GA4 Is Lying to You About B2B Conversions

GA4 reported 4,000 conversions last quarter, and the CRM got 180 leads.

Both numbers are correct, and only the second one belongs in front of your CEO.

Two tools, two different questions

GA4 was built to count things that happen on a website. Lots of them, fast.

B2B demand gen doesn't work that way. A handful of people, a long buying cycle, a committee that never visits your site at the same time.

So when GA4 conversion tracking runs on its defaults, it does exactly what it was designed to do. It counts activity. And activity looks a lot like progress until someone asks how much of it turned into pipeline.

If your conversion rate went up and pipeline didn't, you're counting the wrong thing.

How it drifts

Someone marks scroll depth as a key event. Then a pricing page view. Then a video play. Each one made sense at the time.

Six months later, the "conversions" report is mostly people reading.

The decision that fixes most of it

Before you touch GA4, answer one question:

What does a prospect do right before sales gets involved?

For a B2B team the list is short:

  • Requests a demo or a call
  • Submits a contact form with a company email
  • Registers for a webinar or event
  • Downloads something only a buyer would want

Those are your key events. Everything else is a metric worth watching.

Four to six key events. More than that and the conversion rate stops meaning anything.

Expect the conversion rate to drop when you cut key events down. It isn't your marketing that changed, just what GA4 counts.

Set up conversions in GA4 the B2B way

A quick vocabulary note. In 2024 GA4 renamed conversions to key events. "Conversions" now means the key events you send to Google Ads. Same idea, new label, and it's why the old toggle seems to have disappeared.

Warn leadership before you do this. A number that drops overnight needs a sentence of context.

The setup I'd use:

  1. One event for lead actions. GA4's recommended generate_lead event works. Add a form_name parameter so a demo request and a newsletter signup don't look identical.
  2. Fire it when the form succeeds. Not when the button gets clicked. Button clicks count typos, double clicks and people who gave up halfway.
  3. Mark only that event as key. Admin, then Events, then the key event toggle. Leave page views and scrolls alone.
  4. Test it before you trust it. DebugView and Tag Assistant. Submit a form, watch the event fire once.

An event that fires twice is worse than one that doesn't fire at all, because the inflated number looks believable.

Where your UTMs show up in GA4

Tagged links are easy to lose track of in GA4, because it doesn't have a report called "UTMs." Each parameter becomes a dimension with a different name.

UTM parameterGA4 dimension
utm_sourceSession source
utm_mediumSession medium
utm_campaignSession campaign
utm_contentSession manual ad content
utm_termSession manual term

Three places to look:

  • Traffic acquisition (Reports, then Acquisition). Switch the dimension to Session campaign. Then set the key events column to generate_lead only, so you're looking at leads by campaign and not every event mixed together.
  • User acquisition. Same idea, but "first user" dimensions. Where someone first came from, not where this visit came from.
  • Explore, Free form. For utm_content and utm_term. The standard reports hide them.

Two things that look like bugs and aren't. GA4 is case-sensitive, so "LinkedIn" and "linkedin" are two different sources. That's a naming convention problem. And "(not set)" or "(direct) / (none)" usually means the tag never made it at all, which is its own leak (see why direct traffic is eating your attribution).

If the tags reach GA4 but never reach the CRM, the form is dropping them (hidden form fields fix that). If your funnel crosses into a scheduler or an app domain, check cross-domain tracking too.

Reconcile GA4 with your CRM

They'll never match exactly, because they answer different questions.

GA4 tells you how traffic behaves and which channels bring the right visitors. Your CRM tells you which of those visitors became pipeline.

Once a month, pull generate_lead by campaign from GA4 and new contacts by campaign from the CRM. Put them side by side. If the gap suddenly widens, a tag, a form or a redirect broke. You just found it before the quarterly review did.

That's the sense in which GA4 lies about B2B conversions: it answers the question its defaults ask, and for B2B the defaults ask the wrong one. So give it a better question. Count fewer things, and report pipeline from the CRM.

Skip the build

The B2B Attribution System

The locked UTM system, GA4 and CRM setup, the model decision guide and the dashboard. $399.

See the Attribution System

Not sure your data is ready? Run the free audit first.

The newsletter

Notes from the Field.

Every other week, one broken thing and how I’d fix it.

One email every other week · unsubscribe anytime · Privacy policy