The five numbers worth tracking, and the ones to ignore
Most marketing reports are full of figures that cannot change a decision. Here are the five that can, and how to find them without buying anything.
- Track enquiries, conversion rate, cost per enquiry, average job value and where they came from. Everything else is commentary.
- Impressions, reach, followers and rankings are the four numbers most likely to appear in a report and least likely to change a decision.
- The most valuable question in small business marketing is still "how did you hear about us?", asked on every enquiry and written down.
- You cannot improve a conversion rate you have never measured. Most owners overestimate theirs by two to three times.
- A number is only worth tracking if you can name the decision it would change.
A monthly marketing report arrives. It says impressions are up 34%, reach is up 12%, the account gained 87 followers and the site ranks fourth for a phrase nobody in the business has heard of.
Every number is accurate. Not one of them helps anybody decide anything, which is the only reason to have a number at all.
If you cannot name the decision a number would change, stop collecting it.
The five numbers that matter
One: how many enquiries. Every phone call, form submission, email and message from a potential customer. One row each, with a date. This single number, tracked honestly for six months, is worth more than every dashboard you could buy — because it is the number every other one is trying to influence.
Two: how many became customers. Your conversion rate. Divide two by one. Almost every owner guesses this two or three times higher than it is, because the jobs won are memorable and the quotes that went quiet are not.
Three: what an enquiry cost. Everything spent on marketing in the month, divided by enquiries received. If you spent £1,400 and got 22 enquiries, an enquiry cost £64. That number is what makes every future spending decision straightforward.
Four: what a job is worth. Your average invoice. With three, it tells you the most important thing in your business: if an enquiry costs £64, one in four converts, and a job is worth £4,000, then £256 of marketing produces £4,000 of work. At that ratio the correct action is obvious and it is to spend more.
Five: where they came from. Not from analytics — from asking. Add it to the form and to how you answer the phone.
Why "how did you hear about us" beats analytics
Analytics tells you the last thing somebody clicked. It is often not the thing that made them buy.
A woman sees your van in Cobham. A fortnight later a neighbour mentions you. On Sunday she searches your name and enquires. Your analytics records that as organic search — or worse, direct — and your van and your neighbour get no credit at all.
This is not a flaw you can configure away. It is the fundamental limit of measuring one touch in a journey with five. Asking is the only method that reaches the parts analytics cannot see, and its answers are messier and considerably more truthful.
The numbers to ignore
- Impressions and reach. An advert shown 40,000 times to people who did nothing is not 40,000 of anything you can bank.
- Followers. An account with 2,000 followers and four likes a post is worse off than one with 200 and forty. The number describes an audience that has already stopped paying attention.
- Rankings for phrases nobody searches. Fourth for "premium bathroom installation specialists Surrey" is fourth for nothing. Ask for the search volume alongside every ranking you are shown.
- Bounce rate, mostly. On a local service site, a visitor who lands on your contact page, reads the number and rings you counts as a bounce. It is also a perfect outcome.
- Time on page. Longer is not better. Someone who finds the price in nine seconds and enquires has had an excellent visit.
What to do once a month
Twenty minutes, on the same day each month, with the spreadsheet open.
Count the enquiries. Count the ones that became jobs. Divide the marketing spend by the enquiries. Look at the "where from" column and see if anything has moved. Then ask one question: what would I do differently next month?
Most months the answer is nothing, and that is a legitimate answer. The value is not in constant adjustment — it is in noticing, in month four, that enquiries from one source have quietly halved while you were busy.
That is the real argument for tracking anything: not optimisation, but noticing. A business that measures five numbers monthly will spot a problem in six weeks. One that measures nothing spots it when the year-end accounts arrive.
Questions we get asked
What marketing metrics should a small business track?
Five: how many enquiries you received, what proportion became customers, what each enquiry cost, what an average job is worth, and which channel each enquiry came from. Those five let you decide where to spend next month. Anything that cannot change that decision is commentary.
What is a good conversion rate for a small business website?
For a local service business, 2% to 5% of visitors making an enquiry is a normal range, with well-built pages reaching higher. The useful comparison is not against an industry average but against your own figure last quarter — the average conceals enormous variation in traffic quality.
Why are impressions and reach not useful metrics?
Because they cannot change a decision. Knowing an advert was shown 40,000 times tells you nothing about whether to spend more on it. Cost per enquiry does. Any metric you cannot connect to an action is being reported to fill a page.
How do I track where my enquiries come from?
Ask, on every enquiry, and write it down. Add a "how did you hear about us" field to your form and a line to your phone script. It is unfashionable and it remains the most accurate attribution available to a small business, because it captures the journeys analytics cannot see.
Do I need expensive analytics software?
No. A cookieless analytics tool for traffic, your Google Business Profile for calls and direction requests, and a spreadsheet with one row per enquiry will answer every question in this article. The constraint is almost never the software.