Dealership CRM

Dealership Marketing Attribution: How Canadian Dealer Groups Track What's Actually Working

15+ years in local marketing; Google Ads certified; Shopify Partner.

Most dealers I talk to can tell me how many leads came in last month. Almost none of them can tell me which campaign generated the deal that grossed $4,200 front and back combined.

That gap, right there, is the dealership marketing attribution problem.

And it's not a small one. When you're spending CA$5,000 to CA$15,000 a month on Google Ads, AutoTrader, inventory feeds, email campaigns, and whatever your OEM co-op is funding, not knowing which dollars are producing real gross is a serious problem. You end up cutting the wrong budget lines. You keep paying for leads that don't close. You show up to your 20-group meeting with a dashboard full of impressions and clicks, and the dealer two rows down is talking about cost per sold unit.

This article is about fixing that. Specifically, I want to walk through how dealer groups, whether you're running two rooftops or eight, can build an attribution model that connects marketing spend to actual vehicle sales and gross. I'll cover what attribution actually means in a dealership context, where the data breaks down, and how to wire things together so the numbers you're looking at mean something. I won't be covering CRM software comparisons in depth here, that's handled in our complete guide to dealership CRM for Canadian dealers. And if you're trying to sort out how your DMS and CRM share data, the DMS-CRM integration breakdown covers that in detail.

What "Attribution" Actually Means for a Dealership (And Why It's Harder Than It Sounds)

Dealership marketing attribution means answering one question: which marketing activity caused this customer to buy a vehicle?

Simple to say. Genuinely hard to answer.

Here's why. A customer in Saskatoon sees a Facebook ad for a 2024 RAV4 in February. They don't click. Two weeks later they search "Toyota RAV4 lease Saskatoon," land on your dealer website, browse three VDPs, and leave without submitting anything. A month after that, their spouse calls your BDC after seeing a retargeted display ad. They book a test drive, come in, and buy. Your CRM logs the source as "phone call." Your Google Ads account gets zero credit. Your Facebook campaign looks like it's wasting money. Your BDC looks like a hero.

None of that is accurate. The actual path touched at least three channels before a human ever picked up a phone.

This is what the industry calls the multi-touch attribution problem, and it shows up in every dealership I've worked with. The channels that generate early awareness almost never get credit in a last-touch model, which is what most CRMs default to. So dealers cut awareness spend, double down on bottom-funnel sources, and wonder why their cost per lead keeps climbing.

The other piece that makes automotive attribution harder than, say, e-commerce is the offline close. The deal happens in the showroom, not in a browser. So even if you track every digital touchpoint perfectly, you still need to connect those touchpoints to a VIN-level sale in your DMS. That connection is where most dealer groups fall apart.

The gap usually shows up like this. The agency's monthly report says you got 8,200 leads across 30 rooftops last quarter. Your CRM says you sold 1,240 units in the same period. Of those 1,240 sold customers, fewer than 600 have a documented lead-source tied back to a specific marketing channel. The other 640-plus are attribution dark matter. You spent millions of marketing dollars to drive leads, but you can only honestly trace about half of your closed deals back to that spend.

What we recommend instead is post-sale attribution. On every delivery, the BDC, sales manager, or finance manager runs a 30-second survey with the buyer asking "how did you find us?" and writes the answer in the CRM. Do this on every delivery for six months and you have the only honest data set you'll ever have on which marketing channels are actually producing sold VINs. That's the data that proves to your CFO whether marketing is producing incremental revenue (sales you wouldn't have closed anyway) or just driving leads that would have come in regardless.

Add to that a 24-to-72 hour post-delivery follow-up call, same operator, asking "how's the car so far, anything we missed?" and you catch the problems that turn into 1-star reviews 2-to-7 days post-sale before they go public. Most Canadian dealer groups skip both of these steps. The ones who run them consistently can answer the CFO question that most VPs of Marketing dodge: did our $7M annual marketing spend pay for itself this year, or didn't it?

Where the Data Actually Breaks: The Three Attribution Gaps

I think of dealership attribution as having three distinct gaps. Each one requires a different fix.

Gap 1: The source field in your CRM is wrong.

This is the most common one. A lead comes in from AutoTrader, your BDC agent logs it as "internet," and now you have no idea which platform it came from. Or a customer calls in after clicking a Google Ad, and the CRM logs it as "phone-inbound" with no campaign data attached. Per the research on Canadian CRM usage, this kind of incomplete source logging is extremely common, especially in stores where BDC agents manually enter source data rather than having it flow in automatically from a call tracking or lead routing system.

The fix is call tracking with dynamic number insertion (DNI). Tools like CallRail or CallTrackingMetrics assign unique phone numbers to each campaign and channel, then pass that data into your CRM automatically. When someone calls from your Google Ads landing page, the CRM knows it was Google Ads, which campaign, and which keyword. You stop guessing.

Gap 2: Digital touchpoints don't connect to the DMS deal.

Your Google Analytics shows 400 VDP views on a specific used F-150 last month. Your DMS shows that truck sold. Were those 400 views related to the sale? Who knows. There's no bridge between the web session data and the closed deal unless you've built one.

This is where a customer data platform becomes important for dealer groups. A CDP can stitch together web behaviour, CRM activity, and DMS deal data at the customer level. Without that stitching, you're looking at three separate reports that can't talk to each other.

Gap 3: The CRM and DMS don't sync gross figures.

Most CRM attribution reports count leads, appointments, and sold units. They don't count gross. So when your agency shows you a campaign that "generated 40 sales," you have no idea if those 40 deals averaged $1,200 front gross or $4,800. A campaign that produces 20 deals at $4,500 gross is worth more than one that produces 40 deals at $900 gross, but if your attribution model only counts units, you'll optimize for the wrong thing.

The reason this gap exists is usually a weak DMS-CRM integration. If gross figures from your DMS aren't flowing back into your CRM, you're flying blind on profitability by channel. Getting that integration right is a prerequisite for any serious dealer attribution work.

The Math: What Attribution Should Actually Show You

Let me walk through what a real attribution calculation looks like for a Canadian dealer group, using numbers grounded in what's publicly available.

Say you're a two-rooftop group in Alberta. You're spending CA$6,000/month on Google Ads across both stores, and CA$3,200/month on AutoTrader listings. That's CA$9,200/month in paid marketing, not counting your OEM co-op contributions.

Last month, your CRM shows 80 internet leads total. 22 of those converted to appointments. 14 showed up. 9 sold.

Your blended cost per lead is CA$115 (CA$9,200 ÷ 80). Your cost per sold unit is CA$1,022 (CA$9,200 ÷ 9).

Now here's where attribution matters. If 6 of those 9 sales came from Google Ads and 3 came from AutoTrader, your cost per sold unit breaks down like this:

Google Ads: CA$6,000 ÷ 6 sales = CA$1,000 per sold unit
AutoTrader: CA$3,200 ÷ 3 sales = CA$1,067 per sold unit

On that surface level, they look roughly similar. But now add gross. If your Google Ads deals averaged CA$3,800 combined front and back gross, and your AutoTrader deals averaged CA$2,100 (because AutoTrader shoppers tend to be more price-sensitive and have already cross-shopped multiple dealers), your real return looks like this:

Google Ads: 6 deals × CA$3,800 = CA$22,800 gross on CA$6,000 spend. Ratio: 3.8x.
AutoTrader: 3 deals × CA$2,100 = CA$6,300 gross on CA$3,200 spend. Ratio: 1.97x.

That's the number that should inform your budget decisions. Not cost per lead. Not even cost per unit. Gross return on ad spend by channel.

I want to be clear: the gross figures I used above are illustrative. Your actual numbers will vary by market, vehicle mix, and negotiation culture. Pull your real average combined gross from your DMS by lead source, plug it into this formula, and you'll have something worth acting on.

How to Build Cross-Rooftop Attribution for a Dealer Group

Single-rooftop attribution is already hard. Cross-rooftop attribution for a group adds another layer, because now you have customers shopping multiple stores, campaigns running across different brands, and potentially different DMS and CRM configurations at each location.

Here's how I'd approach building this, week by week.

Month 1, Week 1-2: Audit your source data.

Pull every lead that came into your CRM in the last 90 days. Look at the source field. Count how many are logged as "internet," "phone," "walk-in," or other vague categories with no campaign detail. In my experience, groups that haven't done this before find that 40-60% of their lead sources are either blank or too vague to use for attribution. That number is your baseline. You're not going to fix attribution until you fix source capture.

Month 1, Week 3-4: Deploy call tracking.

Set up unique tracking numbers for each major channel: one for Google Ads, one for your AutoTrader profile, one for your website's organic traffic, one for each OEM campaign if you're running co-op. Make sure those numbers are passing UTM parameters and source data into your CRM automatically, not manually. If your CRM doesn't support automatic source population from call tracking, that's a gap worth flagging to your vendor.

Month 2, Week 1-2: Connect your DMS gross data to your CRM.

This is the hard part. You need your CRM to receive deal-level data from your DMS, specifically front gross, back gross, and the lead source that was on the original CRM record. If you're on VinSolutions and Dealertrack, this connection is more straightforward. If you're on a mixed DMS environment, say Reynolds at one rooftop and PBS at another, you're looking at two separate integration projects. The DMS-CRM integration guide walks through what those integrations typically involve and what they cost.

Month 2, Week 3-4: Build a cross-rooftop reporting view.

Once your source data is clean and your gross figures are flowing, you need a reporting layer that shows performance across all rooftops in one place. This can be as simple as a shared Google Looker Studio dashboard pulling from your CRM's API, or as involved as a proper business intelligence tool. The key fields you want by rooftop, by channel, and by campaign period: leads, appointments, shows, sold units, front gross, back gross, cost per lead, cost per sold unit, and gross return on ad spend.

For groups with four or more rooftops, this is where a CDP layer starts to make sense. It can unify customer records across stores so you're not double-counting a customer who shopped two of your locations before buying at one. If that's where your group is headed, the post-Cox Fullpath overview for Canadian dealer groups is worth reading before you commit to a vendor.

Canadian-Specific Wrinkles in Dealer Attribution

A few things make attribution harder for Canadian dealers specifically, and I want to name them directly.

CASL and consent tracking. Under Canada's anti-spam law, your implied consent window for email and SMS follow-up is six months from an inquiry and two years from a transaction. If your CRM is running automated email nurture sequences, those sequences need to stop when consent expires. That's not just a compliance issue, it affects your attribution too. If you're counting email-assisted conversions from contacts who are technically past their consent window, your email channel will look more effective than it is.

OEM co-op and approved vendors. A number of OEM co-op programs in Canada require you to use approved vendors for certain marketing activities. If your attribution vendor or call tracking tool isn't on the approved list, you may not be able to claim co-op reimbursement for that spend. Before you add any new tool to your attribution stack, check with your zone rep. This is especially relevant in Ontario under OMVIC rules, where advertising claims and pricing disclosures are tightly regulated, and the agency running your ads needs to understand those rules or you're the one who gets the complaint.

Per-lead platforms and margin math. I've talked to dealers who were spending CA$400 per lead on platforms like TrueCar and AutoTrader, on deals that grossed CA$800 combined. That's not a business model. Dealer attribution, done properly, makes these situations visible fast. When you can show that a first-party Google Ads lead costs CA$180 and closes at CA$3,600 gross, and a third-party marketplace lead costs CA$420 and closes at CA$1,100 gross, the budget conversation gets a lot easier.

Multi-DMS groups. If you're a group with stores on different DMS platforms, say CDK at two locations and PBS at another, your attribution data will have different latency and different field structures at each rooftop. Normalizing that data into a single reporting view takes real work. It's doable, but you need to account for it when you're scoping the project. Don't assume the data will just match up cleanly.

What Good Attribution Reporting Actually Looks Like

I want to close with a picture of what you're building toward, because I think a lot of dealers have been shown so many garbage dashboards that they've lost the plot on what a good one looks like.

A good dealership attribution report answers these questions, by channel and by rooftop:

How many leads did this channel generate this month?
How many of those led to booked appointments?
How many appointments showed?
How many sold?
What was the average combined gross on those deals?
What did we spend on this channel?
What's the gross return on that spend?

That's it. Seven data points. If your current reporting can't answer all seven for every channel you're spending money on, you have an attribution gap.

In my experience, groups that build this reporting and actually use it in their monthly budget reviews end up shifting spend toward channels with better gross return within two to three months. Not because the data tells them something shocking, but because they finally have a clear picture of what was already true. The money was already flowing somewhere. Now they can see where it was actually working.

The comparison between VinSolutions, DealerSocket, and Salesforce in terms of how each handles this kind of reporting is worth understanding before you pick a platform. That comparison lives in the CRM comparison for Canadian dealerships if you want to go deeper on which CRM gives you the best native attribution tooling.

3 Takeaways

Last-touch attribution is lying to you. If your CRM is crediting the source of the final touchpoint before a phone call, you're optimising for the wrong channels. The awareness and consideration touchpoints that drove the customer to call in the first place are invisible in that model.
Gross return on ad spend by channel is the number that matters. Cost per lead is a proxy. Cost per sold unit is better. Gross return on ad spend is the actual business metric. Build your reporting around it.
Cross-rooftop attribution requires clean source data, call tracking, and a DMS-CRM gross sync. Get those three things working before you add a CDP or a BI tool on top. You can't analyse data that was never captured correctly in the first place.