Most local marketing is measured badly or not at all
The common situation looks like this. Google Ads reports 40 conversions. Meta reports 25. The owner’s sense is that the phone rang maybe thirty times and perhaps a dozen became jobs. Nobody can reconcile those numbers, so decisions get made on instinct, and the budget goes wherever the last conversation pointed it.
This is not a reporting inconvenience. It is the reason money gets spent on things that do not work for months at a time.
Define conversions as business outcomes
The first thing to fix is almost always what is being counted.
We regularly find accounts where a conversion is a click on the phone number link — not a call that connected, just a tap. Or a visit to the contact page. Or a form submission counted three times because the tag fires on every page load after submission.
Google and Meta both optimise towards whatever you declare a conversion to be. If you tell Google that tapping a phone number is the goal, it will reliably find you people who tap phone numbers and do not call. The machine learning is good. It is just pointed at the wrong target.
So conversions get defined as things that matter: a call that lasted long enough to be a real conversation, a form submission that reached your CRM, a booked appointment.
Call tracking, because calls are most of it
For most trades the majority of inquiries are phone calls, and untracked calls are invisible to every report you look at.
Call tracking assigns numbers by source so you can see that the Google Ads number rang eighteen times, the Business Profile number rang thirty, and the organic site number rang nine. Duration thresholds filter out wrong numbers and hang-ups. Recordings, where you want them, also tell you how inquiries are being handled once they connect — which is frequently where the real problem turns out to be.
Analytics configured for a service business
Google Analytics 4 out of the box is configured for ecommerce. Its default reports are built around products, carts and purchases, none of which you have.
Configured properly for a service business it is genuinely useful: which pages lead to contact, which services get researched most, how traffic from each source behaves, where people drop out of the path. Configured by default it is a bounce-rate chart nobody acts on.
Attribution, honestly
No attribution model is correct. Someone sees a Meta ad, notices your van, searches your name a week later, clicks an ad and calls. Google will claim it. Meta will claim it. Both are partly right and both overstate.
We do not pretend to resolve this perfectly, because nobody can. What we do is measure platform-reported numbers against what actually landed in your CRM and what actually got booked, and report the gap rather than hiding it. When Meta claims 25 leads and 9 reached the pipeline, that is the useful fact.
The report
Once a month, in plain English:
- What you spent, by channel
- What it produced — leads, calls, booked work
- Cost per lead and, where the data supports it, cost per job
- What changed since last month and why
- What we are doing next, and what we need from you
If something is not working, it says so. A report that is good news every month is not a report, it is a sales document.