Measuring marketing means joining what you spend to the enquiries, appointments and work that follow. Define each stage, record it the same way every time, and keep ad spend separate from total marketing cost. A booking is not a customer. Revenue is not profit. A good report helps you decide what to keep, dig into or drop — and shows plainly where the numbers are missing, late or too small to trust.
Agree what the words mean
- Enquiry
- One request from one prospective customer. Three calls about the same job is one enquiry.
- Suitable enquiry
- It meets your service, area and other criteria. Record the uncertain ones separately.
- Booking
- An appointment in the diary. Say whether it is a chat, a quote visit or paid work.
- Attendance
- It actually happened. Treat cancellations, moves and no-shows the same way every time.
- Customer
- Someone who bought the service, under your own definition.
- Revenue
- Sales from that work, on a stated basis — invoiced or collected, VAT handled consistently.
Keep the definitions still. If “suitable” changes meaning, note when. Otherwise a better percentage is just a new label.
Example: a small report
Invented figures to show the sums. Not Job Growth Lab results, not averages, not a forecast. They follow one batch of 30 enquiries to known outcomes, rather than mixing this month’s enquiries with older sales.
| Measure | Amount | Meaning |
|---|---|---|
| Advertising spend | £600 | Paid to the ad platform |
| Other marketing costs | £400 | Management and page costs in this example |
| Total marketing cost | £1,000 | £600 + £400, same accounting basis |
| Enquiries | 30 | Separate requests |
| Suitable enquiries | 18 | Met the criteria |
| Booked assessments | 12 | Appointments agreed |
| Attended assessments | 10 | Actually happened |
| New customers | 4 | Bought and completed work |
| Revenue | £2,400 | Sales from those four |
Ad cost per enquiry: £600 ÷ 30 = £20. Total marketing cost per enquiry: £1,000 ÷ 30 = £33.33. Both are correct. They answer different questions.
Per booking: £1,000 ÷ 12 = £83.33. Per attended visit: £1,000 ÷ 10 = £100. Per new customer: £1,000 ÷ 4 = £250. The stage you name changes the meaning.
Suitability: 18 ÷ 30 = 60%. Of those, 12 booked — 66.7%. Attendance: 10 ÷ 12 = 83.3%. Always say what you divided by.
Revenue ÷ ad spend is £2,400 ÷ £600 = 4. That is not four times profit. It ignores the other £400 and the whole cost of doing the work. Even £2,400 minus £1,000 is not profit until you count labour and materials.
Attribution gives credit, not truth
Attribution decides which click or channel gets the credit. Different tools use different models. Google describes it as assigning credit across the path a customer took. Google Analytics’ attribution guide.
A customer might see your advert, get a recommendation, search your name, then ring. A tracked link sees one slice of that. Asking “how did you hear about us?” helps, but people remember one step.
Never add up every platform’s claimed sales as if they were different customers. Keep one business-level record, remove duplicates, and keep an “unknown” column. An honest gap beats a confident guess.
Small numbers and slow decisions
With four customers, one more takes cost per customer from £250 to £200 on the same spend. That looks like a big win. It is one job. Check the pattern holds before you change anything.
A quote visit can turn into work weeks later. Mark recent enquiries as still open and go back to old batches. Monthly reporting suits spend and workload. It should never imply this month’s sales came from this month’s advert.
Add what the team knows: job value, how hard it was, repeat work, why good enquiries died. Revenue alone flatters big jobs that leave nothing behind.
Use the report to pick the next question
Lots of unsuitable enquiries? Check the offer and the targeting. Suitable ones not booking? Check replies, prices and availability. Bookings not attending? Check the appointment messages. Visits not turning into work? Ask the team, not the ad account.
Job Growth Lab offers reporting and review alongside the marketing. You confirm attendance, work won and what it was worth. Agree the definitions before reading the totals. Nobody can tell from a calendar entry whether a job went well.
- Are the stages and costs defined the same way each time?
- Can you follow individual enquiries to a known outcome?
- Are unknown sources and open decisions visible?
- What decision does this evidence actually support?
Sometimes the answer is to collect the missing outcomes, or leave a promising test alone. A good report can justify patience.