Customer Impact

Leadgeneratie

Lead Source Tracking in Your CRM (Without Losing the Data)

Copy for AI

Lead source tracking in your CRM records which channel brought each lead in.

It only works if a later interaction can’t overwrite that source.

The verdict: lost lead sources are almost never a measurement error. They’re an overwrite problem, and you fix them with two fields, hidden fields and explicit rules.

Here’s the typical symptom. Your marketing team swears most leads come in through LinkedIn.

Your sales CRM says almost everything is “direct” or “unknown”.

Both are right within their own system, and that’s exactly the problem.

The lead source wasn’t measured incorrectly. It was overwritten somewhere along the way.

Below: why that happens and how to get lead source tracking CRM data right, so the source carries through into your deals instead of evaporating at the first follow-up.

This is a follow-up article within our pillar on lead generation.

That piece explains how to build pipeline structurally.

Here we zoom in on the registration layer: how every lead brings its origin along.

The whole chain has four links:

FROM CLICK TO DEAL The four links of source registration 1 Hidden fields capture the context 2 Two fields first and last touch 3 Overwrite rules protect the data 4 Through to the deal attribution per channel Drop one link and you are back to 'direct' and 'unknown'.

Why the lead source disappears

Because one field has to answer for several interactions, and the last one wins.

In many CRMs there’s one single field called “source” or “lead source”.

Some do keep two. HubSpot, for example, has an Original Traffic Source and a Latest Traffic Source: the first known source and the most recent one.

One lead, four interactions

A single field sounds logical until you realise a lead touches your business several times. For example:

  1. someone finds you through a Google search
  2. reads an article
  3. clicks a LinkedIn post later on
  4. fills in a form after receiving your newsletter

That isn’t a lead with one source. Those are four interactions fighting over the same single field.

Here’s what each setup keeps from that journey:

EXAMPLE JOURNEY What one field keeps, and what two fields keep Google search first visit Article reads on LinkedIn post later click Newsletter, form converts One field: source = email (organic search is gone) Two fields: first touch = organic search, last touch = email Same lead, same clicks. Only the field design decides what sales gets to see.

The last interaction wins

What happens next depends on your settings.

In the worst case, every new interaction erases the previous one.

The first real source (organic search) gets overwritten by the last one (the email click).

Or by “direct”, if someone simply types your URL.

By the time sales opens the lead, the source says nothing about how the interest came about.

The data isn’t gone because of a bug. It was neatly overwritten by your own logic.

An overwrite problem, not a measurement error

That’s the core of it: losing the lead source is almost never a measurement error.

Our experience: the recurring CRM problems we see are data out of sync and deals not updated or attributed to the right team.

It’s worse in mid-market teams with a diverse offer.

So you don’t solve it by measuring better. You agree what may happen to the source once it’s recorded.

Split first touch and last touch

Keep two sources instead of one. That’s the first structural decision you make.

The first touch source

The first touch source records how someone discovered you for the very first time.

You fill this field once, at the very first interaction, and after that you never touch it again.

It answers one question: which channel kicked off this relationship?

That’s the information that justifies your marketing budget.

It shows which top-of-funnel channels bring in people you’d otherwise never have seen.

The last touch source

The last touch source records what caused the final conversion.

This field may update with every new interaction.

It answers a different question: which channel pushed this person over the line to fill in a form or get in touch?

That tells you which channels close sales-ready intent.

Analytics tools make the same split

In GA4, user-scoped “First user” dimensions keep the same value on later visits.

Session-scoped dimensions are reassigned each time a user returns.

Here’s a session-scoped view in our own GA4, where each visit gets its own source:

GA4 traffic acquisition report for customerimpact.be filtered on session source chatgpt.com: 58 sessions in 90 days, 0.22% of total, 50% engagement rate and 1 key event, an example of a session-scoped lead source dimension Our own GA4 traffic acquisition report, October 2026: chatgpt.com brought 0.22% of sessions in 90 days, yet 7.14% of all key events.

That’s a last-touch style view: in our account, chatgpt.com brought 0.22% of sessions but 7.14% of key events.

The User acquisition report shows the first-touch side for the same users.

Why one field isn’t enough

Both questions are legitimate, but they have different answers, so they need different fields.

Put them in one field and you implicitly let one question overwrite the other.

Most CRM default settings opt for last touch.

That’s how your first touch source quietly disappears. Two separate fields make the conflict visible and solvable:

First touch fieldLast touch field
RecordsThe very first channelThe channel of the conversion
Write ruleOnce, never againUpdates on relevant interactions
AnswersWhich channel starts relationships?Which channel closes intent?
Used forTop-of-funnel budgetConversion budget

Hidden fields: capturing the source at the moment that counts

At the moment of conversion, the source lives in the visitor’s browser. Your CRM doesn’t see it unless your form sends it along.

That context sits in three places:

  • the UTM parameters of the URL
  • the referring page
  • cookies that kept track of earlier visits

Between the form and your CRM, everything you don’t explicitly send along is lost.

What a hidden field does

A hidden field is a form field the visitor doesn’t see.

It’s filled in behind the scenes with the context of that moment: the source, the campaign, the landing page, the UTM values.

When the visitor clicks submit, those values travel along with the name and the email address.

They land in your CRM as real data instead of guesswork.

Here’s what such a field looks like in a form builder:

The UTM values to capture

Google’s URL builder guide defines the three core parameters:

  • utm_source: the referrer, such as google or newsletter4
  • utm_medium: the marketing medium, such as cpc, banner or email
  • utm_campaign: the product, slogan or promo code

Watch out: values are case sensitive. utm_source=google and utm_source=Google count as two different sources, so agree on lowercase.

Three rules for hidden fields that don’t leak

A few rules make the difference between working and leaking hidden fields:

  1. Fill them on page load, not only on submit. Context can change in a long session. You want the source as it was when the visitor arrived.
  2. Preserve the source across pages. Someone lands on a blog article with UTM parameters and clicks through to your contact page. There the parameters are gone, so store them in a cookie or the session.
  3. Always pass a value, even “direct” or “unknown”. An empty CRM field is ambiguous: did nothing come in, or was it overwritten?

Hidden fields are the bridge between the moment intent arises and the system sales works in.

Without that bridge, your source registration only starts inside the CRM.

By then half the story is already gone.

Running Google Ads? Capture the click ID the same way, as in storing the GCLID in your CRM.

Set your source overwrite rules explicitly

Now that the source arrives safely, decide what your CRM may do with it.

These are your overwrite rules.

Most teams have never set them deliberately; they simply run on the defaults.

Rule for the first touch field: only if empty

The rule is simple: only fill it if it’s empty.

If a value is already there, it stays there, forever. This is a write-once field.

Many CRMs call this “set only if blank” or something similar.

Check that it really is set that way, because the default is often “always update”.

Keep in mind that a person can still change it by hand. HubSpot, for instance, lets you edit the original source on a contact record, which clears its drill-down values.

Rule for the last touch field: update, with limits

The rule here is the opposite: update on every new relevant interaction, but with limits.

An internal click or an email you sent yourself shouldn’t rewrite the source to “email”.

Not when the person actually came in through a paid campaign.

Agree which interactions deserve an update and which ones you ignore. For example:

InteractionUpdate last touch?
New paid campaign clickYes
New organic visit from a searchYes
Internal navigationNo
Click in a sales follow-up emailNo

Write the rules down

Write these rules down literally, in plain language, on a single page that marketing and sales both know. For example:

“First touch is fixed at the first interaction. Last touch updates on every new campaign click. Internal navigation changes nothing.”

Three sentences that stop someone from building an automation six months from now that wipes out your source history.

Without shared rules, your registration is only as strong as the last person who sat at the settings.

Let the source carry through into your deals

A source stored on the contact but nowhere else is half a solution.

The value only appears once it travels to the deal.

That’s where the question that really matters gets answered. Not “which channel delivers leads”, but “which channel delivers deals”.

Copy both fields at deal creation

Copy the first and last touch source from the contact to the deal at the moment it’s created.

Then you can trace won revenue back to the channel that started the relationship and the channel that closed the conversion.

HubSpot’s lifecycle stages show where that hand-over sits:

HubSpot Knowledge Base page listing the default lifecycle stages with the definitions of marketing qualified lead and sales qualified lead, the stages a lead source should travel through to the deal HubSpot’s lifecycle stage definitions, checked 18 October 2026. Marketing hands over the MQL, but only sales decides what counts as an SQL.

The source should survive every one of those stages. Our guide to the lead handoff from marketing to sales covers the moment it most often breaks.

Read deals per channel, not leads per channel

That’s the difference between lead lists and qualified pipeline.

You stop steering on lead volume per channel and start steering on deals per channel.

For example, a channel that produces plenty of leads but rarely deals looks great in your lead report.

In your deal report it looks weak.

Only the second report decides where your budget goes. Our piece on conversion rate by lead source shows how to read it.

From registration to attribution

That way your source registration stops being an administrative formality.

It becomes the basis of your lead to deal attribution.

That’s the core of a measurable growth system in which every euro is traceable to pipeline.

Which model weighs the touches in between? See how to choose an attribution model.

Frequently asked questions about lead source tracking in your CRM

Why does my CRM show so many leads as “direct”?

Usually because a later visit overwrote the original source, or because a form sent no source at all. Two separate source fields and hidden fields with an explicit value fix both.

What’s the difference between first touch and last touch?

First touch is the channel that started the relationship and never changes. Last touch is the channel of the most recent relevant interaction and may update.

Do I need hidden fields if my CRM tracks traffic sources?

Often yes. Hidden fields send the UTM values with the form itself, so the source still arrives.

Source registration is the capture layer, not the destination

A correct lead source in your CRM isn’t a standalone trick.

It’s the capture layer of your entire growth engine.

The four links again:

  • Hidden fields catch the context
  • Two separate fields preserve first and last touch
  • Explicit overwrite rules protect the data
  • Carrying through into deals makes it usable for decisions

Drop one of those links and you’re back to “direct” and “unknown”.

Check your setup with three deals

Open three recent deals in your CRM and ask:

  1. Does each one show a first touch source?
  2. Is that source something other than “direct” or “unknown”?
  3. Does it match what the contact record says?

Three no’s means the data leaks somewhere. Our step-by-step lead tracking setup shows where to start, and a tracking audit finds the leak for you.

How we approach it

At Customer Impact we treat this not as a CRM setting but as the foundation under every campaign.

Want your lead generation built on data that carries through into your pipeline? Then explore how a B2B website that generates leads and your source registration work together.

Or read how to calculate the value per lead once your sources are accurate.

Ready to make your lead registration watertight? Get in touch and we’ll look at where your data is leaking today.

Free website scan

Enter your website and get an automatic scan within minutes, with concrete technical and SEO improvements. No sales pitch.

Where should we send your report?

We only use your details for your scan. No spam, unsubscribe anytime.