The Numbers Said the Marketing Was Underperforming
As we went through it we found two problems. Neither of them was something I thought was an issue until we dug a little deeper, and neither of them was the advertising.
I had been pushing a decent volume of traffic to a landing page built specifically for that location. So the first thing I wanted to see was the enquiry count in his CRM, because if the strategy was not working I needed to change direction quickly rather than keep spending into it.
We pulled up the CRM together and started filtering. The first number came back around two hundred enquiries for the month, but most of those were car sales records rather than new enquiries. Once we stripped those out we were left with about a hundred and sixty.
Then I looked at the sources. Walk-ins, service, and a pile of entries tagged as Facebook campaigns, which was odd because we had turned those off. What I could not find was any lift in website leads for that location. Nothing tagged as social either.
That did not match what I was seeing on my side at all.
The Enquiries Were Never Reaching the CRM
So we tested it live on the call. He went to the vehicle page as a customer would, filled in the enquiry form, and submitted it with a phone number he would recognise.
Two emails came through straight away. One confirmation to him as the customer, one notification to the dealership. Both correct, both fast.
Then I asked him to check the CRM. The lead was not there.
I ran a second one myself on a different vehicle to be sure. He got my notification email within seconds. Again, nothing in the CRM.
That was the moment it clicked for both of us. The form worked. The emails worked. The campaigns were delivering. But the join between the email and the CRM had stopped matching, so every website enquiry for that location had been landing in an inbox and going no further.
Here is the part that makes this hard to catch. Staff were still receiving the emails, so nobody had any reason to think something was broken. A silent failure with a visible workaround is the worst kind, because the workaround hides the failure.
Google Had Quietly Changed His Listing
While we were in there I checked his Google Business Profile, mostly out of habit.
The website button was not pointing at his website. It was sending people to a generic manufacturer stock locator. So anyone searching the dealership by name, finding the listing and clicking through was being routed away from his own site and into a page listing every dealer in the country.
He had audited that account recently, so this was not neglect. We opened the change log and it showed the platform had made the edit itself. His trading hours and social links had been altered too.
We corrected the URL on the call. It went into review, which takes about ten minutes, and we checked his other location while we waited. That one was fine.
It Turns Out This Is Not Rare
I assumed we had found something unusual. We had not. After the call I went looking, and the numbers are worse than I expected.
Demand Local’s February 2026 analysis puts it plainly: 14.1% of automotive sales leads are never logged to dealer CRMs at all. Not slow to be worked. Not badly followed up. Never recorded. They describe the gap between lead capture and CRM integration as pure waste, and that is exactly what we had been looking at.
The same analysis cites research that 43.2% of dealership website sales leads are mishandled, where mishandled covers missed calls, leads not logged in the CRM, lapsed follow-up and slow responses. Close to half.
Foureyes tracked 2.6 million leads across more than 1,150 dealerships in 48 markets through the first quarter of 2026 and found 11.7% of sales leads went on to buy, up from 10.2% the year before. Useful benchmark, but it only counts the leads that made it into a system to be measured. If one in seven never arrives, the real denominator is bigger than anyone’s report shows.
It is not only an automotive problem either. Sopro’s 2026 CRM research found 92% of businesses say valuable customer data sits outside their CRM. Validity’s work, cited widely through 2026, has 37% of organisations losing revenue as a direct result of poor CRM data quality.
The line that stuck with me most was from SuperOffice: they expect 40% of agentic AI CRM projects to fail or stall by 2028 because of data quality, not the AI. Every dealer I speak to wants to talk about AI. Almost none want to talk about whether the lead reaches the database first. You cannot automate follow-up on a record that does not exist.
And on the pressure side, a 2026 survey of more than 250 dealers found 67% say closing is harder now than a year ago. Some of that is the market. Some of it, on the evidence of that call, is leads never landing where anyone can work them.
Every System Was Honest About Its Own Job
Neither of these problems showed up as a failure anywhere, and that is what I keep coming back to.
The website confirmed the form submitted. The mail server confirmed the email sent. The parsing tool was waiting for a format it no longer recognised, so as far as it was concerned there was nothing to process. The CRM cannot report a lead it never received.
Four systems, four different vendors, every one of them reporting healthy. The failure was living in the gaps between them, and the gaps do not have a dashboard.
Same with the listing. No alert, no email, no flag. A platform I do not control changed something on an asset he owns, and the only way to find it was to go and look.
If we had not tested it on that call, the conversation would have ended with me changing the campaigns. New creative, different targeting, maybe a budget shift. I would have been optimising against a number that was never going to move, because the leads were arriving and simply not being counted.
That is the part that should worry any dealer benchmarking their performance. If your reporting is missing a share of your enquiries, then your cost per lead is overstated, your conversion rate is understated, and every decision you make off that data is slightly wrong in the same direction.
Test the Joins, Not the Tools
Integrations are not built once and left alone. A field gets renamed, a template gets updated, a vendor ships a change, and a mapping that worked for two years stops matching. Nothing announces it.
So this is what I run now, and what I would suggest to any dealer:
- Submit a real test enquiry through every live form monthly, and follow that one record all the way to the CRM
- Reconcile the counts across ad manager, website and CRM, and treat any gap as a fault to investigate rather than attribution noise
- Check your Google Business Profile monthly, including the website link, the hours and the change log
- Review who receives enquiry notifications whenever someone leaves, so leads are not going to an inbox nobody opens
- Write down which system owns which step, and who you call when a specific join breaks
Marketing gets blamed for a lot of things that turn out to be plumbing. Before you rewrite creative or move budget, prove a lead can make it all the way through. You cannot judge a campaign on results the reporting was never going to show you.