Mtools

Content

Data

GA4 will never match your CRM: stop reconciling and explain the gap

GA4 counts events in the browsers of people who agreed to be measured; the CRM counts people. See what causes the gap, when it signals an error and the table to bring to the meeting.

A frosted glass 3D tape measure with a red lock button glowing on its round case, on a black background

GA4 vs CRM numbers will never match, and that doesn't mean something is broken. GA4 counts events fired in the browser of people who agreed to be measured; the CRM counts records of people, whether they came through the website or not. Trying to make the two match eats hours and still leaves leadership distrusting both. Credibility comes from explaining the gap before the meeting, with the expected size and the reason behind each part.

Say GA4 shows 312 conversions for the month and the CRM shows 240 new leads. Both can be right at the same time.

What each tool counts

GA4 measures actions in a browser; the CRM stores people and companies. The difference in units explains a good part of the distance:

GA4 CRM
Unit event, session and user per device record of a person (and their company)
When it counts every time the event fires when the record is created or updated
Who is left out anyone who declined cookies or blocked measurement nobody who became a record
What gets added repeated submissions and team tests leads from events, phone calls and imported lists

In GA4, the action that counts as a result is marked as a key event. According to Google's help page on the counting method, the recommended option is “Once per event”: if someone triggers the event five times in a session, five go into the count. For that same person, the CRM keeps one record.

GA4 is still the best source for traffic origin and on-site behavior, as reading ChatGPT traffic in GA4 shows. The mistake is asking it for a head count.

Why GA4 vs CRM numbers never match

Four causes show up in almost every account, and none of them gets fixed for good. It's the same issue you see when Meta Ads data doesn't match the CRM, seen from the website side.

People who decline analytics cookies disappear from GA4, fully or partly. According to the consent mode documentation, when analytics is denied no Analytics cookie is written. In the basic implementation, tags stay blocked until consent is given; in the advanced one, the site sends cookieless pings that feed modeling.

Modeling has a volume requirement. The behavioral modeling page asks for at least 1,000 daily events with analytics denied for 7 days and 1,000 daily users who granted consent on at least 7 of the previous 28 days. A corporate website with a few hundred visits a day falls short, and the gap stays open.

Blockers

Blocking extensions stop the tag from loading. The EasyPrivacy list, used by several blockers, describes itself as a filter that removes all forms of tracking. The form reaches the CRM and the event never reaches GA4.

Switching devices

Without a login, the same person on two devices becomes two users in GA4. According to the help page on reporting identity, without User-ID GA4 uses the device ID, which on websites comes from each browser's client ID. In the CRM, the person who researched on a phone and filled in the form on a laptop is a single record.

Duplicate forms and leads from outside the website

The same lead counts twice in GA4 and once in the CRM, or the other way around. Someone who submits the form twice generates two events; if the CRM identifies people by email, that becomes one record. At the other end, the lead from the trade show booth, the phone call or the imported list enters the CRM without any event on the website.

When the gap is normal and when it's an error

A normal gap is stable: the ratio between the two numbers barely changes from one month to the next. Be suspicious when:

  • the ratio changes size with no change to the website or the cookie banner;
  • GA4 shows zero submissions on a day when the CRM received leads from the website;
  • GA4 starts counting far more than the CRM after a change (an event firing twice, a thank-you page reloading);
  • paid media leads reach the CRM without a source, a sign that a UTM standard is missing or that the click got lost along the way.

The reconciliation table to bring to the meeting

Bring the gap already explained, line by line, before anyone asks. A template:

Gap Which one is bigger Reason Where to measure it
Declined cookies CRM GA4 doesn't measure those who declined banner decline rate
Blockers CRM the tag doesn't load no direct measure, kept as a margin
Repeated submissions GA4 each submission counts as an event duplicate records in the CRM
Leads from outside the website CRM events, phone, imports record source in the CRM

Fill it in with the month's numbers and keep the same table every month. Building this reconciliation, and the dashboard that brings both numbers together, is part of our data and dashboards work.

Talk to usWant help planning this at your company?Tell us what you need. We'll reply by email or WhatsApp and, if it makes sense, put together a quote.Contact us

Which number to use for each decision

Each question has its own source, and the rule is written down for everyone.

  • How many leads marketing delivered and what each one cost: the CRM, which counts people. It's the basis for reading paid media by campaign.
  • Which page or campaign converts better on the website: GA4, comparing rates, without adding it to the CRM.
  • Pipeline, revenue and acquisition cost: the CRM, with the metrics leadership understands.

With the rule agreed, the meeting discusses results, and the GA4 vs CRM gap becomes a known line in the report.

Frequently asked questions

Why does GA4 show more conversions than the CRM?

Because the key event, under the recommended method, counts every submission, including repeated ones and team tests. The CRM merges submissions from the same person into one record.

What is an acceptable GA4 vs CRM gap?

There's no universal percentage. The acceptable gap is the one in your own history: stable from month to month and explained by the causes in the table.

It narrows the gap when the property reaches the volume required for modeling. On low-traffic websites, the gap remains.

Share

shall we createsomething together?

related content