How to track where your leads actually come from
Almost every small business owner can tell you how many enquiries they got last month. Very few can tell you which channel produced them. That gap is expensive, because it means the advertising budget is allocated by feeling rather than by evidence, and the channel that quietly brings the best clients often gets cut first.
Why asking how they heard about you is not enough
The traditional answer to this problem is a dropdown on the contact form asking how the visitor found you. It is better than nothing, and it is also unreliable in three specific ways.
People genuinely do not remember. Someone saw an Instagram post in June, searched your company name in September and arrived through Google. They will answer Google, and they will be sincerely wrong.
People pick the first option. Dropdowns collect a suspicious number of answers matching whatever sits at the top of the list, which tells you about the form rather than about the customer.
And most visitors skip the question entirely if it is optional, which leaves you with data from the minority who happened to be patient that day.
None of this means you should remove the question. Asked as a free text field near the end, it captures the things analytics never sees: a recommendation from a friend, a van with your logo, a conversation at a trade fair. Use it as a supplement, not as your source of truth.
The reliable part of the picture comes from the visit itself, and that is what the rest of this article sets up.
UTM tags: the setup that fixes most of it
A UTM tag is a piece of text added to the end of a link that tells your analytics where the click came from. It takes minutes to set up and it is the single highest value change on this list.
There are three parameters that matter and two that rarely do. Source is where the click came from, medium is the type of traffic, and campaign is which specific push it belonged to.
- Source: google, facebook, instagram, newsletter, partner site name.
- Medium: cpc for paid clicks, email, social, referral, qr.
- Campaign: a short name you will recognise in six months, such as autumn-promo or new-service-launch.
Tag everything you control and nothing you do not. Newsletter links, social profile links, paid ads, QR codes on printed material, links in a partner directory, a signature link in your quotes. Never tag internal links between your own pages, because that overwrites the real source and makes your own site look like the channel that generated the lead.
Keep the naming consistent and lowercase. Analytics treats Facebook and facebook as two different sources, and a year later you will be looking at a report split across four spellings of the same channel.
The most commonly forgotten case is printed material. A QR code on a flyer or a vehicle can carry a tagged link, and it is the only realistic way to learn whether that spend produced anything.
Capturing the source in the form itself
Analytics tells you a channel produced ten conversions. It does not tell you which ten emails those were, which is exactly what you need when you want to know whether those leads turned into paying customers.
The fix is a hidden field in your contact form that records the traffic source and the landing page, then arrives with the enquiry email. It costs one developer hour and it changes what the data can answer.
Store four things: the referring source, the UTM values if present, the first page of the visit, and the page the form was submitted from. The landing page alone is often the most useful of the four, because it shows which article or service page did the persuading.
This also survives the multi visit problem better than analytics dashboards do. If you keep the first recorded source in a cookie, an enquiry from someone who arrived from an ad in July and returned directly in September still carries the ad as the origin.
Then close the loop in whatever you use to track clients, even if that is a spreadsheet. Add two columns to every enquiry: source and outcome. After three months you can answer the only question that really matters, which is not which channel brings the most enquiries but which one brings work you actually want.
That distinction is usually dramatic. It is completely normal for one channel to produce three times the volume and a fraction of the revenue, and no dashboard configured out of the box will tell you that.
Phone calls and messaging apps
For a service business, a large share of enquiries never touch the form. People call, or they write on WhatsApp or Messenger, and those conversations are invisible to standard analytics.
The cheapest partial fix is to make the phone number on the site a click to call link and count the clicks as a conversion. It undercounts, because desktop visitors write the number down instead of clicking, but it gives you a trend and a rough ratio between channels.
A second option is a separate phone number for a specific campaign, printed only on that flyer or used only in that ad. Anything ringing that number came from there, with no software involved.
Call tracking services that swap the displayed number per visitor exist and work well, but for most small businesses the monthly cost outweighs the value unless the phone is genuinely the main sales channel.
For messaging apps, use a link with your own tag on it, and start the prefilled message with a short code that identifies the page. When the message arrives with that code, you know where it came from.
And keep the simplest habit of all: whoever answers the phone asks where the caller found you, then writes the answer down in the same place as everything else. Combined with the tagged data, it fills in exactly the part the tags cannot see.
Reading the numbers without fooling yourself
Once the data starts arriving, there are four traps that catch almost everyone, and knowing them in advance saves months of misdirected budget.
The first is judging by last click. A customer who saw an ad, read two articles, and later searched your company name will be credited entirely to direct traffic or branded search. That does not mean the ad did nothing. It means the ad did its job earlier in the sequence.
The second is counting enquiries instead of revenue. Cheap channels tend to produce cheap enquiries. Ranking your channels by closed work, not by form submissions, frequently reverses the order completely.
The third is judging too early. Below roughly thirty leads per channel, the differences you are seeing are noise. For most small businesses that means a quarter of data before any confident decision, not a fortnight.
The fourth is forgetting the delay. In services with a long consideration period, the enquiries arriving this month often belong to spending from two months ago, and cutting a channel based on this month alone is a reliable way to misread it.
A practical review rhythm looks like this: once a month, list every enquiry with its source and outcome, calculate cost per closed job for the paid channels, and change at most one thing. Once a quarter, look at the full picture and decide where the budget goes next. Anything more frequent is reacting to randomness.
FAQ
Do I need Google Analytics to track where leads come from?+
It helps a great deal, but the most valuable part of this setup does not depend on it, because a hidden field in your contact form that records the traffic source and the landing page will arrive with every enquiry regardless of which analytics tool you use or whether the visitor accepted cookies. Analytics gives you the aggregate view, meaning how many people arrived from each channel and how they behaved, while the hidden field gives you the individual view, meaning which specific enquiry came from where, and it is the second one that lets you connect a channel to actual revenue. If you would rather avoid Google entirely, a privacy focused tool such as Plausible covers the aggregate side perfectly well for a small business site, and the form field works exactly the same way underneath it.
What should I do about leads that arrive as direct traffic?+
Direct traffic is the category analytics uses when it cannot identify a source, so it is a mixture of several very different things rather than a single channel: people who typed your address, people who clicked a link in a messaging app or a PDF, people returning from a bookmark, and people whose original source was lost somewhere along the way. The practical response is to reduce the size of the bucket rather than to interpret it, which means tagging every link you control, including newsletters, QR codes, partner listings and the link in your email signature, so that fewer visits fall through into the unknown pile. What remains after that is genuinely direct, and it is usually a healthy sign, because it consists largely of returning visitors and people acting on a recommendation, both of which tend to convert better than anything you can buy.
How long before this data is worth acting on?+
Plan on a quarter before making budget decisions, because the honest threshold is roughly thirty leads per channel, and a small business rarely reaches that in a month on any single channel. Two forces make short windows misleading: normal random variation, which can easily make a channel look twice as good or twice as bad over a fortnight, and the delay between someone seeing your advertising and actually enquiring, which in service businesses with a long consideration period regularly stretches to several weeks. Start collecting on day one anyway, because the data cannot be recovered retroactively, but treat the first month as setup verification, checking that tags appear correctly and the form field is populated, rather than as a source of conclusions.