Marketing measurement has a bad reputation among people who run local businesses, and it has earned it. Reports arrive full of impressions and engagement rates that answer no question anyone asked. Meanwhile the thing you want to know — is this worth the money — is genuinely harder to answer than anyone admits. It is still worth answering badly rather than not at all.
The four numbers that matter
Most local campaigns can be judged on four figures, tracked over the same period each month.
Enquiries. How many people got in touch. Count every channel: calls, forms, texts, messages. One total.
Suitable enquiries. How many of those were work you actually want, in an area you cover, at a scale you do. This is the number most reports leave out, and the one that separates a busy phone from a good month.
Appointments. How many became a booked visit or call that happened.
Won work and its value. How many became a job, and what those jobs were worth.
Four numbers and the money spent give you a cost per enquiry, a cost per appointment and a cost per job. Everything else is detail underneath those.
What to ignore, and why it is in the report
Impressions. How many times an ad appeared. It tells you the ad ran. It does not tell you anyone looked.
Reach and engagement. Useful to a brand advertising a fizzy drink. Not useful when you need six bookings a week.
Click-through rate on its own. Worth watching as a diagnostic — a collapsed rate means the ad stopped matching the search — but a good rate leading to no enquiries is not success.
Ranking position. A position is not a visit, and a visit is not an enquiry.
These appear in reports because the platforms produce them automatically and because they usually go up. A number that always rises is comfortable to put in front of a client.
Why the numbers never quite agree
Attribution is the attempt to say which marketing caused which enquiry, and it is imperfect for ordinary reasons. Someone sees your ad on a phone at work, searches your name at home that evening on a laptop and calls you two days later from the van. Which of those gets the credit?
Platforms each claim what they can see, so their totals overlap and add up to more than the enquiries you actually received. Browser privacy settings and cookie consent choices mean a share of activity is never linked up at all. Google explains its own modelling in the Google Ads Help documentation.
The right response is not to give up. It is to treat platform figures as directional, and to treat your own count of enquiries and jobs as the truth. If the ad account claims forty conversions and you had eleven conversations, believe the eleven.
Example: the month that looked bad
An illustrative scenario. A bathroom fitter is shown a report: enquiries down from 22 to 17. The instinct is that the campaign is failing.
The rest of the figures say otherwise. Of the 22 the previous month, 9 were suitable — the rest were tiny repair jobs and areas outside his patch. Of the 17, 12 were suitable, because the ads and the qualifying questions had been tightened deliberately. Appointments went from 6 to 9. Won work went from 3 to 5, at a higher average value.
Fewer enquiries, more money, and less of his time wasted. A report showing only the enquiry total would have led to exactly the wrong decision. The figures are invented to make the point, but the shape is common enough to be worth watching for.
Ask where the enquiry came from — carefully
The simplest measurement tool is a question: how did you hear about us? Ask it early, in a natural way, and record the answer somewhere consistent.
It has known flaws. People misremember, say “Google” for anything on a screen, and credit the recommendation from a neighbour rather than the ad that made them look you up. Used alongside platform data, it is still a useful cross-check — particularly for catching sources that no tracking can see, like a van sign or a word-of-mouth referral.
A call answered by a real person can ask it properly. A follow-up system can ask it too, as long as it does not read like an interrogation before the customer has been helped with anything.
How to review a month honestly
Set the review up once, then repeat it the same way. Changing the measurement every month is how campaigns get judged on noise.
- Are you counting enquiries from every channel, or only the ones that are easy to count?
- Do you separate suitable enquiries from all enquiries?
- Is the period long enough for your volume — a fortnight of data for a business with six enquiries a week proves nothing?
- Do you compare against the same period, not against your best ever month?
- When platform figures and your own count disagree, do you go with your own count?
- Did anything else change — a price rise, a holiday, a storm, a competitor opening?
- Can you see cost per suitable enquiry and cost per won job, not just cost per click?
- Are decisions written down, with the date, so next quarter you can tell what you changed?
Honest measurement will sometimes tell you the campaign is not working. That is what it is for. A reporting approach that can never produce bad news is not measurement.