September 30, 2026 · 5 min
How Do You Know Which Marketing Channel Is Bringing Your Leads?
Someone calls, you have a good conversation, they become a client, and a week later you could not say with any confidence what made them pick up the phone. You have a website, a Google listing, a few referral partners, maybe some emails going out, and the honest answer to which one is working is that you do not know. So the budget goes wherever it went last year.
The direct answer is that you find out by asking, and by writing the answer down in the same place every time. Two questions, asked of every new enquiry, will tell you more in six weeks than any analytics dashboard: how did you hear about us, and what made you get in touch today. Add one column to wherever you keep your enquiries, record the answer in the person's own words, and review it monthly. That is the whole method, and it beats guessing immediately.
Why does your instinct get this wrong?
Because you remember the leads that were memorable, not the leads that were typical.
The client who mentioned your podcast appearance sticks in your mind. The eleven people who quietly found you in a search and never mentioned it do not, because there was nothing to mention. So you end up overweighting the vivid channel and underweighting the boring one that actually feeds you.
There is a second problem. The last thing someone did before contacting you is usually a search for your name, which makes almost everything look like it came from Google. The search was how they found your phone number. It was not what made them want to call.
What is the simplest way to start?
Add one column to your enquiry list, whether that list lives in a spreadsheet, a notebook or a CRM. A CRM, if you have not used the term, is just software that keeps a record of every person who has contacted you and what happened next.
Then do these four things.
- Ask every new enquiry how they heard about you, in the first conversation, before you start talking about the work
- Write down what they say in their own words, not in a category you invented
- Ask a second question: what made you get in touch now
- Once a month, read the column top to bottom and group the answers yourself
The reason for recording their words rather than picking from a dropdown is that the useful answers are the ones you did not anticipate. Someone saying "my accountant mentioned you" tells you to go and look after your accountants. A dropdown labelled Referral would have hidden that.
What is a UTM, and do you need one?
A UTM is a short piece of text added to the end of a link that tells your website analytics where the click came from. If you send a newsletter and the link inside it ends with something like utm_source=newsletter, then anyone arriving through that link is counted separately from someone who typed your address in.
You need them for anything you send out yourself: emails, social posts, a link in a directory listing, a QR code on a leaflet. They cost nothing and they turn one lump of traffic into named sources.
You do not need them to get started, and they will not tell you about the referral your accountant made, the person who saw you speak, or the client who has been meaning to call for four months. Most small businesses get further faster with the two questions than with the tracking, and then add the tracking once the questions show which channels are worth measuring precisely.
What about leads that came from more than one place?
Most of them did, and that is fine. Someone hears your name from a colleague, looks you up, reads two pages, forgets, sees a post six weeks later, then searches your name and calls.
You do not need to divide the credit between those touches. You need to know which ones were necessary. The second question, what made you get in touch now, is the one that separates them: the colleague is why they trusted you, the post is why they called this week.
Record both when you get both. In a small business with a handful of enquiries a month, two words in a spreadsheet cell is enough. Do not build a model.
How long before the numbers mean anything?
Longer than feels reasonable, and less long than you fear.
If you get five to ten enquiries a month, give it two to three months before you draw any conclusion. Below about twenty data points you are looking at noise, and one unusual month will send you chasing a channel that was a coincidence.
What you can act on sooner is the absence of something. If you have sent four hundred emails and not one enquiry has mentioned an email, that is informative after a few weeks, not months. A channel producing nothing at all is easier to read than a channel producing a little.
What do you do once you know?
Three things, in this order.
Protect the channel that is already working. If half your clients come through two referral partners, that is not a finished job, it is your most fragile asset. Deliberate, scheduled contact with those partners beats a new channel.
Fix the channel that is nearly working. A source producing enquiries that never convert is usually a message problem or a fit problem, not a volume problem.
Then, and only then, stop the channel producing nothing, and put that time into the first two. Most owners do this in reverse and start something new while the thing that works goes unattended.
The Bottom Line
You find out where your leads come from by asking every one of them and writing it down consistently, not by buying a dashboard. Use two questions: how did you hear about us, and what made you get in touch today. Record the answers in their words, in one place, and read them monthly. Add tracked links for anything you send out yourself. Wait for about twenty enquiries before you conclude anything, then spend your attention protecting what already works before chasing what does not.
This is the kind of record keeping that gets skipped when nobody owns it, which is why it is one of the first things we build into the systems we run for clients at ViralFusion, and it stays a plain list you can read yourself.