The short answer
To track where a website lead came from, store each lead with the page it was sent from, the source of the visit (a referrer, an ad click id or a tagged link) and the pages read before it. Those three facts read every enquiry back to the marketing that caused it, without hidden fields or a CRM project.
What to know
- A count of submissions by channel cannot answer a customer who asks how they found you. The stored lead can.
- Up to four clues name a visit's source: the referrer, an ad click id, utm tags, or a tracking link. Direct is what is left when none is present.
- A Meta ad click from inside the Facebook app arrives with an fbclid and no referrer. Drop the id and it reads as direct.
- The page the form was on and the pages read before it tell you which content did the work, per lead.
- Phone taps are leads with a source too. Count only forms and the report describes a minority of your enquiries.
To know where a lead came from, you need the lead itself stored with three facts attached: the page the form was sent from, the source of the visit, and the pages read before it. A count of form submissions by channel does not do it, because the moment one customer asks how they found you, you need the one row and the total is no use.
This is what I built the Inbox in Venbit Intelligence to hold. Every form submission and phone tap is stored outside the email path, with its page and its source, so when a customer says they filled out your form last Tuesday, you can look. The screens in this article are from the demo site, WebsiteMaintenance.com, and every number read from them is that site's. Before the screens, try the part that decides everything else: paste a link from one of your own ads or emails and see where the visit would be filed.
Read a link
See where a link's visits get filed
Pick a case or paste the address from one of your ads, posts or emails. The answer comes from the same rules the tracker applies to every visit.
Try a common case

This visit is filed under
Social
The fbclid is the clue that decides it. It names Facebook or Instagram, so the visit is filed under Social instead of Direct.
- Referrer
None sent
Apps, text messages, most email programs and typed addresses send no referring page.
- Click id
fbclid (Facebook or Instagram)
Facebook or Instagram adds this to links in ordinary posts as well as ads. It proves the platform and leaves paid or unpaid open.
- Tags
None
No utm_source, utm_medium or utm_campaign on the address.

Make it name its source. If this link is an ad, add utm_medium=paid (or cpc) to the ad's web address. With that tag the same click files under Ads and the lead names the campaign.
The verdict comes from the same rules Venbit Intelligence applies to every visit, which follow Google's published channel definitions. Nothing you paste leaves your browser.
What you see for each lead
The lead above was sent from the home page by someone on an iPhone in Minster, Ohio, on their first visit, after two pages and one minute twenty-three seconds. The source line reads Meta ad, Website Maintenance Campaign 1. Under the message sit a notes box and a button to the visitor trail, which is the step-by-step record of that visit. Those are five facts, from four places.
Device and pages
The tracker's own events.
Town
IP geolocation, so it is approximate.
First visit
From the visitor ID, a random identifier in the browser with no cookie behind it.
Source
The fact this article is about, read from the landing address and the referrer at the moment the person arrived, long before they typed anything.
Where a visit's source comes from
A visit arrives with up to four clues about where it came from.
| Clue | Who puts it there | What it tells you |
|---|---|---|
| Referrer | The visitor's browser, when they clicked a link on another site | The site they came from: google.com, facebook.com, a directory, a news story |
| Click id (gclid, fbclid, msclkid) | The ad platform, stamped on the landing address of every paid click | Which ad platform's link sent the visit. gclid and msclkid are only ever issued for an ad. fbclid alone does not prove a paid click, so a Meta click counts as an ad only when the link is also tagged as paid |
| Tagged link (utm parameters) | You, whoever made the link, or the platform's URL template | The campaign, the ad set and the ad, as named |
| Tracking link (vnbt.app/name) | You, for a flyer, a bio or a mailer you cannot tag | The visits and leads that one link brought, under the name you gave it |
Direct is the row that is not in the table, because it is not a source.
Direct means the browser sent no referrer and the address carried no tags: a typed address, a bookmark, a link tapped in a text message or an email app. It is the absence of information. A lead filed under direct is a lead whose source you do not know, and reading it as a loyal customer who typed your name is wishful.
The click ids are the clue that matters most for paid traffic, because they arrive on sites nobody tagged. A Meta ad click from inside the Facebook app comes with an fbclid, no referrer and no utm parameters.
Click id dropped
- The visit reads as direct
- An owner paying Meta sees their ad traffic filed under came straight to the site
Click id kept
- The id stays with the visit, and then with the lead
- Each lead names the ad platform that sent it
The channel names on the dashboard (Search, Social, Ads, Direct) follow the rules Google publishes for GA4's Default Channel Group, because that is the only taxonomy anybody has standardised. A Google Business Profile link tagged utm_source=google and utm_medium=organic reads as search, which for a local trade is where much of the business starts.
Read a link the way the tracker does
The six cases in the tool are the ones I see go wrong most, and working through them teaches the rules faster than a definition does.
Start with the Meta ad. The address carries an fbclid and nothing else, with no referrer, because the click happened inside the Facebook app. The visit files under Social rather than Ads, and that surprises owners who are paying for the click. Meta's documentation describes fbclid for ad clicks and says nothing about ordinary links, so the tracker reads the id as proof of the platform and not of the payment.
Untagged
- The address carries an fbclid and nothing else
- The click files under Social
Tagged utm_medium=paid
- Add it to the ad's web address in Ads Manager
- The same click files under Ads, with the campaign name on the lead
Google ad
Needs no help. A gclid is issued for ad clicks only, so its presence settles the question even on a link nobody tagged. Microsoft's msclkid behaves the same way.
Business Profile button
The case that costs local businesses the most. A plain website link on a Google Business Profile leaves the visit with a bare Google referrer or with none, and nothing on it says the listing sent it. Tag the link with utm_source=google and utm_medium=organic and the visits read as Search, which is what they are: someone searched, saw your listing and pressed the button.
Text message
The reminder that some sources cannot be recovered afterwards. A link pasted into a text carries no referrer and no id. The only way to know those visits later is to mark the link before you send it.
Tag the links you hand out
Ad platforms and search engines identify themselves. Every other link is yours to label, and the label takes a minute.
- 1
Google Business Profile
Set the website field to your address with ?utm_source=google&utm_medium=organic&utm_campaign=gbp on the end.
- 2
Meta ads
In the ad's URL parameters, add utm_source=facebook&utm_medium=paid and a utm_campaign with the campaign's name.
- 3
Email
Give every link in a newsletter utm_medium=email and a utm_campaign naming that send. Most email tools have a setting that adds these for you.
- 4
Texts, flyers, QR codes and signs
Use a short tracking link with its own name for each one. Tracking links for QR codes, flyers and posts covers how.
- 5
Directory listings and partner sites
Leave these alone. Their referrer names them, and a tag would only overwrite a source that is already correct.
Keep the names boring and identical every time. Facebook, facebook and FB are three sources in most reports, and a campaign called test2-final is no use to you in March.
How much traffic reads as direct
- 32%
- Of visits arrived direct on the median tracked website
- 1 in 3
- Tracked websites where direct was 40% or more of everything
A third of traffic with no stated source is normal, and it is not all loyal customers typing your name. Direct holds the bookmarks and the typed addresses, and it also holds every untagged text, every link opened from an email app, and every click from an app that strips the referrer. The share falls when you tag what you send out, and it never reaches zero.
A report that shows no direct traffic at all is guessing.
See lead sources for the whole site
347 visits
Ads22164% of trafficDirect7020% of traffic
AI assistants267% of traffic
Social247% of trafficSearch51% of traffic
- Other sites10.3% of traffic
The dashboard's sources chart is the same information added up. On the demo site the last 30 days read Ads 221, Direct 70, AI Assistants 26, Social 24 and Search 5, against 12 leads.
The total is useful for proportion and useless for a single enquiry.
The campaigns section does the join, shown here for Website Maintenance Campaign 1 on Meta and on Instagram.
Campaign 1 on Meta
- 137 visits
- 7 leads
- 5% got in touch, as the screen reads it
The same campaign on Instagram
- 61 visits
- 2 leads
- 3% got in touch
Those are leads counted on the site, from the click landing onward, so they are clicks that reached the page and what the visitor did next. The tracker does not see what you paid, so there is no cost per lead here. Whether the ads are working is still answered, in leads and in the rate at which clicks became them.
Which page a lead came from
Every row in the inbox names the page the form was on. The lead opened above was sent from the home page, where the demo site keeps its quote form. On a site with a contact page and six service pages, this column tells you which service page is doing the work, and the trail tells you which pages were read on the way to it. Read the two together.
- 4
- Pages a typical lead reads before writing
- 5m 26s
- About how long a typical lead spends before writing
- 75%
- Of leads came from the ads
A lead that read one page and wrote is a different customer from one that read the pricing page, the reviews and the service area first, and the trail shows which one you have in front of you.
A trail names the fields a person touched by their labels. Started filling the quote form means a field was focused, nothing more. What was typed is on the lead in the inbox, which sits behind the leads permission, and the trail shows it only to the people who hold that permission.
Phone taps and email clicks are leads too
A tap on your phone number
- Lands in the same inbox as the forms, with its page and its source
- Sends a push that reads Someone just tapped your phone number
A click on your email address
- Counts as a lead on that visit
- Has no row of its own in the inbox
- Sends a push that reads Someone just clicked your email address
A tap shows intent. It does not mean the call connected. The tracker knows the number was tapped on that page at that moment and nothing about whether anyone answered.
That matters for source tracking because on a trade site a good share of the enquiries never touch a form. If only the forms are counted, the source report describes the minority of your leads. What counts as a lead on your site is a setting worth deciding on purpose: a page reached, a button, a form, with the phone and the email as signals the tracker already records.
Tracking form submissions in Google Analytics
Two methods are common, and each one leaves something out.
An event in Google Analytics
- Counts a form submission once the form_submit event is marked as a key event. GA4 collects that event on its own for some standard forms, and other forms need it set up through a tag manager or code
- Splits that count by channel group
- Its policies forbid sending it anything that identifies a person, so the name, the email and the message never arrive
A hidden field in the form
- Copies the landing address or the campaign into the submission, so the source arrives inside the email
- Works, and fails without warning when the form is rebuilt, the plugin updates, or the email never arrives
- Tells you the source of the submissions and nothing about the visits that did not submit, which is where the money is being lost
A count by channel is as close as Google Analytics gets. For a marketing team that lives in GA4 that is enough. For an owner asking which enquiry came from which ad, it is a table of totals. What Google Analytics will and will not tell a small business covers the rest.
The same inbox on your phone
Venbit Intelligence for iPhone shows the same leads with the same source line, and a push arrives the moment one lands. Each lead can be marked new, contacted, won or lost, with a note kept on it, so by the end of a month the inbox is also the record of which sources produced the leads you won, which is the number that decides next month's spend.
See this on your own website
Venbit Intelligence installs with one line on your site. Free for 14 days, no card needed.
In Venbit Intelligence
More website analytics guides
Every question in this series, from Website Analytics for Small Business Owners.
Leads inbox4
- Website Traffic but No Leads? How to Find Out Why
- How to Track Where Website Leads Come From (you are here)
- How to Track Phone Calls From Your Website
- Get Notified When Someone Fills Out Your Contact Form
Ad performance4
SEO tracking3
Visitor behaviour4
Feed and notifications3
Goals and measurement4
Privacy, AI and teams3
Venbit Intelligence Research19
- Average Website Visitors per Month for a Small Business
- What Is a Good Conversion Rate for a Small Business Website?
- What Time of Day Do Website Leads Come In?
- Which Day of the Week Brings the Most Website Leads?
- How to Tell If a Website Lead Is Spam
- How Much Website Traffic Comes From ChatGPT?
- Do ChatGPT Visitors Convert Better Than Google Visitors?
- Which AI Bots Read Your Website?
- Why Ad Clicks Don't Match Website Visits
- Mobile vs Desktop Website Traffic for Small Business
- How Long Do Website Visitors Take to Contact a Business?
- Average Time on Site for a Small Business Website
- Average Pages per Visit for a Small Business Website
- How Far Do Visitors Scroll on a Website?
- Why Is My Website Getting Traffic From China and Singapore?
- What Is Direct Traffic and Why Is It So High?
- Where Does Small Business Website Traffic Come From?
- How Many Visitors Reach Your Contact Page?
- New vs Returning Visitors for a Small Business Website
Tony Carnevale
Founder, Venbit
Tony started Venbit in Mill Creek, WA in 2011 and still does the work: web design, SEO, and the website analytics product the company builds and runs for its own clients. The numbers and screens in his articles come from that product, not from a content mill.