How to measure your marketing: the numbers that actually matter

The numbers that actually matter for marketing are cost per lead, cost per customer, repeat purchase or retention rate, and some form of satisfaction score like NPS. Ignore the vanity metrics: follower counts, impressions, traffic for its own sake. If you get the four numbers above right, you will know which marketing is worth doing and which is not, and that knowledge is worth more than most of the marketing itself.

This article walks through each metric, how to set up the tracking behind it as a UK small business, and how to avoid the common measurement mistakes that lead owners to spend money on things that are not actually working.

Why measurement matters more than the marketing itself

Most small business owners underspend on measurement. They will commit a few thousand pounds to a Google Ads campaign, a website refresh, or a content programme, and leave the tracking patchy or absent. The result is they cannot tell, six months later, whether the work earned its keep. So they either keep doing it on faith or cut it on a hunch. Both are expensive.

Proper measurement does not have to be complicated. A small business can get to reliable numbers with a spreadsheet, Google Analytics 4, a basic CRM, and a discipline of asking every new customer where they came from. Total time to set up, perhaps two days. Value over five years, enormous.

The four numbers that matter

Cost per lead. The amount you spend to generate one enquiry. Take the full cost of a channel (ad spend, agency fees, content production, time at an honest hourly rate) over a period, divide by the number of leads that channel produced. A plumber spending £500 a month on Google Ads who gets 20 enquiries has a cost per lead of £25. The number on its own is neutral. Context decides whether it is good.

Cost per customer. The amount you spend to actually win one customer, not just one enquiry. This includes every channel together plus the cost of the lead to customer conversion process (time, tools, any offers you run). If you generate 20 leads at £500 and close 5 of them into customers, your cost per customer is £100. This is the number most worth tracking over time because it tells you whether your marketing and sales combined are sustainable.

Repeat purchase or retention rate. The percentage of customers who buy again or continue to buy within a defined window. For service businesses, this might be "percentage of one-off customers who booked a second job within twelve months". For subscription or ongoing services, it is the retention rate. For ecommerce, it is typically "percentage who made a second purchase within ninety days". This number tells you whether what you deliver is worth the first cost you paid to acquire the customer.

NPS or another satisfaction measure. Net Promoter Score asks one question: "how likely are you to recommend us, on a scale of zero to ten". Add up the promoters (9 or 10) and the detractors (0 to 6), and subtract detractors from promoters to get a score between minus one hundred and plus one hundred. Anything above thirty is good. Anything above fifty is excellent. Satisfaction is a leading indicator of future growth. Unhappy customers do not return and do not refer.

The vanity metrics to ignore (or at least demote)

Website traffic without context. Ten thousand visits a month that produce no enquiries is not a success.

Follower counts. A thousand followers who never engage with you are worth less than fifty customers who know you.

Impressions and reach. These measure how many times something was shown, not whether it did anything useful.

Open rate on emails. Apple Mail preloads images, which skews open rates. Click through and conversion rates are more honest.

Keyword rankings by themselves. Being number one for a term nobody actually searches for, or that does not bring enquiries, is worthless.

These metrics can be useful as secondary indicators. They are not useful as primary ones. If they improve but cost per customer does not improve, nothing useful has happened.

Setting up the tracking

Google Analytics 4. Free, and essential. Install it on every page of your site with a proper events setup. At minimum, track form submissions, phone clicks, WhatsApp clicks, and any booking or checkout actions as conversion events. Link it to Search Console and to any ads accounts. If you want a sharper cut of which numbers matter most, our piece on small business analytics and what to track trims the list further.

UTM parameters on every marketing link. UTMs are tags you add to the end of a URL (?utm_source=newsletter&utm_medium=email&utm_campaign=october) that let you see exactly which link brought a visitor in. Use them on every email, every paid ad, every social post where practical. Google has a free Campaign URL Builder that creates them for you.

Google Tag Manager. If the technical setup is tolerable, Tag Manager gives you a cleaner way to manage tracking. For a very small site, direct GA4 installation is fine.

A simple CRM. Even a well organised spreadsheet will do to start. Record every lead, the source it came from, the date, and the outcome (won, lost, reason). HubSpot has a free tier that covers this. Pipedrive, Capsule, and Zoho are all reasonably priced. Do not try to implement Salesforce as a five person business, the tool will eat your life. Our guide to choosing a CRM for your small business goes deeper on which fits which setup.

Ask every customer. The highest quality attribution data comes from a simple question at the point of first contact. "How did you hear about us." Train whoever answers the phone to ask and record the answer. This catches word of mouth, referrals, and offline sources that digital analytics will never see.

The weekly dashboard

Once a week, spend fifteen minutes reviewing a simple dashboard. It can live in a spreadsheet.

Leads by source this week. Ad spend this week by channel. Cost per lead by channel. Enquiry to customer conversion rate this month. Cost per customer this month. Any customer complaints or NPS responses.

A fortnightly habit is fine for most businesses. Do not obsess daily. The noise is too great and decisions made too fast are usually wrong.

How to know which channel is worth more

Once you have cost per customer by channel and the average lifetime value of a customer, the decision almost makes itself. Any channel with cost per customer well below lifetime value is a channel to scale up. Any channel with cost per customer above lifetime value is a channel to fix or cut.

Lifetime value is worth calculating. For a service business, it is typically average sale value multiplied by the average number of purchases per customer per year, multiplied by the average number of years they stay. A garden maintenance customer paying £500 a year for six years is worth £3,000. If your cost per customer for that kind of buyer is £400, the channel producing them is extraordinarily valuable.

Attribution is imperfect, accept it

Attribution will never be perfect. A customer might see your Instagram post in January, click a Google ad in March, read a blog post in April, and finally phone after seeing your van on the high street in May. Your data will credit one of those touches, usually the last. The truth is they all mattered.

Rather than trying to solve attribution perfectly, accept that it is directional. Look at trends over months, not exact numbers on individual weeks. If an entire channel shows zero attributed revenue for six months straight, it is probably not working, whatever the soft arguments about brand building.

Common measurement mistakes

Not tracking at all and running on gut feel. Guts are often wrong, especially when the owner is the one running the ads.

Over-tracking. Installing twenty tags, getting confused about which number is right, and giving up. Start simple. One source of truth per metric.

Measuring only online activity. If half your customers come from word of mouth and referrals, but your dashboard only tracks digital channels, you are seeing half the picture.

Confusing correlation with causation. "We started doing X and sales went up" is not proof that X caused the rise. Use test periods and comparable controls where you can.

Judging channels too early. Some channels, like content, SEO, and email, compound over months. Three weeks of data is too little to judge. Three months is the minimum, and for SEO and content, twelve months is more realistic.

Not firing things that are not working. The flip side of the point above. If, after enough time, a channel is clearly not producing, stop. Faith is not a marketing strategy.

What to do if numbers look bad

When the numbers look bad, avoid the temptation to immediately cut everything. Work through a short diagnosis first.

Is the tracking broken. Check that conversions are being recorded properly. A surprising amount of "bad performance" turns out to be missing tags.

Is the landing experience broken. A working ad campaign sending traffic to a slow, confusing page will look like a bad campaign. Fix the page first.

Is the product market fit the real problem. If nothing converts, no matter the channel, the offer or positioning might be the issue, not the marketing.

Is the sales process broken. Great leads lost because nobody returned the phone call for three days will drag down every channel's numbers, which is one of the strongest arguments for automating your follow-up, covered in how to automate your small business.

Only after checking these should you conclude that a specific channel does not work for your business.

A realistic measurement plan for a small business

Week one: install GA4 and Search Console, set up conversion events for forms and phone calls, start using UTMs on every campaign, begin asking "how did you hear about us" on every enquiry.

Week two: set up a simple lead tracking spreadsheet with source, date, status, and outcome. Calculate last quarter's cost per customer by channel as best you can from existing records.

Month two: review weekly for the first month to get used to the numbers. Look for obvious problems or opportunities.

Month three: calculate lifetime value for typical customer segments. Compare to cost per customer by channel. Make one decision based on the data (scale, change, or cut a channel).

Ongoing: review fortnightly. Recalculate core numbers monthly. Once a year, revisit lifetime value and the overall channel mix.

Marketing stops being guesswork once you have these habits. It starts being a set of investments you can evaluate and adjust.

If you would like help setting up tracking, or a review of where your current numbers are leaking, try our free audit or book a short call.