service appointment attribution. Linking records end to end is the subject of our guide to closed loop attribution.
Our perspective
We built our work around one observation: dealership leaders are handed several reports that disagree, and every report comes from someone who benefits from the answer. A vendor's match is often a sales tool. We do not sell advertising, manage campaigns or take commissions from vendors, so the analysis does not depend on which channel wins.
Our view is that the sold-VIN record in your own DMS should be the referee. Vendors can bring their data to it, and the store keeps the result. We also think executives should expect a scorecard that separates what was matched from what was proven, because a defensible number survives a challenge from a vendor rep and a challenge from an owner.
The most useful outcome of a match is rarely a single winner. It is a short list of spend that cannot be tied to sales, and a clear question to ask about each line.
— AutoROIQ
Get an independent read on your vendors
If you want the match done without a vendor in the room, our Marketing Intelligence Review evaluates your vendors and channels against your own sales data and gives executive-level recommendations. We also offer ongoing strategic advisory for stores that want continued oversight. See the details on our Marketing Intelligence Review and advisory page, or read more on our automotive marketing intelligence insights page. For a broader plan, see our independent plan to improve automotive marketing ROI.
FAQ
What is an offline sales match in automotive marketing?
It is the process of linking closed deals in your DMS to marketing, lead, call and ad data. The output shows sold units, gross and cost per sale by source. It replaces vendor-reported lead counts with completed sales.
What is VIN-level attribution?
It ties a marketing source to the specific vehicle that sold, identified by VIN. That lets you see which units received ad support, how fast they sold and what they grossed. It is a more precise form of the same sales match.
Is a high match rate proof that a vendor's ads work?
No. A match rate shows how many records could be linked, not that advertising caused the sale. Ask the vendor how it tests incrementality. The Cox Automotive study's 76% IP match rate, for example, reflected linkage under one method, not causation.
How long should the match window be?
We suggest at least 90 days of sales, with a lookback on lead and ad activity. Shorter windows favor quick used-car sales and understate longer new-vehicle shopping cycles. Check the window against your own average days to close.
Can I upload dealership sales to Google Ads?
Yes, if your CRM captured the ad-interaction identifier when the lead came in. Google's documentation describes importing offline conversions, including sales completed by phone or with a sales rep. Without the identifier, the sale cannot be tied back to the click.
Sources
- Google Ads API: Manage offline conversions
- Google Ads Help: About shop visit conversions
- Google Analytics Help: Grow offline sales
- Cox Automotive: Transparency white paper on marketing attribution
- Auto Remarketing: How better marketing attribution can shift the course of the auto industry
- SSRN: Do Virtual Showrooms Drive Offline Brick-and-Mortar Sales?
- Information Systems Research: Online to Offline, the impact of social media on offline sales in the automobile industry